From sol.ctr.columbia.edu!lll-winken!decwrl!shelby!med!hanauma!rick Thu Mar 15 02:17:38 EST 1990
Status: RO

Article 11880 of soc.culture.china:
Path: sol.ctr.columbia.edu!lll-winken!decwrl!shelby!med!hanauma!rick
>From: rick@hanauma (Richard Ottolini)
Newsgroups: soc.culture.china
Subject: chinese computing: direct input computer
Message-ID: <673@med.Stanford.EDU>
Date: 13 Mar 90 00:04:14 GMT
Sender: news@med.stanford.edu (USENET News System)
Reply-To: rick@hanauma (Richard Ottolini)
Organization: Stanford University, Dept. of Geophysics
Lines: 6


Sony Company announced a notebook sized computer last week for Japanese
Kanji word processing without a keyboard.  The user draws up to 3500
characters on the screen to input information.

Perhaps this can be adpated for Chinese use.  It solves the problem of having to learn pinyin, Big5 etc.


From sol.ctr.columbia.edu!samsung!cs.utexas.edu!wuarchive!udel!rochester!rit!ultb!jxy7070 Fri Mar 16 14:21:35 EST 1990
Status: RO

Article 11984 of soc.culture.china:
Path: sol.ctr.columbia.edu!samsung!cs.utexas.edu!wuarchive!udel!rochester!rit!ultb!jxy7070
>From: jxy7070@ultb.isc.rit.edu (J.X. Yih)
Newsgroups: soc.culture.china
Subject: Chinese in Computer
Message-ID: <2470@ultb.isc.rit.edu>
Date: 16 Mar 90 08:44:40 GMT
References: <673@med.Stanford.EDU>
Reply-To: jxy7070@ultb.isc.rit.edu (J.X. Yih )
Organization: Information Systems and Computing @ RIT, Rochester, New York
Lines: 76

This is a long stuff.  If you are NOT interested in Chinese-Computer,
press "n" now!!

In article <673@med.Stanford.EDU> rick@hanauma (Richard Ottolini) writes:
>
>Sony Company announced a notebook sized computer last week for Japanese
>Kanji word processing without a keyboard.  The user draws up to 3500
>characters on the screen to input information.
>
>Perhaps this can be adpated for Chinese use.  It solves the problem of
>having to learn pinyin, Big5 etc.

As far as I know, just like English, Japanese is a symbol-directed
language.  That means all words can be devided into the basic symbol.

I can not find any method to devide Chinese into some certain symbols.

I think that the basic steps of Sony system are as follows:
	1. user writes words on the screen
	2. computer recognizes the words symbol-by-symbol by
	   using some Artificial Intellgence program
	3. combining the symbols into the word
	4. do something (save or other action)

There are lots problem in Chinese:
	Chinese is much more complex than Japanese.  ==>
	It will much harder to recognize (especilly in hand-writing!).

Yes, we can treat every Chinese word as a symbol.  Then can you image
that how many patterns the poor computer has to remember?  Always
aware the limitation of computers!

On the Sony System, all stuff that computer has to remember is the X 
letters (for X<100).
(sorry, I don't know how many symbols Japanese has)  If it has no
spell-checking, it just combines these symbols into a word.  That's
not very difficult in computer.  In fact, I think that Sony System
has certainly included the spell-checking.  Sony is proud of his
technique (pattern-recognization is not an easy tech.) and having
their language in the computers.

Now, back to our topic.  Because Chinese is NOT a symbolic language,
if you have some concepts about computer, you can think about lots
of difficulies.

Why do we use Pinyin or Big5?  Because that's a method to represent
Chinese in a symbolic way.  The basic difficulties of Pinyin are:
	1. No standard
	2. One has try to "quess" (because of the same pronunciation)
	     ex:   "yi(3)" may be "chair"
				  "already"
				  "ant"   ..................

But if I write in this way:
	"yi(3) jin(1)"		  (already)
everyone can understand.  Why?  Why can you understand it?  The reason
is that I said a whole "phrase".

BTW, I got an idea:
	1. standarlize the Pinyin method
	2. build a translator based on a "Pinyin-Phrase Base"
Then we can set up a Chinese system like Sony did on Japanese.

I think that the standard of Pinyin is not very hard.  The key point is:
how should we build the "Pinyin-Phrase Base"?  You know, lots lots lots
of phrases in Chinese....

All the above is my personal opinion.
If you have any suggestion or idea, contact with me by e-mail

jxy7070@ultb.isc.rit.edu




a Chinese-lover


From sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!ohstpy!hjz Sat Mar 17 00:55:32 EST 1990
Status: RO

Article 12029 of soc.culture.china:
Path: sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!ohstpy!hjz
>From: hjz@ohstpy.mps.ohio-state.edu
Newsgroups: soc.culture.china
Subject: Chinese Computing.
Message-ID: <8108.26017e4f@ohstpy.mps.ohio-state.edu>
Date: 17 Mar 90 05:01:18 GMT
Lines: 10



Does anybody know something about IBM 5550 which was developed couple of 
years ago? It is capable of dealing with Chinese characters. I have some
experience with the machine and I don't like its coding system for Chinese
characters. I wonder if it is available here in the US. If it is, its 
Chinese word processing software should also be available, definnately.
There are perhaps hundreds of coding systems for Chinese characters, but as
far as I know, there is not a standard yet, and the so called  "Wu3 Bi3 zi4
xing2" method is the most popular in mainland China.


From CCNET-L@uga.cc.uga.edu Thu Feb 28 22:44:22 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.64/5.10) with SMTP id AA06910; Thu, 28 Feb 91 22:44:05 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 0146; Thu, 28 Feb 91 22:42:25 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 8073;
 Thu, 28 Feb 91 22:42:23 EST
Date:         Thu, 28 Feb 91 23:06:35 EST
Reply-To: Huangxin Wang <wang%PENNMESS.PHYSICS.UPENN.EDU@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L%UGA.BITNET@uga.cc.uga.edu>
From: Huangxin Wang <wang%PENNMESS.PHYSICS.UPENN.EDU@uga.cc.uga.edu>
Subject:      New Chinese input device?
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

A friend told me that a week ago in The Wasgington Post, there is an article
about a small Californian company invented a new Kanji (Japanese) pad
for Macintosh.  Input chinese character is by writing with a special
pen on that pad (rather than on the keyboard with pinyin).  The whole thing
(with the processor) is listed for only $70 in USA.  It said that the Mac's
showroom was crowded like the Tokyo's subway when it was anounced.

This is seem to be the most natural way for Chinese input.  Is the
similar thing for Mac Chinese system avaliable yet?  (Can Ed Lai
comment on that?)

Another question:  what's the status for the newest Chinese system for Mac
(I should say the nuisance of SGB and HGB really fan my enthusiam away,
I just stick to the old Chinese OS)?  Is the Chinese OS 6.05 available
on the D e v e l o p magazine, or other way?  Can Sun Kang or anyone else
comment on that?

		Huangxin Wang, University of Pennsylvania

From CCNET-L@uga.cc.uga.edu Fri Oct 26 22:12:24 1990
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.64/5.10) with SMTP id AA01219; Fri, 26 Oct 90 22:09:25 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 4367; Fri, 26 Oct 90 22:08:23 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 6528;
 Fri, 26 Oct 90 21:52:31 EDT
Date:         Fri, 26 Oct 90 22:03:37 EDT
Reply-To: Huangxin Wang <wang%pennmess.physics%UPENN.EDU@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Huangxin Wang <wang%pennmess.physics%UPENN.EDU@uga.cc.uga.edu>
Subject:      New Chinese OS for Macintosh (6.03)
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

I just obtained the new Chinese OS 6.03 for Mac from the CD disk come
with the July issue (3rd issue of this year) of D e v e l o p Magazine
($10).  I am fortunate enough to find a CD ROM to down load it.  As
everyone believes, update version always better, so my expectation for
it was very high.   But so far it gives me big disappointment.  I
really wish that I did something wrong in the installation.  Here is
what I encounter:

(1)  (a very serious one!)  New system use the GB code (the same as in
CCDOS), while the old system (CH-4.3) use shift GB, so ALL my text
files created by my old Chinese OS cannot be read!!!   They are all
displayed as junk characters, including the file names.

(2)  Very unstable:  Since it comes without an installation script for
the whole system (only for the AppleShare), I had to install it by copy
the system folder to my hard disk.  But somehow it just crash when I am
typing the word "he2" ("river"), no matter whether I am in finder, or
in MacWrite.  Such crash is reproducible.  It even worse that after it
crashes, the CDEV file "pinyin", "bihua", etc, are corrupted.  Next
time when I boot, I cannot type in any Chinese character!  The only way
to cure it is to reinstall the original CDEV files.  I try many
combinations and eventually track down the problem:  if I choose
"pinyin" cdev, and check the option (1) Automatic; (3) "Chuzhi" (Keep
the characters on bar); (4) Choose using number; (5) Beginner, (with these
options the system will rearrange the character which you just
character),  then it will always bomb when you type "he2 ".  Right now
I have to keep the minimal configuration and default options to avoid the
crash (and I'm not sure whether I have really avoided it or not).

(3)  The most wanted improvement for the old system is the character
font, not the input method (which is already easy enough and much
better than a lot of IBM PC package on the market, except, in my
personal point of view, TianMa).  However, this new Chinese OS has no
improvement on the character font.  We still get those large
charatcers, combine with the clumsy looking Chicago font for the
English characters.  I believe that it's possible to use ResEdit to set
the default English character for Beijing to something more beautiful.
But should the user really have to go through all these?   I think the
ideal Beijing font should be Chinese characters combined with New York
14 or other typewriter like font).   At least the application font
should be such.

	Now let me talk about something more encouraging before all
these problem washing away your enthusiasm:  In general, IF the new
system WON'T crash, it has quite some improvements still comparing with
the old system:  (1)  Input is faster, since you can turn on a 64K
cache for character picking.  (2) the bar for picking characters is
very neat, with hierachical menu and tear off keyboard, which is really
helpful for "Bi3hua4" input (or when you mixing pinyin with Bihua
input).  (3)  To improve the input speed a little bit, now you can
group input; also single click (rather than double) to pick a
charatcer.  (4)  I can run MultiFinder, and if it won't crash, I
might consider to totally replace it for the regular English system
(6.05).

	Still it is in Macintosh way --- you don't have a manual (and
you don't need one to use).  I look up and down in the CD disk, and
indeed there are no documents related to this Chinese OS (there is all
the European language system, as well as 12 disk Taiwan system with
Kai-xu and song-ti fonts.  For some reason, there is no Korean  system
on the disk).  I supposed the systems themselves should be FREE (if
Apple should charge for it, it would not be just $10).  Actually it can
be ftp from the AppleLink ftp site, but I just don't know where is such
site (anyone knows?)  Also as it's a MACINTOSH system mean that you can
use a lot of Mac softwares without writing a specific word processor or
graphics program for Chinese.  (BTW, the top level word processor for
Chinese or other foreign language system is Nisus.  MSWord won't be
work properly to display chinese character, but so far I just use
MacWrite).

	Last week I had a chance to see a lot of Chinee word processor
demos at a far east language processing conference at Penn (thank Long
Zhihua!) I am surprised that Mac OS is in such a low key ---- almost
everything is on IBM PC, wth regular price around $200 (e.g.
BrushWriter for IBM: full version $700, student reduce version $80).
And I am also surprised to find that Mac is much better in the
character picking input method, --- for example, Mac is very tolerant,
you can mix pinyin and bihua input.  There is another class of word
processors which use a lot of Macro definition (like "jttqhh" to stand
for "jin-tian-tian-qi-hen-hao"), I think it's only good for fooling
around for demo (some list price $299).   Of course Mac can also do
this kind of macro (like in cihui input method by Ed Lai).  But the
best input method, I think, is TianMa, which does not use character
picking in inputting, --- the major difference is that you don't need
to keep your eye one the character bar to pick your characters, thus it
won't interrupt the user's flow of thinking.  You just keep typing in
pinyin and the program will convert it to Chinese for you (by context
or grammar).   Of course right now Mac OS is NOT a word processor, so
it's not fair to compare it with TianMa, but I think development effort
should be in this direction, and I see that Mac Chinese OS is a very
good platform for doing this.

	And now come back to the problem of the Chinese system 6.03:
Does anyone write a utility to convert the old GB code (shift GB) to
the new one?  Is there an easy way to convert my chinese file created
on the old system?  I look at the Apple File Exchange Utility, but
there is nothing for this purpose.  Also, for text file, it might be
easy to convert, but what about text in HyperCard?  (I think the
convenient way is to write it as a FKey to convert a certain selection
of text).

	Something is very obviously wrong --- it just should not be
this way, but it did happen to me.  Has anyone out there use the new
system?  Have you run into these problems and how to resolve them?

		Huangxin Wang, University of Pennsylvania

From CCNET-L@uga.cc.uga.edu Thu Nov  7 10:41:12 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA22354; Thu, 7 Nov 91 10:40:55 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 5200; Thu, 07 Nov 91 10:39:34 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 7104;
 Thu, 07 Nov 91 10:37:53 EST
Date:         Thu, 7 Nov 1991 10:31:51 EST
Reply-To: Ned Walsh <walsh%CUNIXA.CC.COLUMBIA.EDU@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ned Walsh <walsh%CUNIXA.CC.COLUMBIA.EDU@uga.cc.uga.edu>
Subject:      Radical/stroke input file
Comments: To: CCNET-L@uga.cc.uga.edu
Comments: cc: harbsmei@hedda.uio.no
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

Urs,

It would be great if you'd post the 700-char radical/stroke file.  It
would go a long way toward solving the sorting problem I mentioned a
few weeks ago--i.e. how to get Big5 or GuoBiao files to sort in some
useful order.  That would leave only 5000 characters to assign
radical/stroke numbers.

Has anyone already come up with a solution to this problem?  Christoph,
how did you sort those concordances you were working on?

/Ned

From CCNET-L@uga.cc.uga.edu Thu Nov  7 16:16:18 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA25103; Thu, 7 Nov 91 16:15:46 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 8006; Thu, 07 Nov 91 16:15:46 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 0091;
 Thu, 07 Nov 91 16:15:45 EST
Date:         Thu, 7 Nov 1991 15:16:55 EST
Reply-To: Kam Mok <inmet!kam%UUNET.UU.NET@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Kam Mok <inmet!kam%UUNET.UU.NET@uga.cc.uga.edu>
Subject:      Re: pinyin input in ETEN
Comments: To: uga.cc.uga.edu!CCNET-L@uunet.uu.net
Comments: cc: c@uunet.UU.NET
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

     Re: pinyin input in ETEN
Hi Urs,
Yes I think your PHRASE.BOXes will be quite useful to many others,
I also want to see that you post them to some archive so that people can ftp.
Thanks again for sharing them with us.

Kam
kam@inmet.camb.inmet.com

From CCNET-L@uga.cc.uga.edu Thu Nov  7 16:56:26 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA25414; Thu, 7 Nov 91 16:56:07 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 8319; Thu, 07 Nov 91 16:55:29 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 2887;
 Thu, 07 Nov 91 16:55:28 EST
Date:         Thu, 7 Nov 1991 16:54:24 -0500
Reply-To: Stephen G Simpson <simpson%MATH.PSU.EDU@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Stephen G Simpson <simpson%MATH.PSU.EDU@uga.cc.uga.edu>
Subject:      Re: pinyin input in ETEN
Comments: To: ccnet-l@uga.cc.uga.edu, zhongwen@nada.kth.se,
              A06G%DHBRRZ41.BITNET@psuvax1.cs.psu.edu
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

Regarding Urs Widmer's recent CCNET posting, concerning ETEN phrase
boxes.

I don't really understand ETEN phrase boxes, but it appears from Urs's
posting that the phrase boxes provide a capability similar to the tit
files in CXTERM.  In other words, the ETEN user can create a phrase
box file which defines a customized keyboard input method.

If this is correct, then it should be possible to use the

       PINYIN+TONE ==> ETEN BIG-5

input method, which already exists for CXTERM, and adapt it to ETEN.
This input method is already known to be completely compatible with
ETEN.  In fact, Yongguang Zhang and I developed this input method by a
phonetic conversion process starting with ETPHON.TAB in ETEN 2.03.
So, probably, all that needs to be done is to convert PY-b5.tit to the
ETEN phrase box file format.

I wish somebody who understands the phrase box format would attempt
this conversion and post the result here on CCNET, or make it
available for anonymous ftp.

If anyone would like to attempt this, the file that you need is
PY-b5.tit, available for anonymous ftp at cs.purdue.edu, directory
/pub/ygz/cxterm-dictionary.

I would also like to encourage Urs Widmer to make his own pinyin
phrase boxes available for anonymous ftp.  Urs, Bitt' sch\"on!


Stephen G. Simpson               simpson@math.psu.edu
Department of Mathematics        +1 814 863-0775
Pennsylvania State University    home: +1 814 238-2274
University Park, PA 16802

From CCNET-L@uga.cc.uga.edu Tue Nov 12 19:36:15 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA09081; Tue, 12 Nov 91 19:35:57 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 3786; Tue, 12 Nov 91 19:34:54 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 0578;
 Tue, 12 Nov 91 19:34:52 EST
Date:         Wed, 13 Nov 1991 11:31:30 +1100
Reply-To: dak415%COOMBS.ANU.EDU.AU@uga.cc.uga.edu
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: dak415%COOMBS.ANU.EDU.AU@uga.cc.uga.edu
Subject:      Re:  Cihui Chinese program (2)
Comments: To: ccnet-l%uga.BITNET@munnari.oz.au
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

>>From aaron@ahkcus.org Wed Nov 13 02:22:10 1991
>Date: Tue, 12 Nov 91 09:21:41 CST
>From: aaron@ahkcus.org (Aaron Y.T. Cheung)
>To: dak415@coombs.anu.edu.au
>Subject: Re:  Cihui Chinese program (2)
>Status: R
>
>HI David, sorry for the belated response regarding the use of
>cihui on the Mac.  However, as I myself am not a Mac user with
>Chinese softwares, and particularly regarding cihui, guess I can't
>be of useful assistance in this area.  However, you might be able
>to get some information by mailing to the folks at
>"ccnet@uga.bitnet", where most Chinese information procecssing
>users live.  Perhaps you might've known this already...
>Again, apologize for replying this late.  Best regards, /aaron.
>(operator@ahkcus.org)
>
>
Dear ccnet experts, I did not know about you. The following was my original
SOS to operator@ahkcus.org:

<<Further to my earlier email: I solved the ftp problem and decoded the
program using Stuffit Classic (not Binhex 4.0).

Now a new problem: It successfully created the *Cihui* cdev for simplified
system. But running it as an input method has bugs:

1. it does not show up in the control panel with other input methods.

It can be selected from the "Zhongwen Xitong" control panel submenu, and
can be made the startup method. However

2. When operating, the array of possible compounds is not shown in Chinese,
but in pairs of ASCII characters. It also has a tendency to bomb
completely.

Can anyone help? >>


Best Regards,
        	       	       	       	       	David Kelly
        	       	       	       	       	Contemporary China Centre
        	       	       	       	       	Australian National University

From CCNET-L@uga.cc.uga.edu Mon Nov 18 10:48:49 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA22281; Mon, 18 Nov 91 10:48:38 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 5915; Mon, 18 Nov 91 10:26:31 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 1111;
 Mon, 18 Nov 91 10:19:20 EST
Date:         Mon, 18 Nov 1991 09:55:45 EST
Reply-To: Yolin H u a n g <yih%BRL.MIL@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yolin H u a n g <yih%BRL.MIL@uga.cc.uga.edu>
Subject:      [Carlos McEvilly:  Re: Pinyin input in ETen]
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

----- Forwarded message # 1:

Date:         Sun, 17 Nov 91 10:01:08
From:         Carlos McEvilly <MOEG013%TWNMOE10.bitnet@vtvm2.cc.vt.edu>
To:           Yolin H u a n g <yih@BRL.MIL>
In-Reply-To:  Your message of Sat, 16 Nov 91 12:45:27 EST

This file is PINYIN.ZIP in xxencoded form.  Thanks to Yolin Huang
(yih@brl.mil) for all his help in getting it posted on the net.
The file works with the ETen Chinese system on MS-DOS machines to
provide a pinyin input method for Chinese text entry.  There are
instructions inside the .ZIP file.

- Carlos  moeg013@twnmoe10.bitnet
------------------------- Start of *.XXE --------------------
begin 644 pinyin.zip
...
end

----- End of forwarded messages
******************************************************************

All,
	I just xxdecode-d and ftp-ed this file to:
	ahkcus.org [192.55.187.25], and
	crl.nmsu.edu [128.123.1.14] in their incoming/ directory.

	Please direct your opinions to him (Carlos McEvilly
moeg013@twnmoe10.bitnet), or to the CCNET-L.


Yolin Huang.

From CCNET-L@uga.cc.uga.edu Thu Nov  7 10:41:12 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA22354; Thu, 7 Nov 91 10:40:55 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 5200; Thu, 07 Nov 91 10:39:34 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 7104;
 Thu, 07 Nov 91 10:37:53 EST
Date:         Thu, 7 Nov 1991 10:31:51 EST
Reply-To: Ned Walsh <walsh%CUNIXA.CC.COLUMBIA.EDU@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ned Walsh <walsh%CUNIXA.CC.COLUMBIA.EDU@uga.cc.uga.edu>
Subject:      Radical/stroke input file
Comments: To: CCNET-L@uga.cc.uga.edu
Comments: cc: harbsmei@hedda.uio.no
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

Urs,

It would be great if you'd post the 700-char radical/stroke file.  It
would go a long way toward solving the sorting problem I mentioned a
few weeks ago--i.e. how to get Big5 or GuoBiao files to sort in some
useful order.  That would leave only 5000 characters to assign
radical/stroke numbers.

Has anyone already come up with a solution to this problem?  Christoph,
how did you sort those concordances you were working on?

/Ned

From CCNET-L@uga.cc.uga.edu Tue Nov 19 12:18:06 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA02505; Tue, 19 Nov 91 12:17:52 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 7942; Tue, 19 Nov 91 12:17:02 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 1213;
 Tue, 19 Nov 91 12:16:08 EST
Date:         Tue, 19 Nov 1991 11:13:05 EST
Reply-To: Kam Mok <inmet!kam%UUNET.UU.NET@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Kam Mok <inmet!kam%UUNET.UU.NET@uga.cc.uga.edu>
Subject:      Re: pinyin in Eten
Comments: To: uga.cc.uga.edu!CCNET-L@uunet.uu.net
Comments: cc: kam@uunet.UU.NET, moeg013@twnmoe10.BITNET
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

Thanks a lot to Carlos McEvilly for posting the pinyin.box to a ftp site.
I tried that out and it worked like a charm. Now we can have pinyin method
in ETen.  I guess you would also volunteer to keep it updated if we
report to you any correction or modification. By the way who created this
phrase.box file originally ? A note about it in the readme.pin would
be appreciated.
Thanks again for the good work !

kam Mok

From CCNET-L@uga.cc.uga.edu Sat Nov 23 22:46:55 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA18316; Sat, 23 Nov 91 22:46:43 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 4247; Sat, 23 Nov 91 22:45:48 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 8571;
 Sat, 23 Nov 91 22:45:47 EST
Date:         Sat, 23 Nov 1991 22:44:12 -0500
Reply-To: Stephen G Simpson <simpson%MATH.PSU.EDU@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Stephen G Simpson <simpson%MATH.PSU.EDU@uga.cc.uga.edu>
Subject:      phrase input in emacs
Comments: To: Yu Shen <shen@iro.umontreal.ca>, ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
In-Reply-To:  <9111232328.AA00709@leibniz.math.psu.edu>
Status: RO

Dear Yu Shen,

I was very happy to see your posting with zw.el, and I want to
encourage you to develop this.  My biggest hope for cxterm.el was that
people would learn to use the powerful capabilities of Emacs for
solving the difficult problem of efficient Chinese input.  You are
taking a very big step in this direction.

One small comment on a side issue.  You wrote:

 >  And the transition from ascii
 > input mode to Chinese input is not smooth, beacuse the ESC character
 > has meanings both to emacs and cxterm.

I suppose this is because of the line

   ~Meta <KeyPress> Escape:  insert() switch-HZ-mode(ASCII)

in the cxterm*VT100.Translations: #override\ part of CXterm.ad (or
wherever cxterm is configured).  This line is meant only as a
convenience for users of celvis.  Users of cemacs (i.e. cxterm.el)
should delete this line from the cxterm setup.

Best wishes,

Steve Simpson

Stephen G. Simpson               simpson@math.psu.edu
Department of Mathematics        +1 814 863-0775
Pennsylvania State University    home: +1 814 238-2274
University Park, PA 16802

From CCNET-L@uga.cc.uga.edu Sun Nov 24 11:16:37 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA20623; Sun, 24 Nov 91 11:16:22 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 4986; Sun, 24 Nov 91 11:16:03 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 5319;
 Sun, 24 Nov 91 11:16:02 EST
Date:         Sun, 24 Nov 1991 11:14:29 -0500
Reply-To: Stephen G Simpson <simpson%MATH.PSU.EDU@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Stephen G Simpson <simpson%MATH.PSU.EDU@uga.cc.uga.edu>
Subject:      using abbrev mode to enter Chinese phrases
Comments: To: ccnet-l@uga.cc.uga.edu, shen@iro.umontreal.ca,
              valuck%ucdavis.BITNET@psuvax1.cs.psu.edu
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

There have been some recent postings on how to efficiently enter
Chinese phrases (Ci Hui).  Some of the solutions that people have used
in the past are: Yi Tian box files, and the Chtext self-expanding
phrase dictionary.  I have not used Tian Ma but my impression is that
it has sophisticated methods for entering phrases.

One way to enter Chinese phrases is to use abbrev mode in GNU Emacs
(or Cemacs, running under cxterm).  GNU Emacs abbrev files can be
saved and edited, so this provides a capability something like Yi Tian
box files.  Yu Shen in his zw-mode posting was using GNU Emacs abbrev
mode to create something like a self-expanding phrase dictionary.

A problem facing Yu Shen is that he was trying to use full pinyin
spellings for abbreviations, and these abbreviations are too hard to
type and remember.

To address Yu Shen's problem, perhaps it is practical to follow Carlos
McEvilly's suggestion.  Carlos was talking about Yi Tian box files,
but the same applies to GNU Emacs abbrev files.  Namely, just use the
first letter in the pinyin spelling of each word of the phrase.  For
instance a 4-character phrase like "Chuan1 Shan1 Guo4 Hai3" (pierce
mountain cross sea) could be abbreviated "csgh".  This kind of
abbreviation should be very easy to remember and to type.

Stephen G. Simpson               simpson@math.psu.edu
Department of Mathematics        +1 814 863-0775
Pennsylvania State University    home: +1 814 238-2274
University Park, PA 16802

From CCNET-L@uga.cc.uga.edu Sun Dec 15 23:11:08 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA21867; Sun, 15 Dec 91 23:10:52 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 4416; Sun, 15 Dec 91 23:10:09 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 9233;
 Sun, 15 Dec 91 23:10:08 EST
Date:         Sun, 15 Dec 1991 23:04:33 EST
Reply-To: Yolin H u a n g <yih%BRL.MIL@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yolin H u a n g <yih%BRL.MIL@uga.cc.uga.edu>
Subject:      simplification; PinYin; GB
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

Re:------------------------------------------------------------------
: "Minghui Yao, Dept of Physics, ASU"
: <YAOM%ASUCP1.LA.ASU.EDU@uga.cc.uga.edu> 25 Nov 1991
: ... The evolution of Chinese Characters is a nature simplifying process,
: from complex to simplicity. If a language is unchanged, it is dead.  ...
--------------------------------------------------------------------------

among these dead scripts?: 1) the modern generations of the European
North Sea and Baltic Sea countries can not understand the old Runic
oracle script (Hellenistic/Cyrillic remnants?) after it was replaced by
the Roman script.  2) the modern generations of the Middle East can not
understand the old Egyptian, Philistian, Phoenician, Aramaic, Syriac,
Akkadian, Babylonian, Chaldean, Sumerian, Old Persian, some Turkic, ...
etc scripts; after those scripts evolved and merged into the modern
Arabic, the new lingua franca.  3) new generations of Tannu Tuva (Dr
Feynman's hobby), the Buryat and the Rep Mongol have difficulties to
understand the traditional Mongolian texts, after the Cyrillic alphabet
was enforced.  4) new generations of VietNam can no longer understand the
bulk of historical archives in Han script, after the French style
spelling was enforced. Many people don't even know of their existence
(probably stored mostly in France as archaeological and anthropological
novelty or vanity?)  5) new generations of Korea have difficulties to
understand the bulk of historical archives in Han script (many highly
achieved scholarship), after the phonetic system set in by decree, so as
to refrain from the Han script implied Chinese chauvinism.

Today, after centuries of evolution, English and French spellings have
lost some of their original phonetic purpose, and have become more or
less ideographical ('ped xing'.)  Maybe one day, after Han Solo and Luke
Skywalker graduated into nirvana, the ASCII script would merge with other
ideographical into an yet unknown new script.:)  But right now, both
groups are cautious to take any measure so as not to deliver the Oxford
and Larousse dictionaries into reincarnation.

Contrary to the assertion, the same Han script has been in use, and
stabilized for almost two milleniums without much change.  Different
fancy writing styles have been artfully painted.  But, there was no
significant gradual simplification of this script in the past as
asserted, other than in the hand-writing routine for convenience.  The
Beijing simplification regulation in the 60's, fueled along the craze of
the Cultural Revolution, can probably better be called an 'experiment'
among other 'social experiments'.

Furthermore, the simplification regulation was not a Beijing innovation
either.  The similar regulation in Japan (same idea, not copyrighted)
predated the Beijing regulation by a couple of years, under the post-war
General McArthur Administration.  At that time, the simplification rule
did have some edge over the slower traditional style, in hand-writing
(saved some brain-hemorrhage cases on desk scribe work.)  But with the
advance of PC, this edge dimishes.  For the electronic text processing,
it poses probably more problem than advantage.

Three decades passed.  It is now probably high time to sit down, and
ponder about its worth, run a decent re-evaluation, and learn from the
experience.  I think, for the pupils learning the traditional style
first, and then (say after the elementary school build-up stage) allowed
to use simplified shorthand (as supplement) in taking notes, writing
reports or exams would be better off as a whole; than learning the
simplified style first, and later randomly picking up the traditional
style in the real life (in unscathed family memorabilia, libraries,
museums, parks, exhibitions, memorials, movies, karaoke clubs, etc.)
Doesn't have to go back to the old Confucian-style rigid perfectionism
(in hand-writing tedium), if you will.  In English dominions, Gregg's
shorthand system is widely used to speed up the hand-writing.  But it is
usually not a part of the regular education, as I understand.  And many
people, even secretaries, don't write with calligraphical excellence,
some just stay at the grade school level I guess.  When typed up, not
even the boss would notice.

	-------------------- <o> --------------------

The 'PinYin' system (Beijing, 1950's) is better than 'bopomofo'
(Nanking/Nanjing, 1910's) in that it is readily adaptable to the QWERTY
keyboard everywhere on Earth/Space.  (And also better than the Wade-Giles
system [Cambridge UK, 1850's/1910's], because of a better phonetic
representation, and the abstinence from the excessive apostrophe, ...
among other things.)  But it is not without its pitfalls.  Since 'PinYin'
was romanized from 'bopomofo', it retains some syllabic nature as that in
the bopomofo, and kana.  Despite of the alphabetical appearance, it
doesn't have all the phonetic resolution of an ideal alphabetical
system.  Such as in 'x', 's', and 'q', 'c', (each has a corresponding
symbol in bopomofo), which are usually quite a hurdle to the
uninitiated.  I suppose, it would make better sense for 'x', and 's' to
use the same openning consonant followed by different vowels (for
instance, one sharp (sii) one muted (si)).  Same for the 'q' and 'c' (for
instance, one sharp (cii) one muted (ci); suppose there is probably no
better choice than 'c' other than 'ts').  (And in combinations, some of
the 'sii' can be reduced to 'si' without causing confusions; same for
'tsii', and 'tsi'.  Isn't SiaoPing as clear as SiiaoPing, even though not
as short.)  (QuYuan isn't really that royal as Queen.  Isn't 'Tsisii'
easier for the uninitiated than 'Cixi'?  Not Tootsie; 'Tsaotsao' easier
than 'Caocao'?  No cacao today; 'Siishi' better than 'Xishi'?  Not
exactly Sushi bar. :-) ) In so doing, the 'c', 'q', and 'x' could be
replaced by the less offending 'ts', and 's'; and easier acceptable to
the commoners, ignorant mass, and innocents abroad.

I think the Beijing tongue of Han pronunciation (Portugese mistakingly
called Mandarim) probably have seven vowels ( (x)i, (s)i, e, a, o, u, and
umlaut u), not counting diphtongs.)  For the umlaut u sound, some write
'u" (Umlaut)', some people use 'uu'. ('uu' probably better.  Probably
'ui', 'u:', 'u^' can also be considered.  But it seems there is no easier
solution with the popular QWERTY keyboard.  The current use of 'ui' would
probably be more accurately represented by 'uei'.)

Furthermore, the 'an' in the following (mostly diphtongs)
(group 1:)
bian, dian, jian, juan, lian, mian, nian, pian, qian, quan, tian, xian,
xuan, yan, yuan
\
is pronounced differently from the 'an' in the following
(group 2:)
chuan, chuang, cuan, duan, guan, guang, huan, huang, jiang, kan, kang,
kuan, kuang, liang, luan, nan, nang, niang, nuan, pan, pang, qiang, ruan,
san, sang, shan, shang, tuan, yang, ShaanXi, XiAn
\
but closer to the 'en' in the following
(group 3:)
feng, gen, geng, hen, heng, ken, keng, leng, men, meng, nen, neng, pen,
peng, ren, reng, sen, seng

Does the emperor have a new dress on?  This pattern again betrays its
root in the syllabic bopomofo system, where 'an' has a (phonetic)
symbol.  Probably the 'an' in the group 1 could be modified, and further
evolve to 'en' as in group 3.  (The rest of PinYin is probably as good as
new?  What do you think?)  After that, the 'an' will be consistent
throught out, and easier to teach to kids.  And the PinYin system will be
closer to a full-fledged alphabetical system.  One flintstone two birds.

Responsible administrators (of all ideological pre-destinations) probably
should please look closer into this, and try out in some language labs,
experimental schools, or better educated townships, (HongKong, TaiWan?)
for a few years.  A short pain (in making improvements) would be better
than chronic pain (in accomodating the flaws for the next ?? years?)

There is another practice in mainland I couldn't understand: many
merchandise and signs use the 'PinYin' stand-alone, as-is; as if people
could understand them.  I suspect most tourists would first have to
struggle for a good pronunciation, let alone its meaning.  The native
speakers probably won't even bother to look at them, simply don't give a
damn.  I think 'PinYin' is better used only as a pronunciation aid:
small print next to the Han script, like with 'kana'.  But after all,
'PinYin' is probably lesser a problem than the simplication regulation.

	-------------------- <o> --------------------

ygz%CS.PURDUE.EDU@uga.cc.uga.edu (27 Nov 1991) (cxterm, great job) gave
an impressive report about the status quo of the GB scheme, with
revelatory discussions about problems with GL, GR planes in Big5.  The
problem described with the Big5 is probably related to the present 7-bit
environment, which was originally designed only to handle the ASCII
characters.  (The Europeans complain about the ASCII as well, even though
got more out of it.)  But, according to Jianxun Ding (jding@CS.TAMU.EDU,
14 Feb 91) GB 0-3 also use the high bit; so the full-blown GB scheme
(like the one in LianXian Han Ka system used by AG02005@MVS.UNI-C.DK, 28
Nov 1991) might have the same problem?

Now, with the maturing in the technology (and in the humanity, such as
the New Age), and the necessity to handle texts other than Latin
derivatives, the hardware and software developers are expanding their
horizon.  I guess that the OS is moving toward 8-bit and beyond.  So, the
mission impossible today with the old 7-bit ISO 2022, will probably become
possible with the new ISO standards.  It was said that in a Confucian
style top-down bureaucracy, obedience|revolution is usually the norm;
whereas in democracy the rules will respond to the need of the
constituents, ... in theory anyway.  The expanded bandwidth can also
treat other problems in communication, not just Han scripts.

In the current GB 0-5 arrangement, simplified and traditional characters
seem to have different code numbers.  So, the programmers have to build a
conversion procedure into the software, and exercise contextual analysis
to sort out the 1:n mapping pain with a sufficient phrase database
(Dorothy Day DAY@IUBACS.BITNET, 24 Nov 1991; AG02005@MVS.UNI-C.DK, 28 Nov
1991.)  The overhead is expensive.  I think it is probably not a good
arrangement.  Why should a proper name, printed differently in mainland,
ChoSun, HanKuk, HongKong, JaPan, TaiWan, and elsewhere, have different
pointers?  Don't they in fact mean the same, whether it is in Time-Roman,
Helvetica, or Gothic|Fraktur?  Too complicated.  It is here, I suppose,
that some decent simplification will be helpful.  (Improved efficiency
can improve the economy, i.e. money.)  Therefore, the current GB is
probably not much better than Big5, and is probably a flaky foundation to
build upon for long-term textual database effort.  (Notwithstanding,
cxterm, byx, cwp, njstar, cedit ...  are respectful endeavors.)

IMHO, it would make much better sense to start off with a larger, more
complete generic character set ('logic OR' of all styles; no matter how
written), with considerations to regional sets (JaPan, HongKong etc.),
and then lay out the simplified character set over it (with hundred some
holes), statistically sorted into subsets, and make sure to use the same
code number for the same character for all styles, (in PinYin first,
radical second, stroke count third sequence.)  And then use a background
selector to select fonts (like in DOS Word Perfect, or other Window
softwares.)  In Mcintosh I remember there is even a 'blackmail ASCII
font' for the reckless types.

Then, the user can write many different styles of Han script in the same
text.  For instance, the reporters can write a text, and use the mouse to
mark those personal and locality names into Kanji style in five seconds.
The users (cadre, non bourgeois?) have the freedom to choose the
simplified style (as for 1:n simplified characters, there will be fewer
listed for PinYin choice.)  For some (learned dons, nouveau riche?)
aspiring to stay ahead of the vulgarity, they should have the freedom,
without coercion or harassment, to use other styles.  (The 'blackmail
font' is not suggested.:)

And it can readily be expanded to cover classical script styles, such as
the JiaGuWen (Shell|Bone Oracle script?), ZongDingWen (Bronze Vessel
script?), XiaoZuan, DaZuan, LiShu, etc., and hundreds other brush writing
styles, with minimal effort.  (probably with a lot of holes in the
arrays/subsets as well.)  And it can be readily expanded to handle
different bitmap fonts and Postscript fonts in the future.  Other
languages as well.

With the present demographical distribution, the (Imperium) Han culture
probably better be considered a common heritage to CJK countries plus
Japanese produce growers, Korean grocers, and Chinese carry-outs here and
there, plus Sino philos, enthusiasts.  Anyhow, it's a minority culture on
Earth, with a size less than the Latin derivatives.  As long as the
stability and the self-esteem are not compromised, less chauvinism/
feudalism is probably more constructive.  (Actually, 'Chinese' is more
geographical than ethnic.)

Furthermore, for a mutli-cultural society, supposedly the GB space should
be used, IMHO, to cover other 50+ scripts under the mainland
administration, such as Korean, Manzu, Mongol, Uigur, Tibetan, Yi,
Zhuang, Dai, ... etc.  (some extra-territorial negotiations needed) So
one day, major cultural archives can all be digitized, cross referenced,
and index searched, not only over the time (historical correlations), but
over the space (inter-communal correlations) as well; and start to
demonstrate the usefulness, in solving mystries (what is Li Bai's
original Turkic name?), clearing myths (Shangrila where?), finding
origins of words or expressions (many of foggy regional, peripheral, or
external origin), studying the evolution and migration of dialect groups,
...  etc.  (I wonder how Bernhard Karlgren did this without PC; kicked
off by Sven Hedin?)

The electronic Han text processing is still immature and evolving.  But,
the June Fourth has passed only two years.  I am not sure whether the
present climate is ready for this quibble yet.  However, long hesitation
is probably not good either.  Some people are not as patient.  I like to
see discussions (from linguistic perspective), so that it can evolve
further for the better.

Probably I blow a different tune in this open forum.  But, no blasphemy
or counter-revolution was intended.  Only like to see that the mainland
language policy, as a whole, would move in a better direction, and be more
efficient for every endusers (kind of the copylefted GNU spirit); for
everybody has to survive the ISO.  So then, there is a better foundation
toward the next century.  A good foundation is very important, as I was
told.  It is not an easy job; but probably not much harder than moving
the American measure toward decimal, I guess.  After five hundred years,
Cristoforo Colombo has barely reached Cathay.  I appreciate the progress
the ASCII has brought to the Han text processing.

Happy Holidays.  Now, your turn, please.
(step down the soap box, go get some drink)

============================================================================
Mao: (straight, right arm half raised, stern) Rebel with Cause!
James Dean:             ..... (casual, smile) ..... hn? ..........
Lin Yutang: ... Chinese ... is R4D1H3S3.
Chewbacca: (hoarse) ... R2D2 !! ... hhhrrrrr ... hhhrrr .....
(by the distant Valhalla commune,
a Maitreya/MiLeFo type:) ..... HoHoHo! ..... HoHoHo! .....



Appendix:
-----------------------------------------------------------------------------
From: henstock@ec.ecn.purdue.edu (Peter V Henstock)
Newsgroups: sci.lang.japan
Subject: JIS <-> Guo Bao Cross referencing
Date: 8 Dec 91 19:48:32 GMT

I have been following the threads of Japanese kanji and Chinese hanzi
comparisons.  Does anyone know of a list which cross references the
kanji and hanzi?  There are several standards of codes for both the
kanji and hanzi and many of the characters are used in both languages
though some simplifications have occurred in both languages.  I realize
that there may not be a 1:1 correspondence for all characters, but has
such a correspondence been made?

Thanks in advance,

--Peter
.
-----------------------------------------------------------------------------

From sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!qt.cs.utexas.edu!cs.utexas.edu!sun-barr!ccut!wnoc-tyo-news!sh.wide!astemgw!icspub!uopgw!okamoto Sat Jan 25 20:05:50 EST 1992
Article: 503 of comp.sys.sun.misc
Xref: sol.ctr.columbia.edu comp.unix.sysv386:17225 comp.sys.sun.misc:503
Path: sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!qt.cs.utexas.edu!cs.utexas.edu!sun-barr!ccut!wnoc-tyo-news!sh.wide!astemgw!icspub!uopgw!okamoto
From: okamoto@earth.chigaku.souka.osakafu-u.ac.jp (Kenji Okamoto)
Newsgroups: comp.unix.sysv386,comp.sys.sun.misc
Subject: Re: more about Solaris
Summary: News around Japanese version of Solaris 2.0 appeared also
	 in some Japanese econimic newspapers such as Nikkei.
Message-ID: <1132@earth.chigaku.souka.osakafu-u.ac.jp>
Date: 24 Jan 92 02:01:39 GMT
References: <5@com001.info.com> <1992Jan12.203712.4122@pegasus.com> <468@tokyo07.info.com>
Followup-To: comp.unix.sysv386
Organization: Univ. of Osaka Prefecture, Dept of Earth Science, Sakai, Japan
Lines: 36
Status: RO

These news around Solaris 2.0 also appeared in one of the general
(particularly for economists) newspapers at the end of last year. 
I think many peoples having interests in computers (in Japan) were 
much impacted by this news. 

I can add some information on jimmy's article:

In article <468@tokyo07.info.com>, jimmy@tokyo07.info.com (Jim Gottlieb) writes:

>    Nihon Sun Microsystems K.K. and Just System Co., Ltd. have
> agreed to jointly develop a version of "ATOK," a Japanese kana-to-
> kanji (phonetic-to-ideographic character) conversion software
> module originally developed by Just System.  The module will be
> used in the "Solaris" operating system developed by Sun.

The "ATOK" is the most popular FEP(front end processor for Japanese input) in
the personal computer market, which we can imagine by an analogy to UNIX's 
Wnn jserver + kinput.

>    NEC Corp. and Sun Microsystems Inc. of the US plan to
> make an agreement whereby NEC will be able to use the "Solaris"
> system software on its "PC-9800" series PCs.
>    Sun will develop a Japanese language version of Solaris by the
> middle of next year, and NEC will port the software to the PC-9800
> series PCs.

This PC-9800 series PC governs the Japanese PC market.   Something like
IBM/PC-AT in the world.

Kenji
------------------------------------------------------------------------------
Kenji Okamoto     |             okamoto@earth.chigaku.souka.osakafu-u.ac.jp
 ______  __ |__   |  Universiy of Osaka Prefecture, Dept. of Earth Science,
|__\/__|   /|\    |  Mozu Umemachi 4-804, Sakai, Osaka 591, Japan
||_|_| |  / + \   |  tel(voice) +81-722-52-1161 (ext)2736 (FAX)+81-722-55-2981
------------------------------------------------------------------------------


From CCNET-L@uga.cc.uga.edu Mon Feb 10 10:39:32 1992
Received: from ugav2.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA09115; Mon, 10 Feb 92 10:39:28 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 0422; Mon, 10 Feb 92 10:38:30 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 9752;
 Mon, 10 Feb 92 10:38:28 EST
Date:         Mon, 10 Feb 1992 16:36:27 +0100
Reply-To: "Y. PAN" <pan@VXUWYZ.CERN.CH>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "Y. PAN" <pan@VXUWYZ.CERN.CH>
Subject:      Input methods
Comments: To: ccnet-L%UGA.bitnet@MINT.decnet.cern.ch
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Hi, Netters,
      Could anybody tell me where to find the dictionaries for various Chinese
input methods. They can be either in .tit format or any machine readable
formats. I hope these dictionaries are in public domain, otherwise please
tell me how to get the proper authorizations. They are not going to be used
for any commerical purpose.
Thanks.
Pan.

From CCNET-L@uga.cc.uga.edu Thu Feb 20 03:19:16 1992
Received: from ugav2.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA05365; Thu, 20 Feb 92 03:19:06 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 4907; Thu, 20 Feb 92 03:18:22 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 9494;
 Thu, 20 Feb 92 03:18:20 EST
Date:         Thu, 20 Feb 1992 02:16:10 -0600
Reply-To: Thomas Chin <thomasc@WRI.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Thomas Chin <thomasc@WRI.COM>
Subject:      Cantonese PinYin table?
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I am looking for a Cantonese Pinyin table. (or a program that converts
between Cantonese Pinyin and big5/GB)  I think I came across one before at
a ftp site somewhere in CA.  Can somebody advice ?

Thanks in advance.

-Tom

From CCNET-L@uga.cc.uga.edu Fri Feb 21 05:22:59 1992
Received: from ugav2.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA14892; Fri, 21 Feb 92 05:22:54 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 1791; Fri, 21 Feb 92 05:22:11 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 2529;
 Fri, 21 Feb 92 05:22:10 EST
Date:         Fri, 21 Feb 1992 10:18:20 GMT
Reply-To: wu@CANON.CO.UK
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: wu@CANON.CO.UK
Subject:      Chinese OCR
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Hi,

I am planning to start some work on Chinese OCR for PRINTED
materials. I would be grateful if anyone could send me information
about the technology, products etc.

There is a directory called ocr.fonts in blackbox.hacc.washington.edu
in which I found ocra.mf and similar things. Anyone knows what
they are?

Thanks in advance.

S.F. Wu  wu@canon.co.uk

From CCNET-L@uga.cc.uga.edu Fri Feb 21 11:11:02 1992
Received: from ugav2.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA16250; Fri, 21 Feb 92 11:10:37 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 3648; Fri, 21 Feb 92 11:08:57 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 0168;
 Fri, 21 Feb 92 11:04:57 EST
Date:         Fri, 21 Feb 1992 10:53:02 EST
Reply-To: John McRae <JRM%CORNELLA.BITNET@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: John McRae <JRM%CORNELLA.BITNET@uga.cc.uga.edu>
Subject:      Chinese OCR
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I have a colleague in Kyoto who uses a Japanese NEC 9801 computer and scanner
and software called Saijicho- to do OCR on Buddhist texts in the Japanese
edition of the Chinese canon.  He claims very high accuracy rates and good
efficiency, although the hardware/software combination is slow enough that he
trains operators the system and then has them "mind" two machines running
simultaneously.  He's very happy to distribute the character shape files that
he has generated, as well as all the texts.  Contact Urs App, International
Institute for Zen Buddhism, 8-1 Tsubonouchi-cho, Nishinokyo, Nakakyo-ku,
Kyoto 604, Japan.  You can try e-mail (D54682@JPNKUDPC.BITNET), but his link
is erratic and unreliable for reasons I don't understand; much better to mail
to the address just given or fax to (075) 811-9664.  Voice mail (if you're
into long distance) is (075) 811-5181x280.

One of the things I would like to see happen with this software is some kind
of reverse engineering so that it would work on conventional DOS machines,
probably by using the recently released J/DOS 5.0/V.  The company that makes
Saijicho- (that's a long "o" at the end) may be thinking of doing this them-
selves -- I have no idea at all.  But since the program seems to work pretty
well, it would be a shame for others to have to recreate the wheel again, as it
were.

P.S.  Another thing I don't know about this software (my ignorance is virtually
infinite!) is what format it reads files in.  That is, there should also be
some way to figure out how to use images scanned on other equipment, even if
one had to use the NEC 9801 to do the OCR conversion.



John R. McRae, Asst. Professor, Department of Asian Studies
Director, Cornell University Chinese-Japanese-Korean Computing
     and Database Facility (CJK-CDF), East Asia Program
389 Rockefeller Hall
Cornell University
Ithaca, NY  14853-2502
o:  607/255-1328; fax: 255-1686

From CCNET-L@uga.cc.uga.edu Mon Feb 24 15:07:46 1992
Received: from ugav2.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA07540; Mon, 24 Feb 92 15:07:37 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2192; Mon, 24 Feb 92 15:05:02 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 4065;
 Mon, 24 Feb 92 14:55:59 EST
Date:         Mon, 24 Feb 1992 14:50:51 -0500
Reply-To: Boon Kang <BOONKANG%BCSC02.BITNET@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Boon Kang <BOONKANG%BCSC02.BITNET@uga.cc.uga.edu>
Subject:      Re: [jbm:  Numbers in Ancient Chinese]
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

 *** Reply to note of 02/22/92 22:49

    I don't know who made such claim that "Holo and Hakka are closer
    to classical Chinese" but I thought many students of Chinese
    linguistics are aware that a lot of ancient chinese pronunciations
    and vocabularies have been well preserved in Hakka and Minnan (or heluo)
    dialects. Many scholars have been using Haaka and Minnan dialects as
    studying tools of ancient Chinese linguistics and chinese history.

 Regards,                          VM/PROFS: BOONKANG
 Boon Kang, IMS Support, HTS
 4000 Seymour Pl, W169BLUE         Phone:    604 389-3991
 Victoria, B.C. CANADA   V8X 4S8   Fax:      604 389-3916

From CCNET-L@uga.cc.uga.edu Fri Apr 10 14:24:11 1992
Received: from ugav2.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA21139; Fri, 10 Apr 92 14:23:55 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 4239; Fri, 10 Apr 92 14:16:39 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 0952;
 Fri, 10 Apr 92 13:11:06 EDT
Date:         Fri, 10 Apr 1992 13:08:59 EDT
Reply-To: adlerj@VAX001.KENYON.EDU
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: adlerj@VAX001.KENYON.EDU
Subject:      Chinese in WordPerfect
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

To put Chinese characters into WordPerfect documents, you can run
ETEN (I T'ien, or ET), which converts DOS to a Chinese system.
(It actually runs under DOS, I believe.)  Under this system you
can use WordPerfect, but only up to version 4.2, not 5.0 or 5.1,
unfortunately.  WordPerfect 4.2 works normally under ET, but
printing can be a problem.  The only printer I know that will
definitely print WP files containing ET characters is the
Panasonic KX-P1124 or KX-P1124i (a 24-pin dot matrix in the $300
range).  The trick is to tell ET that you are using an Epson LQ-
2500 (select printer P10 using the etctl command.)  There are
other ways to work around this limitation using other printers.
One is to convert files to ASCII, but then of course you lose the
WP codes for underlining, bold, footnotes, etc.

Concerning the recent message about the ET Pinyin add-on file
available via ftp from ahkcus.org:  If this doesn't work, a
commercial version is available for $20 from Eagon Systems, 10
Hearthstone Circle, Natick, Massachusetts 01760, USA.  They also
have an English manual for ET (simpler and shorter than the
Chinese manual) for $15.

Joseph Adler
Dept. of Religion
Kenyon College
Gambier, OH 43022
ADLERJ@VAX001.KENYON.EDU

From CCNET-L@uga.cc.uga.edu Thu Apr 16 07:04:30 1992
Received: from ugav2.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA29186; Thu, 16 Apr 92 07:04:26 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 0191; Thu, 16 Apr 92 07:03:27 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 7168;
 Thu, 16 Apr 92 07:03:25 EDT
Date:         Thu, 16 Apr 1992 19:08:52 WST
Reply-To: Huang Chih-Hsien <u7911013@CC.NCTU.EDU.TW>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Huang Chih-Hsien <u7911013@CC.NCTU.EDU.TW>
Subject:      Re: Question about differences among input methods
Comments: To: plau%SUNEE.WATERLOO.EDU@TWNMOE10.EDU.TW
Comments: cc: ccnet-l@uga.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <9204160705.AA09096@cc.nctu.edu.tw>; from "Patrick Lau" at Apr
              16, 92 2:52 am
Status: RO

Hi:

> 1. What is the difference between 2 and 4?

Zhuyin, "bopomofo" is widely used in Taiwan and similiar to PinYin
in mainland China.  In fact, Zhuyin is a popular Chinese input method
for "common" users, instead of expert users.  That's because to people
here, we are fimiliar with "bopomofo", although too many characters
with the same pronunciation leads to bad performance.

> 2. What is Simple?

Simple, Jian3 Yi1, is another input method available in Taiwan.  However,
in most cases, people seldom use it.  Jian-Yi = the_1st_TsangChi_input_code
+ the_last_TsangChi_input_code.  It is often applied when people can not
guess out the correct codes.

> 3. What is GB?

GB is not used in Taiwan.  I guess people from mainland China will
know much about it.

> 4. What is Big-5?

Big-5 is NOT a input method but a Chinese coding method.  It is widely
used in Taiwan and maybe also in Hong Kong.  Most Chinese systems
developed here do support Big-5.  However, it is a 8-bit coding method
which is not convenient for spreading out over the 7-bit computer
network.

Best Regards,

--
------------------------------------------------------------------------
|  H u a n g ,    C h i h -    H s i e n   | E.E.83, Sophomore         |
|------------------------------------------|---------------------------|
|    ::   :: .       :;      ,,,,,;.,,,,:. | Dept. of                  |
|  ''::'''::''' ,,,,,;;,,,,;,::.:., ', ,:  | Electronics Engineering   |
|    ;:...:;         ::      ::.,.:  ,':,  | National Chiao Tung Univ. |
|,,,,,,,,,,,,,;,     ::   .  ::.:...'   ':'| 1001 Ta Hsueh Rd.         |
|  :........;;,  '''''''''''  :.......::,  | Hsinchu, Taiwan, R.O.C.   |
|  ::   :   ::    :. :.       ::......::   |---------------------------|
|  ::''':'''::   :::  ;, ';.  ::......::   | E-mail:                   |
|  ;:...:...:;  :'::   ' : :; ::......::   |  u7911013@cc.nctu.edu.tw  |
| ..;''   '';, '' ':::::::'  ..;''   '';,  |  7911013@twnctu02.bitnet  |
------------------------------------------------------------------------

From sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!not-for-mail Tue Jul  7 16:44:43 EDT 1992
Article: 61463 of soc.culture.china
Path: sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!not-for-mail
From: xiong@shape.mps.ohio-state.edu (Chuyu Xiong)
Newsgroups: soc.culture.china
Subject: Chinese Computing
Date: 6 Jul 1992 15:16:59 -0400
Organization: Department of Mathematics, The Ohio State University
Lines: 21
Distribution: world
Message-ID: <13a67bINN428@shape.mps.ohio-state.edu>
NNTP-Posting-Host: shape.mps.ohio-state.edu
Status: RO

Several days ago, I posted some questions about Chinese computing.
There are not many responses for that post. However, I received some
emails. Here is one I would like to share with you:

---------------------------------------------------------------------------
Hi:
	In 1987, I saw a system with optical pen as input in Zhongguanchun,
Beijing.  It was on display there.  I played with it for about an hour. 
The main problem is speed as a PC/AT was used.  I think now this problem
has been solved.  
	I also heard that a scaner system for inputting PRINTED Chinese
have been developed 2 or 3 years ago.
	As to the computer industry, I personally think (or my impression is)
mainland is more advanced on the system aspect and Taiwan on manufacturing.
	Please post this on SCC net if possible.  I don't know how to post
on the net.  Thanks
---------------------------------------------------------------------------

C. Xiong
Math. Dept. Ohio State University.



From @NYX.UNI-KONSTANZ.DE:E2FHVM19@DKNKURZ1.BITNET Tue Jul 14 16:24:55 1992
Received: from nyx.uni-konstanz.de by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA10298; Tue, 14 Jul 92 16:24:41 -0400
Received: from nyx.uni-konstanz.de by NYX.UNI-KONSTANZ.DE (IBM VM SMTP V2R2)
   with BSMTP id 0789; Tue, 14 Jul 92 22:23:12 MEZ
Received: from DKNKURZ1.BITNET (E2FHVM19) by nyx.uni-konstanz.de (Mailer R2.08
 R208004) with BSMTP id 9051; Tue, 14 Jul 92 22:23:12 MEZ
Date:         Tue, 14 Jul 92 22:13:57 MEZ
From: Wang Wei <E2FHVM19@DKNKURZ1.BITNET>
Subject:      Chinese System for PC
To: Yuan Jiang <yjj@ctr.columbia.edu>
Message-Id:   <920714.221357.MEZ.E2FHVM19@DKNKURZ1.BITNET>
X-Acknowledge-To: <E2FHVM19@DKNKURZ1.BITNET>
Status: RO

Hello,

I've long been a listener of the group CCNET-L and I've just read your
message about that Mr. Yan YongXin was trying to write a Chinese system
for PC all from scratch. I think I might be of some help. I have been
trying something on my own for some time. Till now I've done a Chinese
phrase input module for Microsoft Windows (3.0/3.1). It may still has
bugs. But it has already the functionality of PJS (I'm not sure if you
know it in the US. PJS is a China made system based on PinYin input. It
is quite comfortable to use but dies all the time) or looks somewhat
like the CiHui for MAC (I'm not sure of it myself). I haven't showed it
to anyone yet. But please let know if it is of any interest to you.


Regards.


Wang Way      Bitnet:   E2FHVM19@dknkurz1.bitnet
              Internet: E2FHVM19@nyx.uni-konstanz.de

From sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!news.acns.nwu.edu!casbah.acns.nwu.edu!tan Sun Aug  2 00:14:30 EDT 1992
Article: 374 of alt.chinese.text
Newsgroups: alt.chinese.text
Path: sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!news.acns.nwu.edu!casbah.acns.nwu.edu!tan
From: tan@casbah.acns.nwu.edu (Shufeng Tan)
Subject: CWP-style input method for cxterm
Message-ID: <1992Jul30.232602.16599@news.acns.nwu.edu>
Followup-To: tan@casbah.acns.nwu.edu
Sender: ~{L4Jw7e~}
Organization: Northwestern University, Evanston Illinois.
References: <Bs86BB.6ux@ecf.toronto.edu>
Date: Thu, 30 Jul 1992 23:26:02 GMT
Lines: 51
Status: O

~{::SoF4RtJG4s6`J}HKSC@4JdHkVPND5D7=7(!#NR>u5CSCJ}WV<|@4JdHkRt5w:MQ!~}
~{TqM,RtWVJG8v:\1?5D0l7(!#0=P!F=5D~}CWP~{KySC5DJdHk7=7(>M1H=O4OCwR;P)#:~}
~{MjH+2;SCJ}WV<|!#RuF=!"QtF=<0IOIy7V1pSC~}hxv~{1mJ>#,H%Iy<0GaIy2;1j#,M,~}
~{RtWVSCLX1pLtQ!5DIyD8Gx7V!#NR8y>]Ub8vT-TrP^8D~}cxterm~{5DF4RtJdHk1m#,4r~}
~{WVKY6HSPKyLa8_!#2;V*VnN;JG7q8PPKH$!#Hg9{SP5D;0#,GkV(R;Iy!#~}

~{@OL4~}

~{8=JdHk1mM7HgOB#:~}

# $Id: TSF.tit,v 1.0 1992/7/30 06:28:34 tsf Exp $
# HANZI input table for cxterm
# To be used by cxterm, convert me to .cit format first
# .cit version 1
ENCODE:	GB
MULTICHOICE:	YES
PROMPT:	~{::WVJdHk!KIq1JF4Rt!K~}
#
COMMENT	~{Iq1JF4Rt7=08~}
COMMENT ~{WwU_#:L4Jw7e~}
COMMENT ~{P!P4S"NDWVD84z1mF4Rt7{:E#,~}umlaut u ~{SC~} yu ~{1mJ>!#~}
COMMENT ~{Rt5wR2SCWVD81mJ>#,~}hxv~{7V1p4z1mRuF=!"QtF=<0IOIy#,H%Iy<0GaIy2;1j!#~}
# define keys
VALIDINPUTKEY:	abcdefghijklmnopqrstuvwxyz
SELECTKEY:	b\040
SELECTKEY:	c
SELECTKEY:	d
SELECTKEY:	f
SELECTKEY:	j
SELECTKEY:	k
SELECTKEY:	l
SELECTKEY:	m
SELECTKEY:	p
SELECTKEY:	q
BACKSPACE:	\010\177
DELETEALL:	\015\025
MOVERIGHT:	.>
MOVELEFT:	,<
REPEATKEY:	\020\022
# the following line must not be removed
BEGINDICTIONARY
#
ah	~{0!0":G_9kgo9~}
ax	~{0!:G`D~}
av	~{0!:G~}
a	~{0!:G0"~}
aih	~{0#0$0%0&0'^_oM~}
aix	~{0$0(0)4t^_~}
aiv	~{0+0*0%`Hv0~}
ai	~{0.0/0-0,0%0&`H`IfHh(jSmA~}
anh	~{020001bVhqpF3'9cZO~}


From @cuvmb.cc.columbia.edu:CCNET-L@UGA.BITNET Sun Apr 22 15:39:14 1990
Received: from CUVMB.COLUMBIA.EDU (cuvmb.cc.columbia.edu) by sirius.ctr.columbia.edu (4.1/5.10) id AA08521; Sun, 22 Apr 90 15:39:10 EDT
Received: from CUVMB.COLUMBIA.EDU by CUVMB.COLUMBIA.EDU (IBM VM SMTP R1.2.1) with BSMTP id 8130; Sun, 22 Apr 90 15:37:12 EDT
Received: from PUCC.PRINCETON.EDU by CUVMB.COLUMBIA.EDU (Mailer R2.05) with
 BSMTP id 3486; Sun, 22 Apr 90 15:37:11 EDT
Received: from PUCC.BITNET by PUCC.PRINCETON.EDU (Mailer R2.07B) with BSMTP id
 9492; Sun, 22 Apr 90 15:38:43 EDT
Date:         Sun, 22 Apr 90 15:22:47 EDT
Reply-To: yjj%SIRIUS.CTR.COLUMBIA.EDU@cuvmb.cc.columbia.edu
Sender: Chinese Computing Network <CCNET-L%UGA@cuvmb.cc.columbia.edu>
From: yjj%SIRIUS.CTR.COLUMBIA.EDU@cuvmb.cc.columbia.edu
Subject:      Chtext's input method
To: Multiple recipients of list CCNET-L <CCNET-L@uga>
Status: RO

Chtext is the line editor, interpreter and viewer in the J.B. Wang's
Ch-series.  It demonstrates many new concepts for Chinese input.
One of them is to combine PinYin and some strokes (represented)
by symbols like t, s...).  There seems to be 1-1 correspondence between
such combination and GB code.  It's certainly very attractive if
what you type on the keyboard correspond to only on Chinese word
(PinYin alone cannot do that).  In speeds up the input.  Most other
(simple) input methods cannot achieve this.  Here are my questions.

1) How natural is this method to a typist comparing to other methods?
What is the limitation of this method.
(Roman text typing is close to speacking, the speed of which is limited
by the speed of thinking, while writing relates to the visual nerval
system.  The speed of typing is faster than writing and can be close
to the speed of thinking.  But The incorporating of strokes in
J.B. Wang's method requires you to imagin how the character is written.
It can never match the speed of English typing.  But with the enbeded
problem of Chinese, there may not be a better way than this.)

2) Is the way it uses to represent strokes the best?  Is there
any possibility to standardize how we represent the radicals and strokes?

Yuan

From SPO_JWANG2@sitbat.stevens-tech.edu Sun Apr 22 20:07:26 1990
Received: from SITBAT.STEVENS-TECH.EDU ([192.12.216.38]) by sirius.ctr.columbia.edu (4.1/5.10) id AA09044; Sun, 22 Apr 90 20:06:35 EDT
Date: Sun, 22 Apr 90 17:06:56 EST
From: SPO_JWANG2@vaxb.stevens-tech.edu (J.B. WANG)
To: yjj@sirius.ctr.columbia.edu
Subject: RE: Chtext's input method
X-Vms-To: EDU%"yjj@SIRIUS.CTR.COLUMBIA.EDU"
Message-Id: <9032217657.2040bc8f.SPO_JWANG2>
Status: RO


As Mr. Wang, Yongmin (5-stroke input method inventer) said in People's
Daily, there are currently 500 Chinese inputting schemes. Far sighter
all know that according to the law of natural competition, only one
of these 500 inputing schemes will survive. Althought 1/4 of world's
population is speaking and using Chinese, and the main part, i.e., the
mainland China is ruled by the communist party which has the absolute
power in setting up a standard (in contrast, USA and other democratic
states have to reply on defact standards, like ASCII). GB is the encoding
standard (not de facto in China). However, no one dared to set an inputing
standard. Therefore, significant effort is spent in China, Taiwan, especially
Singpore and overseas to find a inputing standard which will be the de facto
one.

As Yuan Jian pointed out, one of the Chtext inputing method is to append
abreviations of radicals to the end of a PinYin. This, however, was an old
idea I had in 1986 summer when I built the PinYin dictionary which is later
used in Chtext. The mapping is one-to-one, because Chtext is designed for
a mainframe dummy terminal, where graphics is not supported. In order to
edit an existing Chinese file (in GB), the way is ChText is to translate
GB into all PinYin (plus some escape commands). After editing, the file
is translated back to GB. If one 1-1 mapping, one has to do picking again
and again. This feature is only noticed by careful users like Yuan Jian,
because I did not document it. My believe is that it is a way for more
professional typist, if a table of such abbreviations is provided and ideally
standardized. For general users, I don't encourage using it. Nevertheless,
it is harmless even if you use it (probably inproperly) since auto-correction
function exists in ChText. For instance, to input Jian Yuan, the simplest
and best way is,

jian1-yuan2

or if you like,

j-y

where in the second instance, you have to browse through many characters in
the initial picking, but then it is saved in the Ci2Dian3. Therefore, 
afterward every time you type 

j-y

there is no more picking. Alternatively, you can also use single word input
mode, i.e.,

jian1ss yuan2ss

(I am sorry, it should be Yuan Jiang everywhere about, thus, it is

jiang1ss yuan2ss

)

where 'ss' means San1 Dian3 Shui3 (three dot water). There then is no picking.
Or since you know it is already unique, you can type

Jiang1ss Yuan2ss

or

jiang1ss= yuan2ss=

to tell chtext that you don't need to comfirm. Some then would say that
who knows 'ss' stands for 'three drops of water'. Prof. S.K. Chang, the
department chairman of Pitt CS department, showed me his paper that he'd do

jiang1sandianshui yuan2sandianshui

where, you can also try that out with Chtext, you will find that the
result is exactly the same as

jian1ss yuan2ss

except that Chtext worked a little harder to throw away the extra trailer.


In summary, PinYin plus CiZu and AI of context is definitely the way to
go. PinYin plus radicals is also a natural. Why, let's consider the situation
one telling the other over the telephone his name and address, in English,
he would pronounce and spell to avoid ambiguity, but in Chinese he would
pronounce each character and also provide things like radicals or common
context the word would appear. Therefore, to describe jiang1 yuan2 accurately,
over the phone, one would probably say,

jiang1, san dian shui yi ge gong
yuan2, san dian shui yuan-lai de yuan


What do you think folks?



J.B. Wang
------------

From @cuvmb.cc.columbia.edu:CCNET-L@UGA.BITNET Mon May 14 12:22:03 1990
Received: from CUVMB.COLUMBIA.EDU (cuvmb.cc.columbia.edu) by sirius.ctr.columbia.edu (4.1/5.10) id AA13613; Mon, 14 May 90 12:21:58 EDT
Received: from CUVMB.COLUMBIA.EDU by CUVMB.COLUMBIA.EDU (IBM VM SMTP R1.2.1) with BSMTP id 0899; Mon, 14 May 90 12:19:18 EDT
Received: from PUCC.PRINCETON.EDU by CUVMB.COLUMBIA.EDU (Mailer R2.05) with
 BSMTP id 4614; Mon, 14 May 90 12:19:17 EDT
Received: from PUCC.BITNET by PUCC.PRINCETON.EDU (Mailer R2.07) with BSMTP id
 4978; Mon, 14 May 90 12:21:31 EDT
Date:         Mon, 14 May 90 08:08:45 PDT
Reply-To: zhang qing <qing%tantalum.cchem.berkeley.edu%UCBVAX.BERKELEY.EDU@cuvmb.cc.columbia.edu>
Sender: Chinese Computing Network <CCNET-L%UGA@cuvmb.cc.columbia.edu>
From: zhang qing <qing%tantalum.cchem.berkeley.edu%UCBVAX.BERKELEY.EDU@cuvmb.cc.columbia.edu>
Subject:      Any info on speech recognition in China?
To: Multiple recipients of list CCNET-L <CCNET-L@uga>
Status: RO

        I got a letter from my mother who quoted 4.30. People's Daily
an article titled "Use voice to replace keyboard -- Voice controled Chinese
typewritter Succeed". The article reports that Guangdong Foshan Electronic
Music and Sound General Co. (guang3dong1 fo2shan1 dian4zi3 yin1xiang3 zong3
gong1si1) and Beijing Xinghe Electronics Co. developed a Chinese infomation
processing system which is capable of listening, recording, writing and
speaking. When the reporter read character byx character, the machine can
with "reasonably high (jiao4gao1)" accuracy, but also can automatically
select correct character from homophones(tong2yin1zi4). At present, the
machine can reach 30-70 characters per minute. After some training, it can
identify speech with accents. The more you use, the better it gets.
        Now, anybody have further infomation on this machine? On what kind
computer is this machine implemented? What's the CPU speed, amount of memory
? What is the sound identification accuracy, and what is the tong2yin1zi4
distinction accuracy? What research institution is behind it? Any place in
the US we can see it?
                        Yan Yong-Xin

From @cuvmb.cc.columbia.edu:CCNET-L@UGA.BITNET Tue May 15 18:06:06 1990
Received: from CUVMB.COLUMBIA.EDU (cuvmb.cc.columbia.edu) by sirius.ctr.columbia.edu (4.1/5.10) id AA21605; Tue, 15 May 90 18:06:01 EDT
Received: from CUVMB.COLUMBIA.EDU by CUVMB.COLUMBIA.EDU (IBM VM SMTP R1.2.1) with BSMTP id 2484; Tue, 15 May 90 18:03:19 EDT
Received: from PUCC.PRINCETON.EDU by CUVMB.COLUMBIA.EDU (Mailer R2.05) with
 BSMTP id 9345; Tue, 15 May 90 18:03:17 EDT
Received: from PUCC.BITNET by PUCC.PRINCETON.EDU (Mailer R2.07) with BSMTP id
 9544; Tue, 15 May 90 18:05:20 EDT
Date:         Tue, 15 May 90 08:08:16 -0700
Reply-To: Ed Lai <lai%APPLE.COM@cuvmb.cc.columbia.edu>
Sender: Chinese Computing Network <CCNET-L%UGA@cuvmb.cc.columbia.edu>
From: Ed Lai <lai%APPLE.COM@cuvmb.cc.columbia.edu>
Subject:      various subjects
Comments: To: CCNET-L%UGA.BITNET@cunyvm.cuny.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga>
Status: RO

>The speech recognition system suffered high failure rate. The fact
>is, you'll have to "train" it virtually every time you want to
>use it. The joke that the group had was that system would recognize
>your voice in the morning and it would reject you in the afternoon;
>it would recognize your happy voice, but would be confused the
>minute you change your mood. I would not be surprised if the newly
>introduced system behaved the same way.

A story which I have heard from MIT media lab may be interesting. At MIT
they have devloped a speech recognization system that seems to work quite
well, but when they demostrate to their sponsor, it completely breaks down.
Finally they figure out why. The person who is working on it is a graduate
student, and the sponsor you can guess it is the miltary, so you have a
few big shot military brass watching a graduate student doing a big demo,
you can see the anxiety level of the graduate student. The stress introduce
changes to the voice that may not be detectable by the human ear, but it is
very significant for the system, so it was a failure. Their solution is
suprisingly simple. When you are talking on the telephone, if the other
party is completely silent, soon you would be very uneasy, so they usually
make some sound like ha-huh so that the other party know you are still
listening. So at MIT they have their software recognize gaps in the speech,
and then generate the ha-huh sound as if the machine is listening. This
calm the speaker and the experiment can be conducted in normal.

>Actually, in the Chinese metafonts I've been developing (very slowly, and
>please don't write for information, as I don't have any), I've been putting
>in stroke counts and have been considering adding radicals (`bu4shou3').
>The problem with radicals is that there are so many radical systems around.
>I'm satisfied with the one used in _Kang1 Xi1 zi4dian3_, but it won't work
>at all for jiantizi.

For fanti, Kangxi's radical grouping is definitely the standard. For jianti
the one used in the GB document is pretty much the standard.

Scott, does that means that you have stroke count information for each
character in jiantizi in electronic readable form. If you do I am interested
in getting that information into my dictionary stack. I have it already for
fanti.

>Ed Lai of Apple wrote a Chinese dictionary for Macintosh HyperCard.  It is
>not a real dictinary, what it does is exactly what we are talking here:
>the Chinese indexing.

When I started the dictonary stack, my interest indeed is indexing, however
the indexing is indexing in an electronic media rather than a printed media
as is dicussed lately in CCNET. For me, indexing in printed media you cannot
do any better than what they are doing in dictionary or telephone book now.
so it is not very interesting. For electronic indexing it is wide open and
we can go much further. In version 0.1, I am just imitateing a paper
dictionary, i.e. by radical, by sound or by stroke count (stroke count is
for fanti version only since as I said I do not have that data). However, at
that time the superioity of electronic media is evident because of the rapid
naviagation that is possible, a mouse click is all it takes. So if you want
to find one(yi), but you want the complex version used on cheques, you have
no idea what radical it belongs to, you have no idea how to find it by sound
(not very likely?, but that version use bopomofo for fanti, so if you do not
know the bopomofo symbol, you cannot find the sound), but you can still find
yi (the one stroke version) by stroke count, after getting one stroke yi
click on the sound to get all the yi sound, then click on the mulitple stroke
yi to arrive the character you want. Since each step takes about one second,
it is not too bad even you have to take a big detour. For version 0.2, what I
did was to let you use either pinyin/bopomofo/zhuyin2/wade-gile/yale. Unlike
somebody else's philosophy, I believe that the user is the best judge of what
is the best for himself. Also the sound table can be arrange in alphabetical
order or as a two dimension matrix of consonant and vowel part, with a
representative character as the matrix element. In version 0.3, you can do
boolean search such as "all characters with water radical and yi sound",
"all characters with water compoent and wood component".

Is this the ulitmate? Hardly. I do not even call it version 1.0. There is a
lot more we can do when we have more data. I need stroke count data,
component data, type of stroke of each stroke data. Timothy Huang suggested
adding rhyming data and he sent me an article on what is rhyming but I cannot
understand it. Also I am interested in data on what sound are more similiar
and hence more easily confused (which incidently is useful for speech
recognization use). Also I am interesting in ci data rather than zi data so
that we can do search such as "give me the character X where X-dian is a ci.
The possiblity of electronic indexing is infinite.

/* Disclaimer: All statments and opinions expressed are my own */
/* Edmund K. Lai                                               */
/* Apple Computer, MS77-E                                      */
/* 20525 Mariani Ave,                                          */
/* Cupertino, CA 95014                                         */
/* (408)974-6272                                               */
zW@h9cOi

From @cuvmb.cc.columbia.edu:CCNET-L@UGA.BITNET Tue Jun  5 17:51:10 1990
Received: from CUVMB.COLUMBIA.EDU (cuvmb.cc.columbia.edu) by sirius.ctr.columbia.edu (4.1/5.10) id AA06207; Tue, 5 Jun 90 17:51:03 EDT
Received: from CUVMB.COLUMBIA.EDU by CUVMB.COLUMBIA.EDU (IBM VM SMTP R1.2.1) with BSMTP id 9315; Tue, 05 Jun 90 17:50:41 EDT
Received: from PUCC.PRINCETON.EDU by CUVMB.COLUMBIA.EDU (Mailer R2.05) with
 BSMTP id 5354; Tue, 05 Jun 90 17:50:39 EDT
Received: from PUCC.BITNET by PUCC.PRINCETON.EDU (Mailer R2.08A) with BSMTP id
 1779; Tue, 05 Jun 90 17:50:56 EDT
Date:         Tue, 5 Jun 90 12:56:00 EDT
Reply-To: "J.B. WANG" <SPO_JWANG2%STEVENS@cuvmb.cc.columbia.edu>
Sender: Chinese Computing Network <CCNET-L%UGA@cuvmb.cc.columbia.edu>
From: "J.B. WANG" <SPO_JWANG2%STEVENS.BITNET@cuvmb.cc.columbia.edu>
Subject:      RE: Mouse for stroke input
To: Multiple recipients of list CCNET-L <CCNET-L@uga>
Status: RO

There had been a number of discussions on Stroke input methods.

Lately I have also come across some papers on this topic. Of course, almost
everyone tried to say it was the best and would be the future.

To draw a summary:

(I mean brief sub-summary)

1) most of the stroke methods classify over 100 basic Chinese strokes into
 5~10 groups, and assign them to a key on the keypad or mother key board.
2) most of the stroke methods narrow down corresponding Chinese words
 as one increases number of strokes, and displays the selections on the
 screen.
3) most of the stroke input methods need strokes in sequence to make correct
 judgments.

The commercially most sucessful stroke input methods include:
1) Wang Yongming's WuBieXing (five-stroke method),
2) Pressure writing board input method

I personally have seen and or used 3 kinds of stroke inputing schemes, i.e.,
Wu-Bi-Xing and Wu-Bi-Hua, and one designed by a friend of mine using
nine keys on the keypad or mouse (mostly for creating input index).

My understanding is:

1) Only the pressure-sensitive writing board will have the real practical
 significance to survive.
2) All the rest of keyboard stroke input methods on any kind of machine
 have the same fatal problem as PinYin input method one~two years ago, i.e.,
 extremely high brain fatigue rate. In the view that you have to switch
 between the screen, key board and memory library (of stroke grouping, and
 sound -> stroke conversion). From this aspect, PinYin had less problem,
 because there was no stroke grouping and sound->stroke conversion steps.
 This alone may give stroke input the death sentence or life-time secondary
 input method sentence (just a little bit better than GuoBiao, or DianBao
 input methods).

Now PinYin+CiZu+Statistics+Grammer(AI) input method which are seen in
TianMa, FeiMa, Byx 2.0 and Ch-series is viewed to be the most prospective
input method in an article appeared on 5/25(?) issue of People's Daily.
If more and more effort is devoted, I personally believe it will become
THE method for keyboard inputing. Speech input and writing input do not
belong to keyboard input methods, but the speech recognition will be very
closely related to the above method PinYin input method.

Just like with English typing, there is a standard, and many auxilary
ones, e.g., two-finger typing (instead of 10), stroke input, radical
input, Bu-Shou input and etc may not become obsolete in Chinese keyboard
input in a short time. But one thing for sure, a brute-force table checking
input method will always be necessary, and today it is the GuoBiao input
method.


Sincerely yours,

JB Wang

p.s.: I will still be in Stevens for another 0.5 months
(201)-420-5814

From sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!uwm.edu!bionet!agate!ames!data.nas.nasa.gov!taligent!apple!lai Sun Sep  6 13:54:59 EDT 1992
Article: 592 of alt.chinese.text
Path: sol.ctr.columbia.edu!zaphod.mps.ohio-state.edu!uwm.edu!bionet!agate!ames!data.nas.nasa.gov!taligent!apple!lai
From: lai@Apple.COM (Ed Lai)
Newsgroups: alt.chinese.text
Subject: Re: About HanZi Input Method
Message-ID: <71788@apple.Apple.COM>
Date: 28 Aug 92 17:14:31 GMT
References: <1992Aug26.204534.17979@leland.Stanford.EDU> <Btnpsx.5zy@mentor.cc.purdue.edu>
Organization: Apple Computer Inc., Cupertino, CA
Lines: 16
Status: RO

>Is it some thing similar to the glossary input I described above?

Yes it is, I was using . instead of -, where . is meant to be abbreviation.
So you would be typing in j.s.j. In the special case where all abbreviation
is single letter, it can also be type as ~jsj  Abbreviation can also be
multiple letter, so you can have j.s.j. j.su.j. j.sua.j. Abbrevation can also
be no letter, so you can also have j..j. or (~j.j), this will be you wildcard
? so it is equivalent to your j-?-j

/* Disclaimer: All statments and opinions expressed are my own */
/* Edmund K. Lai                                               */
/* Apple Computer, MS37-UP                                     */
/* 20525 Mariani Ave,                                          */
/* Cupertino, CA 95014                                         */
/* (408)974-6272                                               */
zW@h9cOi


From CCNET-L@uga.cc.uga.edu Thu Apr 11 20:26:09 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.64/5.10) with SMTP id AA22817; Thu, 11 Apr 91 20:25:59 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 8676; Thu, 11 Apr 91 20:25:17 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 6828;
 Thu, 11 Apr 91 20:25:13 EDT
Date:         Thu, 11 Apr 91 17:23:48 -0700
Reply-To: Fung Fung Lee <lee%MILO.STANFORD.EDU@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L%UGA.BITNET@uga.cc.uga.edu>
From: Fung Fung Lee <lee%MILO.STANFORD.EDU@uga.cc.uga.edu>
Subject:      On BIG5 input methods
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

The lastest cxterm (11.4.2) is indeed a very nice program.

As for BIG5 input methods, one easy way to build something better than IC
is to use hcu (Hanzi Converter)(available in june.cs.washington.edu:pub/yeung).
For example, one can generate a pinyin input method for BIG5 by converting
"PY.tit" bundled with cxterm as follows:

% hcu -m GBtoBIG < PY.tit | sed -e 's/ENCODE: GB/ENCODE: BIG5/' > PY-B.tit
% tit2cit PY-B.tit > PY-B.cit

Insert a line such as the following into .Xdefaults:
 	<KeyPress> F9:	switch-HZ-mode(PY-B)	\n\

Similary, one can generate CTCPS3-B.tit, the BIG5 version of a Cantonese
input method, and other phonetic based input methods back and forth
between GB and BIG5.

Due to the imperfect mapping between GB and BIG5, input methods generated
this way would be far from 100% satisfactory, since you may not be able
to input some BIG5 hanzi, the JianTi (GB) form of each of which correspond to
more than one FanTi (BIG5) forms.

You may also want to replace the "PROMPT" line in PY-B.tit, as there is
no exact correspondence of the "double colon" used in PY.tit, hence hcu
translates it into an empty box symbol.

Fung F. Lee


>From: Yongguang Zhang <ygz%CS.PURDUE.EDU@uga.cc.uga.edu>
>Subject:      New release: cxterm (rev 2), celvis (bug-fixes)
>
>* New Version of Chinese xterm (a Chinese terminal emulator on X11R4)
>
>
>The following are new in cxterm 11.4.2 (which means X11R4 cxterm Rev 2):
>
>  1)  Input methods are separated from the code of cxterm program.
>      They are stored in external files and are loaded at run time
>      on demand.
>
>  2)  There are programs to convert external input methods between
>      human readable format (.tit) and cxterm loadable format (.cit),
>      so that users can change the existing input methods or add their
>      own input methods without recompiling cxterm.
>
>  3)  The user can redefine keys to switch input mode (without
>      recompiling cxterm).
>
>  4)  Encodings other than GuoBiao (GB) are supported, such as BIG5.
>      i.e. cxterm is encoding independent.
>
>  5)  For GB coding, we have 3 builtin input methods, and 5 external ones,
>      including IC (internal coding in hex), pin yin, pin yin with tones, etc.
>
>  6)  For BIG5 coding, only input method IC is provided. I don't know
>      any popular input method in BIG5.

From CCNET-L@uga.cc.uga.edu Mon Nov 12 03:07:33 1990
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.64/5.10) with SMTP id AA26223; Mon, 12 Nov 90 03:05:26 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 5800; Mon, 12 Nov 90 03:02:23 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 8671;
 Mon, 12 Nov 90 03:02:21 EST
Date:         Mon, 12 Nov 90 00:54:27 MST
Reply-To: mleisher%NMSU.EDU@uga.cc.uga.edu
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: mleisher%NMSU.EDU@uga.cc.uga.edu
Subject:      Pinyin parser
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

Below is a short Pinyin parser that doesn't depend on tone numbers for
most cases.  It only parses Mainland Pinyin spellings right now, but I
am working on parsers for Cantonese and Taiwan Pinyin spellings.  It
does not do any spell checking, but this would be very easy to add.

The parser takes a phrase and breaks it into the individual Pinyin.
There are some cases that are ambiguous, like "huan" vs. "hu an" and
"bin an suo" vs. "bi nan suo".  In cases like these, some sort of
divider is needed to get the correct parse; like a tone number, space,
tab, or carriage return.

I am working with a small sample of phrases(about 3000) that have no
tone numbers.

This code is sort of a first draft, so it's VERY messy, but if you
have comments, bug reports, or notice that I missed something, please
send me email.  Please feel free to modify the code for your own use.

BTW, this code was written to be part of a modified X11R4 Athena
AsciiText Widget that has Chinese and Japanese support.  I hope to get
it out soon.

----------------- Code Start -------------------------
/*
 *        Pinyin Parser
 *
 * Copyright (C) 1990 Mark Leisher.
 *
 * Author:   Mark Leisher
 * Internet: mleisher@nmsu.edu
 * Bitnet:   mleisher@nmsu.bitnet
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 1, or (at your option)
 * any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * A copy of the GNU General Public License can be obtained from this
 * program's author (send electronic mail to mleisher@nmsu.edu) or from
 * the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA
 * 02139, USA.
 */

#include <stdio.h>
#include <stdio.h>
#include <ctype.h>

#define isvowel(c) (c == 'a' || c == 'e' || c == 'i' || c == 'o' || c == 'u')

char *
downcase_string(str)
char *str;
{
    int c;
    char *out = (char *)malloc(strlen(str) + 1),
         *ptr = out;

    while((c = *str++)) *ptr++ = isupper(c) ? tolower(c) : c;
    *ptr = '\0';
    return out;
}

#define INITIAL_ARRAY_SIZE 5
#define ARRAY_INC 5

char **
parse_pinyin(str, n)
char *str;
int *n;
{
    int found = 0, c, len, i = 0, arr_size,
        vcount = 0;
    char *ptr, **out = NULL;

    *n = 0;
    while (*str) {

/* skip initial non-alphanumeric chars */

        while (!isalnum(*str)) str++;
        ptr = str;

        while (!found && (c = *str++)) {
            if (!isalnum(c)) {
                found = 1;
            } else {
                if (isvowel(c))
                  vcount++;
                else
                  vcount = 0;

                if (vcount == 2 && *str != 'i' && *str != 'o' && *str != 'n') {
                    if (isdigit(*str))
                      str++;
                    if (*str == 'r') {
                        if (*(str+1) && !isvowel(*(str+1)))
                          str--;
                    }
                    found = 1;
                    vcount = 0;
                } else {
                    switch(c) {
                      case 'e':
                        if (*str == 'r') {
                            if (str - ptr <= 1)
                              str++;
                            if (isdigit(*str))
                              str++;
                            found = 1;
                        }
                        break;
                      case 'b': case 'w':
                        if (*str == 'o' || *str == 'u') {
                            str++;
                            found = 1;
                        }
                        break;
                      case 'c': case 'r': case 's': case 'z':
                        if (*str == 'i') {
                            str++;
                            found = 1;
                        }
                        break;
                      case 'f': case 'm': case 'p':
                        if (*str == 'u') {
                            str++;
                            found = 1;
                        }
                        break;
                      case 'y':
                        if (*str == 'e') {
                            str++;
                            found = 1;
                        }
                        break;
                      case 'n':
                        if (*str == 'g') {
                            str++;
                            c = *str;
                            if (c && isvowel(c)) {
                                switch(c) {
                                  case 'e':
                                    if (!*(str+1) || *(str+1) != 'r')
                                      str--;
                                    found = 1;
                                    break;
                                  default:
                                    if (c != 'a')
                                      str--;
                                    found = 1;
                                    break;
                                }
                            } else
                              found = 1;
                        } else if (!isvowel(*str) && !isdigit(*str))
                          found = 1;
                        else if (isvowel(*str) && str - ptr > 1) {
                            found = 1;
                        }
                        break;
                    }
                    if (!found && isvowel(c) &&
                        !isvowel(*str) && *str != 'n') {
                        if (isdigit(*str))
                          str++;
                        found = 1;
                    }
                }
            }
        }
        if ( !found ) str--;  /* probably on end of string */
        found = 0;
        len = str - ptr;
        if ( out == NULL ) {
            out = (char **)malloc(INITIAL_ARRAY_SIZE * sizeof(char *));
            arr_size = INITIAL_ARRAY_SIZE;
        } else if ( i == arr_size ) {
            arr_size += ARRAY_INC;
            out = (char **) realloc(out, arr_size * sizeof(char *));
        }
        out[i] = (char *)malloc(len+1);
        strncpy(out[i], ptr, len);
        *(out[i]+len) = '\0';
        i++;
        ptr = str;
    }
    *n = i;
    return(out);
}

main(argc, argv)
int argc;
char **argv;
{
    char *str = argv[1], **pinyin = NULL;
    char buff[50];
    int i, n;

    do {
        scanf("%[^\n]\n", buff);
        pinyin = parse_pinyin(buff, &n);
        for ( i = 0; i < n; i++ )
          printf("%s ", pinyin[i]);
        putchar(10);
        free(pinyin);
        pinyin = NULL;
    } while (!feof(stdin));
}

From CCNET-L@uga.cc.uga.edu Mon Jul 29 23:19:47 1991
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.61/5.10) with SMTP id AA14597; Mon, 29 Jul 91 23:19:40 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP R1.2.2MX) with BSMTP id 4324; Mon, 29 Jul 91 23:19:29 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.07) with BSMTP id 2272;
 Mon, 29 Jul 91 23:19:22 EDT
Date:         Mon, 29 Jul 1991 22:17:00 EST
Reply-To: Dorothy Day <DAY%IUBACS.BITNET@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L%UGA.BITNET@uga.cc.uga.edu>
From: Dorothy Day <DAY%IUBACS.BITNET@uga.cc.uga.edu>
Subject:      RE: seeking info. on Chinese computer systems
To: Multiple recipients of list CCNET-L <CCNET-L@UGA>
Status: RO

The following reply is to Carlos McEvilly's query about Chinese
software from the mainland and Hong Kong.
        --Dorothy Day, Indiana University
        DAY@IUBACS

-----------------------------------------

Date:         Mon, 29 Jul 1991 12:17:38 EDT
From:         LCKPO@fulton.crc.uno.edu
Subject:      Re: seeking info. on Chinese computer systems

Dr McEvilly:
   When I was in China in 1986 People's Daily ran a story about a
PC based program that was similar to IBM's "learning to read". Local
peasants could speak into the system and the software would return the
appropriate character. They claimed an 80% success rate in acheiving
rudimentary literacy with this software.
   Also, I believe that China's foreign language's press has a catalog
of Chinese language software available. I have this somewhere and it
contains about 40 titles, but I don't have it readily available.
   Finally, Morgan-Kaufman, the computer science/artificial intelligence
publisher has a Hong Kong office for their PRC/ Taiwan/ Hong Kong operations.
They are at Transglobal Publishers Service, Ltd.
27/F Unit E, Shield Industrial Centre
Tsuen Wan N.T.
Hong Kong
Tele: 413 5322
Fax: 413 7049.
     I'm sorry I couldn't identify specific software, but this material
should give you a good set of leads.
                               Lawrence C. Katzenstein
                               Political Science
                               The University of New Orleans

From CCNET-L@uga.cc.uga.edu Tue Sep 22 10:56:34 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA11776; Tue, 22 Sep 92 10:56:27 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 4156; Tue, 22 Sep 92 10:54:25 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 4658; Tue, 22 Sep 92 10:54:22 EDT
Date:         Tue, 22 Sep 1992 22:56:13 SST
Reply-To: "(Guo Jin)" <guojin@ISS.NUS.SG>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "(Guo Jin)" <guojin@ISS.NUS.SG>
Subject:      Re:  input method
Comments: To: CCNET-L@uga.cc.uga.edu, oass@OAS.UNIZH.CH
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

> > Dear reader
> The so-called Pinyin input-method for Chinese Script T/7 drives me to
> desperation. Does  anyone know a better method similar to Ed Lai's cihui or
> is anyone working on it?
> Informations would be very much appreciated.
> Brigitte Koella
>

Chinese input method is alway a hot topic on this forum, and
the most popular ones, I think, are PinYin-based.  Of course
the Pinyin-and-selection loop is of 'desperate', and maybe 'cihui'-like
ones are more or less the same.  Thus, a dozen of developers truned to
so-called 'continuous input' which tries to convert or trancript
a string of pinyin (pinyin phrase or sentence) into a string of hanzi
(Chinese graphic character phrase or sentence).  'Excellent' examples
are:  TIANMA (Canada), InSun (PRC), PJS (PRC & Singapore), XLBR (US),
cWnn (Japan) and THED (PRC).   (Most of them are commecial software with
a special hardcard for copyright protection)

The problem is that:  no convertion or transcription method can achieve
100% accuracy.  Even worse, all these sentence-based input methods are
still employed in an isolate-character-input style and thus hard to use.
(Someone has already complained the XIN TIANMA in this forum.)

Does anyone interested in shaping out an user-freindly sentence-based input
routine (or style, approach, usage, standard) under the assumption of the
 existance
of a pinyin-to-hanzi convertion subroutine with an accuracy about 95% ?
(If needed, I will contribute such a subroutine late this year or early next
 year.)
Again, if we already have such a convertion subroutine in hand, HOW TO USE IT??

I am sure that such a discussion will both encourage the Chinese software
developers to explore and/or improve their products and civilize our
end-users the existance and goodness of such continuous input approaches.
A reference fact is : to input Chinese characters (kanji) into computers,
Japanese people have used kana (Japanese pinyin ?) to
Kanji (hanzi) conversion method for decades.

GUO Jin, ISS, NUS, Singapore.

From CCNET-L@uga.cc.uga.edu Tue Sep 22 16:43:51 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA16118; Tue, 22 Sep 92 16:43:48 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6641; Tue, 22 Sep 92 16:41:46 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 9341; Tue, 22 Sep 92 16:40:44 EDT
Date:         Tue, 22 Sep 1992 16:39:45 EDT
Reply-To: Yuan Jiang <yjj@ctr.columbia.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yuan Jiang <yjj@ctr.columbia.edu>
Subject:      PCs in Taiwan and Chinese input
Comments: To: ccnet-l@uga.bitnet
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I visited Taiwan last month.  I was really impressed by the wide use of PCs
in offices (I did not have a chance to visit any house hold).  Before I
went there, I thought the difficulty in keying in Chinese would prevent the
wide use of computers.  But when I visited the offices and a hospital
there, PCs are used everywhere from word processing to hospital
administration, just like in the US.

CangJie is the most popular input method for typists, >100 characters per
minute.  Most people who are not professional use BoPoMoFo input.  One
office worker told me the she could key in ~40 characters per minute.

To my luck, the World Trade Center had a computer exhibition, Taipei
International Computer Applications Show.  I saw many interesting products
there.  I collected many brochures and will write summaries for each
category when I find the time.  I regret that I did not have a chance to
meet some of the CCNETers in Taiwan.

From CCNET-L@uga.cc.uga.edu Tue Sep 22 21:12:27 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA19066; Tue, 22 Sep 92 21:12:16 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 8099; Tue, 22 Sep 92 21:10:06 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 2261; Tue, 22 Sep 92 21:10:04 EDT
Date:         Tue, 22 Sep 1992 18:10:13 -0700
Reply-To: Ed Lai <lai@APPLE.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ed Lai <lai@APPLE.COM>
Subject:      Re:  input method
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

>Chinese input method is alway a hot topic on this forum, and
>the most popular ones, I think, are PinYin-based.  Of course
>the Pinyin-and-selection loop is of 'desperate', and maybe 'cihui'-like
>ones are more or less the same.  Thus, a dozen of developers truned to
>so-called 'continuous input' which tries to convert or trancript
>a string of pinyin (pinyin phrase or sentence) into a string of hanzi
>(Chinese graphic character phrase or sentence).  'Excellent' examples
>are:  TIANMA (Canada), InSun (PRC), PJS (PRC & Singapore), XLBR (US),
>cWnn (Japan) and THED (PRC).   (Most of them are commecial software with
>a special hardcard for copyright protection)

This is a very strange view of how good an input method is. What is called
continous input is called inline input in Mac system 7.1, whereas the Mac
system 6 input is called bottomline input. Both are possible under system
7.1. But I would hardly judge an input method to be better solely because
it is inline. In fact I consider inline input to be "Batch" mode whereas
bottom line input could be (but not necessary) "Interactive" mode.

The fact it is not a real sentenced-based input unless the syntax and
the context of the whole sentence is used to help the conversion. And
the input methods mentioned above, the ones I know are certainly not
really sentenced based, but are really single phrase conversion where
the conversion are delayed until it batches up.

As for the question of Cihui under ChineseTalk II on the Mac, I have no
plan to do it since I do not even have the information on how to write
an input method in ChineseTalk II. However the plan is to port Cihui to
System 7.1

/* Disclaimer: All statments and opinions expressed are my own */
/* Edmund K. Lai                                               */
/* Apple Computer, MS37-UP                                     */
/* 20525 Mariani Ave,                                          */
/* Cupertino, CA 95014                                         */
/* (408)974-6272                                               */
zW@h9cOi

From CCNET-L@uga.cc.uga.edu Tue Sep 22 23:17:57 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA20840; Tue, 22 Sep 92 23:17:42 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 8682; Tue, 22 Sep 92 23:15:26 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 6211; Tue, 22 Sep 92 23:15:25 EDT
Date:         Tue, 22 Sep 1992 21:46:35 CDT
Reply-To: yuan <A10RYT8%NIUCS.BITNET@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: yuan <A10RYT8%NIUCS.BITNET@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

 2 year(s) ago(not very sure), I noticed some news regarding hand-writing
 input method while I was in taipei. Several days ago, I got the
 news from China Daily News, Oversears Ed., saying that, this
 product is started marketing to local market and overseas market
 like Singapore and Hongkong, and has a hot sale. I don't have the news
 in detai ls, hopefully somebody who knows this tech can add more to this.
 Here I roughly discribe the so called "hand-writing" input method in
 the following (please send corrections, if something wrong, should be
 a lot).

     The user use a pen-like stuff, write characters down on a
 pad-like stuff. This is exactly the way Chinese's writing behavior.
   The software will recognize what is writing
 down. According to the report, the corrent precise recognization
 rate is 98(?). The software has the ability of "self-learning"
 which means, it will get more familiar with user's "hand-writing
 style" while it is kept using.

  /\/\/\/\/\/\/\/\
 /  --yuan         -------
  || How have you been?---------------
  !! Your CORRECTIONS & RESPONSES are appreciated! ------------
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

From CCNET-L@uga.cc.uga.edu Wed Sep 23 01:06:34 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA21652; Wed, 23 Sep 92 01:06:15 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9127; Wed, 23 Sep 92 01:04:13 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 9507; Wed, 23 Sep 92 01:04:11 EDT
Date:         Wed, 23 Sep 1992 13:05:31 SST
Reply-To: "(Guo Jin)" <guojin@ISS.NUS.SG>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "(Guo Jin)" <guojin@ISS.NUS.SG>
Subject:      Re:  Hand-writing Chinese input
Comments: To: A10RYT8%NIUCS.BITNET@uga.cc.uga.edu, CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

>
>  2 year(s) ago(not very sure), I noticed some news regarding hand-writing
>  input method while I was in taipei. Several days ago, I got the
>  news from China Daily News, Oversears Ed., saying that, this
>  product is started marketing to local market and overseas market
>  like Singapore and Hongkong, and has a hot sale. I don't have the news
>  in detai ls, hopefully somebody who knows this tech can add more to this.
>  Here I roughly discribe the so called "hand-writing" input method in
>  the following (please send corrections, if something wrong, should be
>  a lot).
>
>      The user use a pen-like stuff, write characters down on a
>  pad-like stuff. This is exactly the way Chinese's writing behavior.
>    The software will recognize what is writing
>  down. According to the report, the corrent precise recognization
>  rate is 98(?). The software has the ability of "self-learning"
>  which means, it will get more familiar with user's "hand-writing
>  style" while it is kept using.


To my knowledge, at least half a dozen such "on-line handwritten Chinese cha
character recognition systems" have been marketed in mainland China,
especially in the famous "electronic street" of the "silicon valley"
(Haidian Zhongguancun) in Beijing, for about 5 years.
The prices are about RMB 1,000 - 2,000 (US$ 150 - 300, at current official
'tiaoji' exchange rate).

To my knowledge, the 'hot sale' never goes beyond the order of 1000 sets,
and almost no buyer keeps in using it to 'write essay'.  In my opinion,
the main obstacles are:

(1) Too slow.  Not the computer to recognize your writing, but your writing!
Just imagine copying an article through writing palm-sized Chinese character
on a palm-sized plate one at a time.  Moreover, to ensure a 'good' recognition
rate, you have to write 'regularly, neatly, clearly, stroke-by-stroke, ...'

(2) Not so accurate.  98% ?  It can achieve 99.9% or more, but by the
developer in the lab!  For 'unfamiliar' end user, less than 80%.  Ok, you also
can get a 'very high' recognition rate through careful practice to fit the
machine.  Hint: don't write 'too fast'.

(3) Hard to use.  This is in the same spirit as I triggered yestoday on the
'continuous pinyin input'.  Yes, no 100% accuracy.  Thus, when you write down
a palm-sized character on the palm-sized plate, you are always anxious to
look at your CRT to check the correctness.  If it is wrong, you are willing to
correct it imidiately.  Atter two or three times repeatition, you may
congratulate yourself of writing down a 'correct' Chinese character.  At the
same time, however, you have forgotten what to write next!

Anyway, on-line hand-writing is a new thing to both the developers and our
end user.  The developers may do more R&D to let their systems more
 'intelligent't
In my opinion, however, the strong feedback of our end users on how to use
such a new 'high tech' might be much more important than the R&D effort to
improve the recognition rate.  The problem is not the recognition approach
itself but HOW TO USE IT!!!

Based on the above BIASED comments, if you still want to buy one, I will
provide a little more COMMERCIAL info.

Discussions are extremely welcomed.
GUO Jin, ISS, NUS, Singapore

From CCNET-L@uga.cc.uga.edu Thu Sep 24 00:37:46 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA03291; Thu, 24 Sep 92 00:37:25 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6302; Thu, 24 Sep 92 00:35:19 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 7888; Thu, 24 Sep 92 00:35:17 EDT
Date:         Thu, 24 Sep 1992 00:35:45 EDT
Reply-To: Yuan Jiang <yjj@ctr.columbia.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yuan Jiang <yjj@ctr.columbia.edu>
Subject:      Pen-based input
Comments: To: A10RYT8%NIUCS.BITNET@uga.cc.uga.edu, CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I saw the demos of Genius pen input pad and some other companies.
I'll write more details later.  They only work if you
write neatly.  They are not popular.  Most people think they
are slow because their inaccuracy.

From CCNET-L@uga.cc.uga.edu Thu Sep 24 02:46:47 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA03798; Thu, 24 Sep 92 02:46:35 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6608; Thu, 24 Sep 92 02:44:38 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 0208; Thu, 24 Sep 92 02:44:35 EDT
Date:         Thu, 24 Sep 1992 14:36:59 SST
Reply-To: "(Guo Jin)" <guojin@ISS.NUS.SG>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "(Guo Jin)" <guojin@ISS.NUS.SG>
Subject:      Re: input method
Comments: To: ccnet-l%uga.BITNET@nusvm.nus.sg
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Two days ago, following a question on "input method", I asked: How to use
"continuous input". (guojin@iss.nus.sg, CCNET-L, Sept. 22, 92)

Instantly, knowledgeable responses come up.  Ed. Lai, the contributor of
the familiar "cihui" package, felt, "This is a very strange view of how
good an input method is." (lai@apple.com, CCNET-L, Sept. 22, 92)  Y. Jiang,
one of the list owners of the CCNET-L, asked me "What do you mean ..."
(yjj@ctr.columbia.edu, personal comm., Sept. 22, 92).  Agreeing with A.
Boretz's idea of "There hasn't been much discussion on the net of this
particular area" (xhgy@crux2.cit.cornell.edu, CCNET-L, Sept. 22, 92), I am
aware of the necessity of further clarifying my question.

Isolated Input
=============

Almost all Chinese character input methods we are currently in use, in my
previous term, are "isolated" ones.  No matter "character-based" or "word-
based", no matter "shape based" (e.g., WBZX, CangJi) or "sound based (e.g.,
pinyin, BoPoMoFo), no matter "explicitly-coded" (e.g., WBZX) or
"implicitly-coded" (e.g., LianXiang), and no matter "pure" or a "hybrid" of
some of the above, they possess a common characteristics: the current
entering of a character or a word is INDEPENDENT of both the characters or
words already entered (no matter they are "correct or wrong") and the
characters or words to be entered.  Moreover, the character or word is a
single unit that leads to only two possible results of your input: correct
or wrong, succeed or fail.  That is, a character or a word is "isolatedly"
inputted.  Of course you can postpone the correction of an input error (the
so-called batch mode).  You, however, have to agree that you have already
failed to input the character or word due to your mistakes such as the
forgetting of the special character code or whatsoever.

Continuous Input
================

In contrast, in "continuous" input methods, until the end of the entering
of a "large" block (a phrase, a sentence, a paragraph, or even a whole
book!) of codes (pinyin or whatever you like), you may have no chance to
know confidently even the first Chinese character you really entered.
Moreover, even if your code-inputs are 100% accurate, it is guaranteed that
you may NOT make 100% perfect Chinese-character-inputs.  In other words,
the Chinese character you are entering is both depending on its previous
character(s) and on its following one(s).


The core ...
============

According to the above description, the core for continuous input is NOT
"sentence based" but "having an imperfect batch conversion process behind".
Sentence-based input, of course, is one kind of continuous input.  It is
also very clear that there is no reason to convince our end-user the
superior of continuous input approach over isolated ones.  In fact, we even
do NOT know an approapriate way to use the so called continuous input
approach, not to mention its advantages.

My Question
===========

"I am already tired to read your wordy jargons, what do you really want to
convey?!"  My objective, in this first stage, is only to raise a "simple"
question:  HOW TO USE CONTINUOUS INPUT METHOD???

Why ...
=======

This is a meaningful yet crucial question.  Intuitively, at least in my
mind, the continuous input sounds much better than isolated input.  For
instance, with continuous pinyin input, you may pay no attention to the
tedious homonym problem.  The current application practice, however, is the
opposite.  In spite of the existence of so many pro-continuous input
packages in the market and/or on the network, how many users really employ
them as their prime Chinese character input means?!  Why? We even do not
know an appropriate way to do continuous input, not to mention to take
their full advantages.

To the developers end ...
=========================

To the developers end, in my opinion, little serious and/or extensive R&D
has been made on the usage of continuous input.  To certain extent, the
attachment of a conversion input routine is only a marketing strategy:
attracting unaware customers (sorry!).  In some products, the conversion
input routine is simply embedded totally in the same way as an isolated
coding routine, not to mention the far from satisfactory conversion routine
itself.

To the end-users end ...
========================

To the end-users end, the confusion of isolated and continuous input
approaches block them to take full advantages of the goodness some products
already provided.  For example, early this year, a netter wrote:  "I have
had a chance to use ... and frankly I am disappointed. ... I found that
that 'self converting pinyin->hanzi' not to be all it is cracked up to be.
... when it didn't work you had to move the cursor back and select change,
then select the new character and then move forward again, so even though
you don't have to select every word, you sill have to read what you type as
you type it, and when you do have to make a change it is not all that
quick."  Clearly, he used the continuous input routine in the same way as
he uses an isolated input routine and thus did not capture its
conveniences.

My Imagination
==============

"What kind of answer you are looking for?"  An imagination of the operation
procedure of using a continuous input routine to type in an article.  The
one now in my mind is:  first, blindly key in the article in pinyin.  Then,
double-click a mouse to switch on the batch conversion process.  When the
conversion of a sentence finished, two accompanied subroutines are
automatically triggered.  One is a Chinese spelling checker and/or grammar
watchdog that will highlight any possible input/conversion error. (Ref.:
it is informed that such a checker will be in market this December in
Beijing.  Estimated price: RMB 2,000)  The other is a Chinese speech
synthesizer witch will take the results of the linguistic analysis of the
sentence from the converter and read the sentence phrase by phrase with
clear intonation. (Ref.:  tens of Chinese speech synthesizers have already
in market for at least a decade.  Price ranges from as low as RMB 300 to
far more than RMB 10,000, depending on varieties of quality features and
configurations.)  The idea is, if the converter makes a wrong conversion,
in certain cases, the linguistic analysis of this sentence is also improper
that naturally leads the speech synthesizer produces a strange reading
which in turn let you easily capture the bug.  What ever means with which
you find the error, you may simply move the mouse arrow to the position of
the wrong Chinese character and double-click the mouse button to trigger a
small window of all possible candidates (homonyms), then click the correct
Chinese character to replace the wrong one and the whole process goes on.
It is better than any existing input methods, isn't it?!

Brainstorm
==========

Of course, shaping a new input style is never an easy job.  Too many
factors have to be taken into account.  The one I imaged above is
technically feasible (even not too difficult for a Chinese software
developer to implement) but may not be economically affordable.  Anyway, if
we can make clear our desire and/or demand, I am sure that, dozens, if not
hundreds, of experts and amateurs are already there waiting for developing
and/or integrating such a product.  What kind of continuous input product
we are really looking for ?!  Brainstorm!

guojin@iss.nus.sg

From CCNET-L@uga.cc.uga.edu Thu Sep 24 18:17:58 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA10708; Thu, 24 Sep 92 18:17:40 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 1402; Thu, 24 Sep 92 18:15:19 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 1331; Thu, 24 Sep 92 18:15:17 EDT
Date:         Thu, 24 Sep 1992 18:15:48 EDT
Reply-To: Yuan Jiang <yjj@ctr.columbia.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yuan Jiang <yjj@ctr.columbia.edu>
Subject:      Re: input method
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I'm very glad input method is taking the stage again.  I believe
Chinese input still remains the biggest obstical in Chinese systems.
All the other problems will be solved given enough time and labor
poured in.  But Chinese input needs innovation and revolution.

I may agree with many points GuoJin raises.  However, many of terms
he uses are ambiguous.  (That's why I asked what he meant when saying
"continuous input.")  I'd like to stress the following.
1) Zi (character) is very often mistakenly translated into "word."
2) Ci is closer than Zi to "word" in English.
3) A Ci is usually composed by 2 ZIs.

Input methods can be devided into 3 most important categories.
A. Pen based (stroke recognition) or character recognition based.
  Stroke recognition and character recognition are still not
  mature technologies.  On one hand Chinese characters have
  simple regular strokes and Chinese might be easy, on the other
  hand they average have more strokes than Latin aphabets.
B. Speach based.  It has the same and more technical problems
  than phonetic keyboard based methods.
C. Keyboard based.
 a) Map Chinese character set <-> key stroke sequences
    Canji, 5 Stroke and more many more methods have invented on
    this line.  It is basically mapping written Chinese into
    a "foreign language."  If Chinese can all learn such a
    "foreign language," they could have learned English to make
    life easier.  I believe this is a dead end.
 b) Phonetic input methods.  PinYin, bopomofo, etc.

A good method should be judged by 2 criteria.
1) 1<->1 mapping of an input (sequence) and Chinese character set.
2) Easy to learn, convenient, price, ..., in a word, User Friendly.

Method A meets 1) very well, but has many technical problems
remain to be sovled.  It may take years.  Method C.a does not
meet 2).

Method C.b may be the most promising one in the near future.
It started from inputing PinYin (or Bopomofo) -> Zi (character).
This method fails criterion 1) because of homonyms in Chinese.
It fails 1) almost 100%.  This might be close to the "isolated input"
method mentioned by Guojin.
TianMa started PinYin sequence -> Ci and phases.  It meets 1),
say 60%-95% (any one has the statistics?).  This might be the
beginning of the "continuous input."

Again, the core os the issue is not "isolated input" versus
"continuous input," or "interactive" versus "batch mode," it's
rather meeting the 2 criteria.  TianMa nad XLBR have basically
the same input method.  But XLBR is more popular because it's
more user friendly -- much cheaper and very good page/text formating.

Where should we go from PinYin sequence -> Ci and phases to
remove the remaining 40%-5% ambiguity?  I imagin 2 ways:
1) dynamic change of PinYin sequence -> Ci and phases mapping
  library to adopt the writing style of the user;
2) make PinYin sequence -> Ci and phases mapping more context
  sensitive.  This may be closer to Guojin's continuous input.
  XLBR makes the most frequently used Zi and Ci as the defualt
  ones when homonym occurs, making use of statistics.  It's
  still not context sensitive.
Any more suggestions?  Let's not simply brushing off other guys'
ideas.  Brainstorming is good for everyone.  I'll try to pursuaed
Yan YongXin to buy an MCI MAIL so that he can get onto the net.

Comments to Guojin's
>An imagination of the operation
>first, blindly key in the article in pinyin.  Then,
>double-click a mouse to switch on the batch conversion process.  When the
Pressing space bar or any punctuation is a more natural way to trigger
the convertion.  TianMa and XLBR all adopt this.

>automatically triggered.  One is a Chinese spelling checker and/or grammar

Yes, PinYin checking (spelling check) is another thing need to be added.
The software should be Ci, phrase and context sensitive when
doing the mapping (converting).

From CCNET-L@uga.cc.uga.edu Thu Sep 24 18:29:47 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA10840; Thu, 24 Sep 92 18:29:40 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 1456; Thu, 24 Sep 92 18:26:35 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 1814; Thu, 24 Sep 92 18:25:55 EDT
Date:         Thu, 24 Sep 1992 18:25:57 EDT
Reply-To: Yuan Jiang <yjj@ctr.columbia.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yuan Jiang <yjj@ctr.columbia.edu>
Subject:      Re: Chinese Windows in Taiwan
Comments: To: ccnet-l@uga.bitnet
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Comment to my previous postings
====================
> - provide 5 input methods: CangJie, bopomofo, Quick learn(Su4cheng2), Inner
                             ^^^^^^^  ^^^^^^^^
                            Changjei  Phonetic
 Correction of spelling, It was used in Taiwan ( I think it is also the
standar spelling.) C-Win 3.1 hasn't been announced yet. We use version 3.0
and it is good.

From CCNET-L@uga.cc.uga.edu Thu Sep 24 18:31:21 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA10857; Thu, 24 Sep 92 18:31:16 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 1471; Thu, 24 Sep 92 18:29:23 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 1918; Thu, 24 Sep 92 18:28:06 EDT
Date:         Thu, 24 Sep 1992 18:28:00 EDT
Reply-To: Yuan Jiang <yjj@ctr.columbia.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yuan Jiang <yjj@ctr.columbia.edu>
Subject:      Re: Pen-based input
Comments: To: ccnet-l@uga.bitnet
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Comment to my previous posting
========================
I had a pen-based input device (not Genius, cost
NT$3990). I bought it not only because it's inexpensive but it's more
clever than Genius. It attracted me because I can do a "continuous write".
Even though, I still don't think it is good enough. i.e. I must write
carefully and slowly and must check what happened on screen. For a
experienced user it takes too much time.

From CCNET-L@uga.cc.uga.edu Fri Sep 25 02:08:13 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA13774; Fri, 25 Sep 92 02:08:04 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 3315; Fri, 25 Sep 92 02:06:05 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 6449; Fri, 25 Sep 92 02:06:01 EDT
Date:         Fri, 25 Sep 1992 14:08:07 SST
Reply-To: "(Guo Jin)" <guojin@ISS.NUS.SG>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "(Guo Jin)" <guojin@ISS.NUS.SG>
Subject:      Re: input method
Comments: To: ccnet-l%uga.BITNET@nusvm.nus.sg
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

This is my third mail on Chinese input method.  Again, it is to
clarify some "ambiguous" in my previous mail :)

character/word
==============

In my mind,

	(Chinese) character == zi4
	(Chinese) word == ci2

I do understand the existence of so much ambiguous on the term
"word".  In fact, until this century, no one differentiates zi/ci.
The modern concept of "word (ci)" was first discussed by Ma3
Jian4Zhong1 in 1898 in his famous book "ma3 shi4 wen2 tong1" ("Ma's
grammar"? The first Chinese grammar book).  Even today, so many word
dictionaries are still titled "(da4) zi4dian3" in Taiwan, Hong Kong
and other places.  So many people accept

	zi4 == (Chinese) character
	    == (Chinese) word
	ci2 == (Chinese) phrase

However, before an "operational" definition of the term "word" (read:
with this definition, given a "zi4chuan4", i.e., a Chinese character
string, anyone can make sure if it is a "word".) being widely
accepted, no one can rule out any of the translations above.
Example, "mei3 li4 jian1 he2 zhong4 guo2 (United states of America)".
Word or phrase? How many words in it?  Endless argues!

Discussion:  Does there exist the so called "word" in Chinese
language or just a "cheating" from western language?!

One-to-one mapping
==================

Then, why do I always use the term "Chinese character or word (read:
zi or ci)"?  No matter character (zi) input or word (ci) input, both
are in the ward of "one-to-one mapping".  In these two input
paradigms, an identical ID code is always tried to be marked on the
unit (zi/ci).  When you key in the correct ID code, you are
guaranteed you will get the zi/ci you intend to get.  If the computer
sends you a strange zi/ci, that is yours fail: you typed a wrong ID
code!  This is one aspect the term "isolated input" implied.
Continuous input?  No one-to-one mapping!

Discussion:  Why one-to-one mapping?!

Touch input
===========

Classified on the type of input devices, Y. Jiang listed yesterday 3
Chinese character input categories:  pen-based, phone- (speech)
based, and keyboard-based.  As a reference, the fourth category, in
my term, pointing-device-based, or touch input, is added here.  Its
three popular sub-categories are:

(1)  mouse-based

"Kai3Te4 Ren2Ren2 Xi4Tong3 (KT everyone system?)", a Chinese system
developed by Dr. Hou in Xi'an JiaoTong Univ. and popular in mainland
China, supports "keyboard-less" Chinese character input.  To input a
Chinese character, the user, using a mouse on a computer screen,
first selects the first stroke or the radical of the character from a
stroke or radical window, then, after the emerging of the window with
all Chinese characters of the same first stroke or radical (even with
some symbols with the 'similar' first stroke or radical.  Wa!
Treating western symbols as they are derived from Chinese characters!),
selects the exact character.  In a word, a way similar to peak up a
character from a dictionary.  A similar package is also developed in
Singapore on UNIX stations.

Still keyboard input?  Yes, ONE key keyboard!

(2)  screen-based

Instead of the nimble mouse, using your finger to point on the
special sensitive screen.

(3)  plate-based

Instead of the harmful screen, using a large array of Chinese-
character-marked touch buttons.

Still keyboard input?  Yes, ONE key a character!

Discussion:  Possible classifications other than device-based?

I will go on the discussion of Chinese character input later.  Thanx.
guojin@iss.nus.sg

From CCNET-L@uga.cc.uga.edu Fri Sep 25 17:10:03 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA22912; Fri, 25 Sep 92 17:09:58 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7969; Fri, 25 Sep 92 17:08:11 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 8946; Fri, 25 Sep 92 17:08:10 EDT
Date:         Fri, 25 Sep 1992 17:08:47 EDT
Reply-To: Yuan Jiang <yjj@ctr.columbia.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yuan Jiang <yjj@ctr.columbia.edu>
Subject:      Re:      Re:      Re: input method
Comments: To: ccnet-l@uga.bitnet
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  Mail from 'Yuan Jiang <yjj@ctr.columbia.edu>' dated:         Fri,
              25 Sep 1992 16:38:09 EDT
Status: RO

The Byte articles says that the most advanced Japanese input
methods are using context sensitive schemes.  His $450 Japanese
Sharp wordproessor guess right 80% of the time.  Well, I don't
think this is a high percentage.  It will frustrate me.

From CCNET-L@uga.cc.uga.edu Tue Sep 29 01:14:57 1992
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA27803; Tue, 29 Sep 92 01:14:53 -0400
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 1475; Tue, 29 Sep 92 01:12:41 EDT
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 9772; Tue, 29 Sep 92 01:12:39 EDT
Date:         Tue, 29 Sep 1992 13:14:31 SST
Reply-To: "(Guo Jin)" <guojin@ISS.NUS.SG>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "(Guo Jin)" <guojin@ISS.NUS.SG>
Subject:      ? on Chinese grammar
Comments: To: ccnet-l%uga.BITNET@nusvm.nus.sg, wujianhu@iscs.nus.sg
Comments: cc: wuzhibia@iscs.nus.sg, zhangyiq@iscs.nus.sg
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Hi, Jianhua,

Yesterday, I sent to you through the inner mailbox parts of HAN Di's
Ph.D dissertation, "computer speech recognition and expert system methods
for entering Chinese characters to computers".  It includes about 30
pages dictionary codes and ATN graphs.

The main problem he encountered is:  the resolution of homonyms in
pinyin to hanzi transcription.  Two accompanied ambiguouses are:  less
of tonal and word boundary information.  The main idea (not new, even in
Chinese) is to use lexical, syntactic and semantic constraints as
filters.  I don't think his work a breakthrough.  It is, however,
representative.

There are two kinds of NLP systems:  constructive and destructive.  The
ultimate goal of a constructive system is to build an inner (deep)
representation (logic form or whatsoever).  The destructive system is
simply to rule out any "impossible" hypotheses.  The former is always
try to say "it is ..." while the later is always to say "it is not ...".
But I never understand the real difference between them.

Theoretically, the logic for constructive system is:  if you can not say
it is impossible, it is possible.  On the other hand, the logic for
destructive system is:  if not possible, it is impossible!

I think the logic is very important in grammar rule formation.


Each Chinese NLP system has its own linguistic rules.  It is not hard to
collect them.  But how to use them?  Each system has its own assumptions
and emphases.  It is hard, if not impossible, to merge them together.


You see, how many ambiguous in my mind!  Any info. about grammar and any
of your progress in this area, pls share with me.  Talk to you later.

guojin@iss.nus.sg

From @UGA.CC.UGA.EDU:CCNET-L@UGA.BITNET Mon Feb 15 11:54:49 1993
Return-Path: <@UGA.CC.UGA.EDU:CCNET-L@UGA.BITNET>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AB29852; Mon, 15 Feb 93 11:54:40 CST
Message-Id: <9302151754.AB29852@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 5066; Mon, 15 Feb 93 11:50:02 EST
Received: from UGA.BITNET (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail
 V1.1c/1.7e) with BSMTP id 2236; Mon, 15 Feb 1993 11:26:40 -0500
Date:         Sat, 13 Feb 1993 15:50:17 EST
Reply-To: Liedong Zheng <lz17@CUNIXB.CC.COLUMBIA.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Liedong Zheng <lz17@CUNIXB.CC.COLUMBIA.EDU>
Subject:      Wang Yongmin and his Wang Code
Comments: To: ccnet-l@uga.bitnet
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Dear CCNET netters:

        There was a meeting here at Columbia University
yesterday (Feb. 12) with Professor Wang2 Yong3 Min2,
and general manager of WANGMA Computer Company, Mr.
ZHENLU XU. They were invited by the Chinese Division
of the United Nations. I would like to share some of the
info with you:

Wang Code (Wu3 Bi3 Zi4 Xing2 input method):

        Wang Code input method is a planar input method.
A chinese character is a planar figure, you catch the image of
the whole character when you read it. On the other hand,
other input methods including PinYin for chinese and all
western languages are linear, you have to read it from
one direction to the other. This somehow slows down your
speed. As a simple example, you can test the time you
spend counting the number of "o"'s in the following two
arrays:



                o   o   o   o   o   o   o   o   o



                                o    o    o

                                o    o    o

                                o    o    o

Usually, the second array is easier to count, and that
is the core of the Wang Code. The average key strokes
needed to input one chinese character are 2.6, and no
more than 4 key strokes to input one character. It takes
one day to learn Wang Code for a high school student,
and after some typing training, as you need to
be a secretory in western countries, for example, the
average input speed is most probably greater than 100
chinese characters per minute. Consider the typing
speed for an average secretory in the U.S., this is not
that slow though. With Wang Code, you can create new
characters, new phrases. It has Lian2 Xiang3
functionality, it can be used for both simplified and
traditional characters. They are pushing it to be the
national standard now.

In mainland China:

        Wang Code input method is installed on eighty-five
percent of the computers, and it is being used in almost
all the publishing houses, newspapers and all chinese
embassies and consulates and the Chinese Division of the
United Nations. About two million people in China are
trained to use Wang Code, and the average input speed for
most of them is greater than 100 chinese characters per
minute.  It is one of the curricular requirements in about 1200
elementary schools, high schools, including "Shanghai High
School" (SHANGHAI ZHONG XUE). About 10% of the reporters
use Wang Code in their NoteBook or desktop computers.

In overseas:

        Wang Code is being used by Wen Hui Daily in Hongkong,
Lian2 He2 Zao3 Bao4, Lian2 He2 Wan3 Bao4 in Singapore.
It became part of the curriculum of some chinese
elementary and high schools in Malasia.

WANGMA 480 Desktop System:

        This is their latest software package.  It has bitmap
48 song, fang3song4, kai, and hei (black) (GB48) in
firmware,  vector fonts from 8x8 to 480x480, pull-down
menu, up to nine windows for editing. It uses high memory,
and is compitable with PC/XT/AT, 286,386,486, supports HP
laser printers, and X.25 protocal, and some other features.
        Their earlier products cost from $95 (software only)
to $300 (software + card)

        Interested users my find the following useful:

        Mr. Wang Yongmin, or Mr. Xu Zhenlu
        R00m 401 Beijing Science Hall (Ke1 Xue2 Hui4 Tang2)
        Friendship Hotel
        Beijing, China 100873

        Tel:    861-8498736
        Fax:    861-8498737

        Store:

        WANGMA Computer Company
        Room 209, No.37
        BaiShiQiao Lu
        Beijing

        Tel:    861-8421346
        Fax:    861-8421291

        Autherized dealer in North America:

        COMAX SYSTEM INC.
        43-14 Main Street
        Flushing, NY 11355

        Tel:    (718)-762-2822
        Fax:    (718)-762-4320

--------------------------------------------------------
Disclaimer:  I have no connection with the companies
mentioned above. This is only for the information of
networkers of CCNET and I assume no liability to the
accuracy of the information provided here.


Liedong Zheng
Chemistry Department
Columbia University

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Feb 19 13:48:23 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA20164; Fri, 19 Feb 93 13:48:11 CST
Message-Id: <9302191948.AA20164@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9488; Fri, 19 Feb 93 14:44:59 EST
Received: from UGA.BITNET (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail
 V1.1c/1.7e) with BSMTP id 8318; Fri, 19 Feb 1993 14:42:25 -0500
Date:         Fri, 19 Feb 1993 14:42:31 EST
Reply-To: Dong-Ping Deng <deng@OWL.RHIC.BNL.GOV>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Dong-Ping Deng <deng@OWL.RHIC.BNL.GOV>
Subject:      Pig in Mule
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

To: ccnet-l@uga.cc.uga.edu
Subject: Pig in Mule
--text follows this line--

Are there any netters using Pig inside Mule?  If there are some, we may start
of a mailing list to exchange tricks and tips.

Here are some quotes from Pig

Version: 1.1
Author:  Ping Zhou  zhou@okra.fnal.gov
Date:    November 16, 1992
...
Several ways can be used to input a CiZu:
As an example all following can be used for the input of Ji4 Suan4 Ji1
j-s-j j.s.j j-su-ji ji4-suan4-ji1 ji-suan4-ji jsj ji4suan4ji1
you can see there are more possibilities. The input is case insensitive.

...
If you have added new entries to your CiDian, you have to save the CiDian
 permanently using the command M-x pig-save-CiDian. Otherwise you will loose
 the entries for the new CiZu once you quit Emacs or Mule.
If you want a copy of your CiDian in the PinYin-GuoBiao format at any time
  use the command M-x pig-extract-CiDian.
Consequently there are two ways to build up a CiDian:
 either as you editing files by inputing unknown CiZu
 or edit the CiDian file directly and then use the command
 M-x pig-build-CiDian to convert it to the internal format.
...


I just cooked up one way to build a large CiDian quickly ;-). Here you are:

1) Grab a HanYu article from anywhere, then extract your favorite CiZu and
arrange them as follows.

~{>-<C~}
~{?*J<~}
~{?FQ'~}
~{?`DU~}
~{?l@V~}
~{@,;x~}
~{C@9z~}
~{C@V^~}
~{CqVw~}
~{DOK9@-7r~}
...

2) Use TommiKaikkonen's gb2pinyin.c to find  pinyin for these cizu. You need
to modify gb2pinyin.c a little bit (put "-" instead of " "). Then use your
favorite Emacs utilities (like kill-rectangle, yank-rectangle etc.) to
arrange them as follows:

jing1-ji3-                         ~{>-<C~}
kai1-shi3-                         ~{?*J<~}
ke1-xue2-                          ~{?FQ'~}
ku3-nao3-                          ~{?`DU~}
kuai4-le4-                         ~{?l@V~}
la1-ji1-                           ~{@,;x~}
mei3-guo2-                         ~{C@9z~}
mei3-zhou1-                        ~{C@V^~}
min2-zhu3-                         ~{CqVw~}
na1-si1-la1-fu1-                   ~{DOK9@-7r~}
....

3) Follow instructions in Pig to build your internal CiDian.el.

Have fun.

Caveat: Duo-yin-zi is not well taken.

--

~{5K6,F=~}

~{5K~}@bunny.rhic.bnl.gov

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Feb 23 16:43:43 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA13095; Tue, 23 Feb 93 16:43:37 CST
Message-Id: <9302232243.AA13095@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 0865; Tue, 23 Feb 93 17:40:48 EST
Received: from UGA.BITNET (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail
 V1.1c/1.7e) with BSMTP id 4741; Tue, 23 Feb 1993 17:40:47 -0500
Date:         Tue, 23 Feb 1993 17:41:09 -0500
Reply-To: Stephen G Simpson <simpson@MATH.PSU.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Stephen G Simpson <simpson@MATH.PSU.EDU>
Subject:      ChangJie Su Ru fa
Comments: To: "L.S.Ng" <L.S.Ng1@southampton.ac.uk>
Comments: cc: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <9302231716.AA00308@leibniz.math.psu.edu>
Status: RO

"L.S.Ng" writes:
 > To any users of ChangJie Su Ru fa, I seek answers to these questions:
 >
 > 1. Do you know if ChangJie or ChangXie is the correct pronunciation ?


This question came to my attention also, because of the [][] input
method in ET (CHANGJEI.TAB).  I did a little research, and here is
what I found out.

Since the mailer chews up Big-5 Chinese, I replace the Chinese
characters by [] and append a uuencoded version below.  (I wish there
were something like hz for Big-5!)

I got the following information from: A New Practical Chinese-English
Dictionary, Editor in Chief Liang Shih-Chiu, Far East Book Co. Ltd.
(I translate the pronunciation into pinyin.)

6797 [] (jie2) -- to deduct; to omit.
[] [] (jie2 hang2) -- 1. to match; to contest.  2. (of birds)
to fly high and low.  3. haughty; conceited.

191 [] (cang1) -- 1. a granary. 2. a cabin, as in the ship.
3. green.  4. a Chinese family name.
...
[] [] (cang1 jie2) -- Tsang Chi, a legendary figure credited
with the invention of the written Chinese language.
...

However, according to Ross Paterson <rap@doc.imperial.ac.uk> the jie2
character is pronounced this way only in Cang1 Jie2, the name of the
legendary hero, but elsewhere it is pronounced xie2.  Here is what
Ross says:

> > 6797 -- [] (jie2) -- to deduct; to omit.
> > [] [] (jie2 hang2) -- 1. to match; to contest.  2. (of birds)
> > to fly high and low.  3. haughty; conceited.
>
> Hmmm, my little Xinhua Zidian says it's pronounced xie2 for these meanings,
> and jie2 in Cang1 Jie2.  The zidian file has an entry
>
>       f2a1    BE65    xie2    fly upwards, neck, (surname)
>
> Best regards,
>
> Ross

begin 644 jie2.big
M1G)O;3H@02!.97<@4')A8W1I8V%L($-H:6YE<V4M16YG;&ES:"!$:6-T:6]N
M87)Y+ I%9&ET;W(@:6X@0VAI968@3&EA;F<@4VAI:"U#:&EU+"!&87(@16%S
M="!";V]K($-O+B!,=&0N"@HV-SDW("TM(+YE("AJ:64R*2 M+2!T;R!D961U
M8W0[('1O(&]M:70N"KYE(.#>("AJ:64R(&AA;F<R*2 M+2 Q+B!T;R!M871C
M:#L@=&\@8V]N=&5S="X@(#(N("AO9B!B:7)D<RD*=&\@9FQY(&AI9V@@86YD
M(&QO=RX@(#,N(&AA=6=H='D[(&-O;F-E:71E9"X*"C$Y,2 M+2"MW" H8V%N
M9S$I("TM(#$N(&$@9W)A;F%R>2X@,BX@82!C86)I;BP@87,@:6X@=&AE('-H
M:7 N"C,N(&=R965N+B @-"X@82!#:&EN97-E(&9A;6EL>2!N86UE+@HN+BX*
MK=P@OF4@*&-A;F<Q(&II93(I("TM(%1S86YG($-H:2P@82!L96=E;F1A<GD@
M9FEG=7)E(&-R961I=&5D"G=I=&@@=&AE(&EN=F5N=&EO;B!O9B!T:&4@=W)I
M='1E;B!#:&EN97-E(&QA;F=U86=E+@HN+BX*"E1H:7,@8V%M92!T;R!M>2!A
M='1E;G1I;VX@8F5C875S92!O9B!T:&4@K=P@OF4@*$-H86YG:F5I*0II;G!U
0="!M971H;V0@:6X@150N"B!O

end

Best wishes,

Stephen G. Simpson
Department of Mathematics, Pennsylvania State University
333 McAllister Building, University Park, State College, PA 16802
Office phone: +1 814 863-0775      Internet: simpson@math.psu.edu
Home phone: +1 814 238-2274        Bitnet: T20@PSUVM
FAX: +1 814 865-3735

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Mar  2 10:00:18 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA09041; Tue, 2 Mar 93 09:59:17 CST
Message-Id: <9303021559.AA09041@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2202; Tue, 02 Mar 93 10:54:15 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU
 (LMail V1.1d/1.7f) with BSMTP id 2218; Tue, 2 Mar 1993 10:54:12 -0500
Date:         Tue, 2 Mar 1993 14:57:17 GMT
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      The merits of fast typing
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Summaray : Discuss the merits of fast typing and introduce Chang Jie.

This
message uses HZ encoding. Some mailers may distort this by changing tilde ^ to
carat ˆ or percent ^. If so when you read it, try to change carat or percent
back to tilde, then encode by HZ.

^{4rWVFU<0;/^}  ^{2Vr!JdHk7(^}
^{#^#^#^#^#^#^#^#^#^#^#^^}

^{HtR*J95gDTND;/N*VPND5DR;2?7]#,DGC4JWR*Lu<^>MJG#,4rWVVAIYR*SkP4WVR;Qy?l!#749}@4
 K5#,514rWVFU1i;/AK#,VPND>M;a8|=xR;2=5XOV4z;/AK!#::SoF4RtSk<rLeWVJGVPNDOV4z;/5DO
 HG}#,4rWV>MJGVPNDTY=x;/5D1X>-V.5@!#^}

^{NRCGS&8C7EFzNe1JWVPNSk2Vr!JdHk7(JGW(R54rWVT15DW(@{5D>I9[Dn!#Ok51DjS"ND4rWVR2R;
 6H1;51WwR;VVW(R5#,5+JG5=AK=qLlUb8v<<Ju<:>-OsWT@4K.R;0c#,3IN*FdND;/V.R;2?7]AK!#K
 yRT#,R;VV?l=]JdHk7(JGHKHKS&Q'5D!#V;Ph;(7Q<88vP!J15DJ1<dQ';aR;VV?l=]JdHk7(#,1cD\
 VUImJ\SC!#Pm6`Q'U_TZVPND5gDTMxBg7"1mB[ND#,6<RTS"NDN*Vw#,FdVPR;8vT-Rr#,R2JGRrN*2
 ;6.5CR;VV?l=]JdHk7(#,6x2;5C2;SCS"ND4rWV!#^}

^{GR2;B[Ne1JWVPNSk2Vr!JdHkJkSEJkAS#,NROkTZ4KL8L82Vr!JdHk7(5D<8On;y1>#,HCSPPKH$5D
 EsSQ@4Q'O0!#^}

^{OH@4R;8vP!A7O0#:#(DcPhR*R;8vSP2Vr!JdHk7(5DHm<^#,NRMF<vDO<+PG!##)^}
^{V8J>#:R@WE::WVOB1_5DS"NDWVD8GC<|#,C?R;WV04R;OB?U8q!#^}

^{=q^}  ^{Ll^} ^{Ll^} ^{Fx^}  ^{Gg^}   ^{@J^}, ^{L+^}  ^{Qt^}  ^{8_^}   ^{8_^}
  ^{TZ^}  ^{IO^}
OIN MK MK OMN AQMB IIB KI NLA YRBR YRBR  KLG  YM

^{Dc^}  ^{NJ^}  ^{NR^}   ^{N*^}  ^{J2^} ^{C4^} ^{5t^}   ^{Q[^}   ^{@a^},   ^{DQ^
 }  ^{5@^}   ^{Dc^}  ^{2;^} ^{Cw^} ^{0W^} ^{JG^}   ^{N*^}  ^{AK^}  ^{0.^}
ONF LSR HQI IKSI OJ HI QYAJ BUAV EBU , EOG YTHU ONF MF AB HA AMYO IKSI NN BBKE

^{SC2Vr!JdHk7(4rWV#,JG8y>]WVPN04<|5D!#NRAm8=IOR;2Vr!WV8y1m#(#P#C#XND<^#)#,GkDcR;
 1_2NUUWV8y1mR;1_Q'4rWV!#N*0o<GFpDGP)WV8y#,Gk748413KP?Z>w#(HUTB=pD>K.;pMA#,P15c=
 ;2fW]:a94#,HKPDJV?Z#,2`2"QvE&7=27#)!#^}

^{!.=q!/WVH!Bk7=7(:\<r5%#,<404!.HK8j9-!/H}8vWVPN!#!.Ll!/WVH!!.R;4s!/!#!.Fx!/H!!.
 HKR;9-!/!#^}

^{!.@J!/WVH!BkJ10QWsSR7V?*#,Ws1_5DH!JWN2A=8vWVPN#,VP<d5D?IJ!BT#,<4H!!.8j8j!/#,J!
 BT!.HU!/#,OB@4H!SR1_5D!.TB!/#,H+2?N*!.8j8jTB!/!#^}

^{!.L+!/WVM,IO!#^}

^{!.Qt!/WVR27VN*WsSR#,Ws1_5D6zETH!!.9-VP!/#,2;R@1JK3#,V;R@WVPNIOVAOB#,JG2Vr!5DLX
 5c#,UbR2JG1_!.?4!/1_!.4r!/5DWTH;7=J=!#Sk!.@J!/M,@m#,H!!.9-VPHU!/!#^}

^{!.8_!/WVSIIOVAOBH!!.27?ZTB?Z!/!#^}

^{!.TZ!/WVSIMbVADZH!!.4sVPMA!/!#^}

^{!.IO!/WVM,@m!#^}

^{!.Dc!/WV7VWsSR#,WsH!!.HK!/#,SRH!!.9-;p!/#,<4!.HK9-;p!/^}

^{!.NJ!/7VMbDZ#,MbPNH!A=8v6T3F5DWsSRWVPN<4!.VPJ,!/#,2;H!5c#,DZH!H!!.?Z!/#,<4!.VP
 J,?Z!/^}

^{!.NR!/WVSI!.F2#(Vq#)JV8j!/H}8vWVPNWi3I!#^}

^{!.N*!/WVR@PrSIIOVAOB#,MbVADZH!!.8j4sJ,8j!/#,!.A&!/WVSI!.4sJ,!/Wi3I#,UbR2JG2Vr!
 5DLX5c#,<4R@WVPN2;R@1J;-!#@`KF5DSP!.>E#(4s9-#)!/!.56#(J,Vq#)!/5HWV!#JTH!!.0l!/!
 .Mh!/6^WV!#^}

^{!.J2!"C4!/V1H!!#^}

^{!.5t!/7VWsSR#,WsN*!.JV!/#,SRN*!.27HUJ.!/#,<4!.JV27HUJ.!/^}

^{!.Q[!/7VWsSR#,WsN*!.TBI=!/3ID?WV#,SRN*!.HUE.!/3ItˆWV!#^}

^{!.@a!/H!K.ET#,D?DK!.TBI=!/^}

^{!.DQ!/WV#,Ws1_!.SV!/WVJt!.K.!/#,SR1_!.v?!/WVH!!.HKMA!/#,RrN*4KWV3#SC#,Ih<FU_0Q
 K|6(ReV;H!FdJWN2#,@`KF5DSP!.CE!/!.9m!/!#^}

^{!.5@!/OHH!4,ET#(27#)#,JG2Vr!5DLX5cV.SIWsVASR#,2;R@1JK3!#!.JW!/WVN*!.X%VqTBI=!/
 #,5+WwN*WVIm#(SR1_#)#,9f6(V;H!H}Bk#,<4R;!"6^!"D)5D!.X%VqI=!/!#^}

^{!.2;!/N*!.R;;p!/!#!.Cp!/R`N*!.R;;p!/#,Ub3FN*VX84WVOVOs!#0QIYSC5DAPN*VX84WV#,H!
 BkJ1TZT-1>!.R;;p!/<SIOR;8v#X#,<4!.R;;p#X!/!#^}

^{!.Cw!/V1H!#,!.0W!/N*!.VqHU!/!#^}

^{!.JG!/N*!.HUR;27HK!/#,UbJG3#SCWV#,RK6`A7O0!#^}

^{!.AK!/H!!.9-9-!/#,!.0.!/H!!.TBTB4sK.!/!#^}


^{A7O0AKRT:s#,SPAKR;P)8EDn#,>M?IRTQ'O02Vr!5DT-@m#,D\H!R;GPDQWV!#@}Hg#:!.2X!/!.N"
 !/!.=E!/S&Hg:NH!#?!.P!!/!.9$!/!.UI!/!.F_!/TuH!#?5H5H!#^}

^{2Vr!JdHk7(8EB[^}
^{#^#^#^#^#^#^#^^}
^{2Vr!JdHk7(9fTrSIT-TrSkM(TrA=2?7VWi3I!#4rWVJ1R@T-Tr=+5%WV7V8nN*;y1>WVPN#,TYR@M(
 TrH!Bk!#;y1>WVPN#,N;VC#,WV8y#,?Z>w5D6(Re#,Gk2NUUWV8y1m!#KyN=!.Bk!/7:V8WV8y<0FdN
 ;VC#,H!Bk<4!.04OB8w;y1>WVPNKyJt5DN;VC!/#,<S04?U8q>MMj3IJdHkAK!#^}

^{9fTr^}
^{#^#^^}
^{R;!"T-Tr#:>+<r!"MjU{!"LXUw^}
          ^{>+<r^}  ^{HtSP6`VVH!Bk?ID\#,BkJ}WnIYN*U}H7!##(<{76@}#)^}
          ^{MjU{^}  ^{HtBkJ}O`5H#,OHH!71845DWVPN!##(<{76@}#)^}
          ^{LXUw^}  ^{1#3VT-WVLXUw#,>!A?2;VX5^H!PN<01\CbTZW*=G4&7V8n!#^}

^{6^!"M(Tr#:<rM<^}            ^{5%^}  ^{WV^}
                            ^{#|^}
                  ^{#|#^#^#^#^#^#^#^#^#^#^#^#^#|^}
                ^{A,LeWV^}                    ^{7VLeWV^}
              ^{#(OˆKDBk#)^}                    ^{#|^}
                                      ^{WVJW^}  ^{#+^}  ^{WVIm^}
                                   ^{#(Oˆ6^Bk#)^} ^{#(OˆH}Bk#)^}

    ^{VPND5%WV=a99;y1>IOSPIOOB!"WsSR!"DZMbH}4s@`!#H!BkJ15DK3PrJGSIIOVAOB#,SIWsVA
 SR^}    ^{SIMbVADZ!#^}

    ^{A,LeWV5DH!Bk7=7(<r5%V1=S#:OˆH!KDBk#:KDBkRTOB5DWVH!H+WV8y#,3,9}J1H!R;6^H}D)
 !#^}
    ^{7VLeWVH!Bk#,7VN*WVJWSkWVIm!#^}
^{#(2Vr!5DWVJW6(ReN*Ws!"IO!"Mb5HF+ET#,2;B[JG7q2?JW!#WVImN*SR!"OB!"DZ5D2?7]#)^}
    ^{H!WVJWOˆ6^Bk#,6^BkRTOB5DWVJWH!H+WV8y#,3,9}J1H!JWN2!#^}
    ^{H!WVImOˆH}Bk#,M,@mH}BkRTOB5DWVImH!H+WV8y#,3,9}J1?<BGRTOBLu<^#:^}
        ^{R;!"HtWVImDKR;A,LeWV#,H!FdR;6^D)!#^}
        ^{6^!"HtWVImDKR;7VLeWV#:?<BG^}
            ^{#A#.FdWVJW5%Bk#:H!WVJW5D5%Bk#,H!WVIm5DJWN2^}
            ^{#B#.FdWVJW84Bk#:H!WVJW5DJWN2#,H!WVIm5DN2Bk^}

    ^{H!BkJ1#,S&J!BT2;1XH!5DBk#(<{@}#)#,1XPkJ!BTJ1#,OHJ!BT0|:,TZDZ5DWVPN!#^}


^{76@}^}
^{#^#^^}
^{#1#.A,LeWV#:Ll!"N*!"Bm!"Dq!"9m!"V1!"OD^}
^{#2#.FdS`6`N*7VLeWV!#^}
^{#3#.2;1XCcG?7V@`!#RTIO5D9fTr?4KF84TS#,FdJ5K|5D8EDn<r5%5C:\!#Ub:\DQSCSoQT=bJM#,
 Ph^}    ^{R*6`A7O0>MD\Al;aAK!#^}
^{#4#.>+<r#:!.@<!/#,?IH!!.X%R;R;!/;r!.=pR;R;R;!/#,5+RT!.X%R;R;!/N*>+<r!#^}
^{#5#.MjU{#:!.7r!/#,?IH!!.JVHK!/;r!.J.4s!/#,5+RT!.JV!/=O!.J.!/7184!#^}
^{#6#.LXUw#:Ub8vT-Tr=OIYSC4&#,@}Hg!.?w!/N*!.R;R;E.J,!/^}
^{#7#.J!BT#:@}!.G=!/#=MAMA=pLo#,J!BT!.?Z!/^}

^{@}Mb^}
^{#^#^^}
^{5+72SoQTW\SP@}Mb#,5gDT4rWVR22;@}Mb#:^}
^{#1#.3#SC84:OWV#:CE#(VPJ,#)9m#(Vq8j#)v?#(HKMA#)5HSCSZWVIm;rWVJWJ1#,V;H!6^Bk!#^}
^{#2#.4s!";p!"D>9a4)WVPN#,2;R@IOOBK3Pr!##(<P#=4sX%#,@4#=D>X%#,3K#=VqD>VPPD#)^}
^{#3#.!.DQWV!/#:>J#(Vq#X#)#,K`#(VP#X#)#,B9#(8j#XPD#)#,Im#(Vq#XVq#)#,<f#(X%#X=p#)
 !##4#.VX84WV#:IYSC5DVX84WV#,H!T-Bk<S#X!#HgRTIO5D!.Cp!/#(R;;p#X#)!#^}
^{#5#.SPJ1R;P)WV5D1`BkDQRT=bJM#,5+UbV;JtIYJ}#,>M51WwSoQT5DR;0c5D@}Mb:CAK!#^}

^{5sWjWV^}
^{#^#^#^^}
^{KyN=5sWjWVRbV8DQRT=bBk5D5%WV#,5+V;R*?49}R;4N#,>MD\;PH;4sNr!"@N@N<GWEAK!#?49}R;
 F*V.:s#,>MD\Cw0W2Vr!JdHk7(5D>+R*AK!#^}

^{WVJW#:#(IO#)^}
^{Bˆ^}  ^{LoVP9-8j^}      ^{5[^}  ^{27TBVPTB^}    ^{U_^}  ^{J.4sHU^}    ^{>m^}
 ^{;pJVJ,I=^}
    ^{#^#^^}              ^{#^#^^}            ^{#^#^^}          ^{#^#^^}
^{4:^}  ^{JV4sHU^}
    ^{#^#^^}
^{WVJW#:#(Ws#)^}
^{=+^}  ^{VPR;9-8j8j^}    ^{4N^}  ^{8jR;9-HK^}    ^{Iq^}  ^{8j;pVPLoVP^}
    ^{#^#^^}              ^{#^#^^}            ^{#^#^^}
^{UI^}  ^{J.4s^}        ^{SV^}  ^{9-4s^}        ^{7r^}  ^{JVHK^}        ^{F_^}
 ^{J.I=^}
^{6z^}  ^{J,J.^}        ^{>.^}  ^{X%X%#X^}      ^{Hk^}  ^{HKVq^}        ^{0K^}
 ^{VqHK^}
^{D)^}  ^{D>J.^}        ^{1>^}  ^{D>R;^}        ^{Jx^}  ^{D>VP^}        ^{N4^}
 ^{J.D>^}
^{1X^}  ^{PDVq^}        ^{=m^}  ^{VPTB^}        ^{1R^}  ^{HKTB#X^}      ^{3$^}
 ^{PDHK^}
^{R2^}  ^{PDD>^}        ^{J@^}  ^{PDX%^}        ^{D?^}  ^{TBI=^}        ^{Cs^}
 ^{TBX%^}
^{J?^}  ^{J.R;^}        ^{9$^}  ^{R;VPR;^}      ^{JK^}  ^{HKMA^}        ^{U'^}
 ^{HKJ,^}
^{<8^}  ^{Vq9-^}        ^{6y^}  ^{VqI=^}        ^{3'^}  ^{R;Vq^}        ^{9c^}
 ^{27Vq^}
^{A&^}  ^{4sJ,^}        ^{>E^}  ^{4s9-^}        ^{P!^}  ^{9-=p^}        ^{PW^}
 ^{I=4s^}
^{35^}  ^{4sJV^}        ^{6+^}  ^{4sD>^}        ^{Ij^}  ^{VPLoVP^}      ^{Nˆ^}
 ^{R;4sI=^}
^{DK^}  ^{9-J,Vq^}      ^{<0^}  ^{9-9-HK^}      ^{SH^}  ^{8j4sI=^}      ^{8%^}
 ^{VPVP9-^}
^{U_^}  ^{J.4sHU^}      ^{4:^}  ^{JV4sHU^}      ^{FU^}  ^{X%=pHU^}      ^{Ji^}
 ^{8jD>J,^}
^{11^}  ^{VPR;PD^}      ^{?`^}  ^{X%J.?Z^}      ^{Ht^}  ^{X%4s?Z^}      ^{C+^}
 ^{VqJVI=^}
^{G.^}  ^{=p8jJ.^}      ^{W(^}  ^{JV9-8j^}
^{F,^}  ^{VPVPR;VP^}    ^{GI^}  ^{R;R;E.J,^}    ^{QG^}  ^{R;X%=p^}      ^{5\^}
 ^{=p9-VPVq^}
^{<R^}  ^{J.R;J,HK^}    ^{N*^}  ^{8j4sJ,8j^}    ^{Bm^}  ^{VPE.J,R;^}    ^{Dq^}
 ^{PD27J,R;^}
^{H4^}  ^{MA8jJ,VP^}    ^{=E^}  ^{TBMA8jVP^}    ^{<Y^}  ^{HK?Z27K.^}    ^{Cf^}
 ^{R;LoJ,VP^}
^{2X^}  ^{X%8jR;J,^}    ^{UW^}  ^{VPR;I=HK^}    ^{Jt^}  ^{J,VqVPTB^}    ^{JY^}
 ^{JV4sD>8j^}
^{9a^}  ^{LoJ.TBHK^}    ^{>w^}  ^{8jE.D>4s^}
^{4x^}  ^{X%J.TBVPTB^}  ^{7G^}  ^{VPR;J,27^}    ^{T5^}  ^{E.R;E.9-HK^}  ^{7l^}
 ^{E.R;27VqJ.^}

^{=aB[^}
^{#^#^^}
^{2Vr!JdHk7(2;?IN=MjC@NˆH1#,5+R2?IRTKc5CIOR;8v:CJdHk7(AK!#R*4&@m<8G'8v::WVJk7GRW
 JB!#Q'O04rWVCb2;AKR;7,?`9&#,5+1HFp51DjP!J1Q'P4WVDGR;1JR;;-5DPA?`#,J5TZKc2;AKJ2C
 4!#7QR;7,PD;z#,R;IzJ\SCNˆGn#,TYR2C;SP8|;.Kc5DM6WJAK!#K-=LNRCGIzN*::HKDX#?^}

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Mar  2 10:08:06 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA09142; Tue, 2 Mar 93 10:05:30 CST
Message-Id: <9303021605.AA09142@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2283; Tue, 02 Mar 93 11:00:51 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU
 (LMail V1.1d/1.7f) with BSMTP id 2970; Tue, 2 Mar 1993 11:00:48 -0500
Date:         Tue, 2 Mar 1993 14:57:42 GMT
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Chang Jie Table cjb.pcx
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

begin 644 cjb00001.pcx
M"@4!`0````!_`M<!@`+@`0```/___]($&`5C`&\U?TS/.V,,+`##%04`__]X
M`&<:M1UG&@&``8#5.PU,-@,QEP`!4``!````9QK5.Q%,)V1#"V<:T@L%!#&7
M,9<!`"),B`\!````M1VD`#&7.DSX$0$````!````9QIB"/X#``#!_W_![\'_
MP?Y_P>[!ˆ\'ˆ?___Q__!_C_!YP#![SW!Y@'!_GO__\?_P?Y_PN?!P`#!YG/!
MP`'__\?_P?R_@<'OC\'YP>9SP?Y____'_\'YP<_!YP#!W\'K@`/!X`/__\?_
MP?/"YSS!ˆ`?!YG/!_GW__\?_P>/!X<'#),'_P<_!QG.``/__Q__!R`#!_R3!
M_Y_!P@/!_'O__\?_N<'-@23!_SW!P%?!ˆ+'__\?_P?G!SYDDP<``I$G!\:?_
M_\?_P?G!SYDDP?\_IF/!X<'?___'_\'YCYDLP?\_ID?!R<'/___'_\'YP=N!
MP>/!_S_!YE.YP>/__\?_P?G!ˆ[W!R<'_/\'F,<'Y,/__Q__!ˆ`'!_YC!_#_!
MY'C!\,']___'_\'\`\'_?<'_?\'NP?W!ˆ___R/___]'____1_\'ˆ?\'NP?O"
M_\'W?\'``<7_O\+_P?O!W\'_/\'ˆ?\'_N_;_P>\]P>8!PO_!\Y_!S)G$_\'?
MG\+_P?G!S\'_/\'O/<'`&?;_P<``P>9SPO_!\\'/P<R9PO_![\'[P<ˆ;PO_!
MX`/!_S_!P`#!\G/V_X_!ˆ<'F<\+_P>?!V\'``<+_P>`!P>8!PO_!ˆ<'/P?\]
MC\'YP?)W]O_!W\'K@`/"_\'L`<'ˆ?\+_P>?!\\'OD\+_P?@/P<``P=_!Z\'R
M9_;_P?@'P>9SPO_!Q\'_@`#"_\'GP?/!]Y/"_\'YP<_!_A_!ˆ`?!\EOV_\'_
MP<_!QG/"_\''P??!]G?"_\'GP?.#D\+_P?@/P?P/P?_!SX`9]O_!_Y_!P@/"
M_Z7!\\'P`\+_P>?!\\'GD\+_P?G!R\'\+\'_G\'R<_;_P?\]P<!7PO_!Y,'G
MP?/!Y\+_P>?!\\'D`,+_P<`!P?DGP?\]P?)G]O_!P`"D2<+_P>9GP?`'PO_!
MY\'SP>>?PO_!ˆ,'/P?LWP<``P?)-]O_!_SˆF8\+_P>9OP?/!Y\+_P>?!\\'G
M#\+_P?)GP?,SP?\_P?(\]O_!_SˆF1\+_P>;!S\'P!\+_P>?!\\'E)\+_P>9S
MP?<YP?\_P>9Y]O_!_S_!YE/"_\'GP=_!\\'GPO_!X`/!XG/"_\'(&,'L",'_
M/\'F<_;_P?\_P>8QPO_!Y[W!\`?"_\'GP?/!QG'"_[Y]P=\]P?\_P<YG]O_!
M_#_!Y'C"_\'@`,'SP>7"_\'OP??![,'XPO_!_G?!_S_!_#_!WL'/]O_!_W_!
M[L']PO_![\'_@`#$_\'[P?W"_\'@`\'_/\'_?[\_]O___]'____1____T?_!
M[\'[RO_![\'[Q/_![\'[QO_!W\'[]/_!X`'%_W_$_\'@`<3_P>`!QO_!P`'T
M_\'GP?/$_\'ˆ/\3_P>?!\\3_P>?!\\;_P<YS]/_!Y\'SQ/_!_K_$_\'GP?/$
M_\'GP?/&_\'.<_3_P>?!\\3_P?V?Q/_!Y\'SQ/_!Y\'SQO_!SG/T_\'GP?/$
M_\']G\3_P>?!\\3_P>?!\\;_P<YS]/_!X`/$_\'[P<_$_\'@`\3_P>`#PO_!
MW\']PO_!SG/T_\'GP?/$_\'[P<_$_\'GP?/$_\'GP?/"_\'``,+_P<`#]/_!
MY\'SQ/_!ˆ`_$_\'GP?/$_\'GP?/"_\'/P?G"_\'?P??T_\'GP?/$_\'SP<?$
M_\'GP?/$_\'GP?/"_\'/P?GX_\'GP?/$_\'WP>?$_\'GP?/$_\'GP?/"_\'`
M`?C_P>?!\\3_P??!Y\3_P>?!\\3_P>?!\\+_P<_!ˆ?C_P>`#Q/_!X\'#Q/_!
MX`/$_\'@`\+_P<_!ˆ?C_P>?!\\K_P>?!\\3_P>?!\\+_P<`!ˆ/_![\'WRO_!
M[\'WQ/_![\'WPO_!W\'[ˆ/___]'____1_\'WP?O*_\'WP?O$_\'WP?O*_\'O
MP?W#_\'GPO_!\ˆG_P?`!RO_!\`'$_\'P`<+_P??![\;_P<``PO_!_@/"_\'V
MZ?_"\\3_P>`/Q/_"\\3_PO/"_\'P!\;_C\'YPO_!X#_"_\'@?ˆC_PO/$_\'Q
MP<?$_\+SQ/_"\\+_P?/!S\+_P=_!_<+_P=_!ˆ\+_P?;!]\+_P>3I_\+SQ/_!
M\\'GQ/_"\\3_PO/"_\'SP<_"_\'``,;_P?-CPO_!S.G_P?`#Q/_!\\''Q/_!
M\`/$_\'P`\+_P?/!S\+_P<_!ˆ<;_P?D'PO_!T.G_PO/$_\'SC\3_PO/$_\+S
MPO_!\`_"_\'/P?G&_\'[P=_"_XGI_\+SQ/_!\#_$_\+SQ/_"\\+_P?/!S\+_
MP<_!ˆ<K_P>'I_\'P`\3_P?./Q/_!\`/$_\'P`\+_P?/!S\+_P<_!ˆ<K_P?/I
M_\+SQ/_!\\''Q/_"\\3_PO/"_\'P#\+_P<_!ˆ<K_P>?I_\+SQ/_!\\'GQ/_"
M\\3_PO/"_\'SP<_"_\'/P?G*_\'/Z?_"\\3_P?/!Y\3_PO/$_\+SPO_!\\'/
MPO_!S\'YROˆ_Z?_!Y\'SQ/_!\<'/Q/_!Y\'SQ/_!Y\'SPO_!\\'/PO_!S\'Y
M]/_!Y\'SQ/_!X!_$_\'GP?/$_\'GP?/"_\'SP<_"_\'/P?GT_\'/P</*_\'/
MP</$_\'/P</&_\'/P>'T_[_!]\K_O\'WQ/ˆ_P??&_\'?P?OT____T?___]'_
MP?\_R_\_Q?\_QO_!_<'?]/_!_G_*_\'ˆ?\3_P?Y_QO_!ˆ,'/QO_!W\'W[/_!
M_C_$_\'\'\3_P?X_Q/_!_C_&_\'YP<_&_\'OP>/L_\'\G\3_P?G!Q\3_P?R?
MQ/_!_)_"_\'YG\+_P?G!S\;_PN?L_\'YP<_$_\'SP>?$_\'YP<_$_\'YP<_"
M_\'YG\+_P?/!Y\;_P?/!SˆS_P?/!X\3_P>?!]\3_P?/!X\3_P?/!X\+_P?F?
MPO_!\\'CQO_!][_L_\'GP?#$_\'GQ?_!Y\'PQ/_!Y\'PPO_!ˆ9_"_\'GP?'T
M_Y`)Q/_!Y\7_D`G$_Y`)PO_!ˆ9_"_\'/P?C"_\']O_#_P?Y_Q/_!Y\7_P?Y_
MQ/_!_G_"_\'SG\+_O\']PO_!ˆ,'?\/_!_F_$_\'GQ?_!_F_$_\'ˆ;\+_P?.#
MQO_!ˆ<'/\/_!X`?$_\'GQ?_!X`?$_\'@!\+_P??!Q\;_P?G![_#_P?9OQ/_!
M\\'WQ/_!]F_$_\'V;\+_P>_'_\'SP>?P_\'R9\3_P?G!S\3_P?)GQ/_!\F?*
M_\+S\/_!ˆ$_$_\'\'\3_P?A/Q/_!ˆ$_*_\'GP?'P_\'Z&\K_P?H;Q/_!ˆAO*
M_\'OP?GP_\'``<K_P<`!Q/_!P`'*_\'?P?OP____T?___]'_P?[+_\'ˆQ?_!
M_L3_O\+_P?[U_\'ˆ?\K_P?Y_Q/_!_G_#_X_"_\'ˆ?_3_P?Y_Q/_!X!_$_\'ˆ
M?\3_P?Y_P_ˆ?PO_!_G_T_\'ˆ>\3_P?'!S\3_P?Y[Q/_!_GO#_YO"_\'ˆ?\+_
MP?SQ_\'``<3_P?/!Y\3_P<`!Q/_!P`'"_X`!PO_!_F?"_\'\\?_!ˆC_$_\+S
MQ/_!ˆC_$_\'Z/\/_G\+_P?X#PO_!_,'O\/_!ˆ!_$_\+SQ/_!ˆ!_$_\'X'\/_
MG\+_P?!GPO_!P`?P_\'X7\3_PO/$_\'X7\3_P?A?P_ˆ?POˆ&9\+_P?S!S_#_
MP?)/Q/_"\\3_P?)/Q/_!\D_#_Y_"_\'ˆ9\+_P?S!S_#_P?)GQ/_"\\3_P?)G
MQ/_!\F?#_Y_"_\'ˆ1\+_P?S!S_#_P>9CQ/_"\\3_P>9CQ/_!YF/#_Y_"_\'ˆ
M;\+_P?G!S_#_P>9QQ/_!\\'GQ/_!YG'$_\'F<</_G\+_P?[#_\'YP<_P_\'.
M>,3_P?'!S\3_P<YXQ/_!SGC#_Y_&_\'YP<_P_[Y]Q/_!X!_$_[Y]Q/ˆˆ?<+_
MP?Z?QO_!ˆ<'/\/_!_G_*_\'ˆ?\3_P?Y_P_\?ˆ/_!_G_*_\'ˆ?\3_P?Y_P_ˆ_
Mˆ/___]'____1_\'ˆR__!_L7_P?[#_\'?ˆ?_!_G_*_\'ˆ?\3_P?Y_PO_![\'[
MP__!]_3_P?Y_Q/_!X`?$_\'ˆ?\3_P?Y_PO_!Y`'"_\'@`_3_P?YWQ/_!\<'G
MQ/_!_G?$_\'ˆ=\+_P>?!\\+_P??!Y_3_P?ISQ/_!\\'WQ/_!ˆG/$_\'Z<\+_
MOL'SPO_!]\'G]/_!P"?$_\'SQ?_!P"?$_\'`)\+_P=[!\\+_P?O!S_3_P?@/
MQ/_!\\'?Q/_!ˆ`_$_\'X#\+_P<]GPO_!ˆ<'/]/_!ˆ%_$_\'P'\3_P?A?Q/_!
Mˆ%_"_\'-9\+_P?V?]/_!\D_$_\'SP=_$_\'R3\3_P?)/PO_!ˆR_"_\'\G_3_
MP?)OQ/_!\\7_P?)OQ/_!\F_"_\'SC\+_P?X_]/_!]F?$_\'SQ?_!]F?$_\'V
M9\+_IY_"_\'ˆ/_3_P>9SQ/_!\\'WQ/_!YG/$_\'F<\+_P<>/PO_!_)_T_\'.
M<<3_P?'!Y\3_P<YQQ/_!SG'"_\'/)\+_P?G!S_3_P=YXQ/_!X`?$_\'>>,3_
MP=YXPO_!QG'"_\'SP>/T_[A]ROˆX?<3_N'W"_\'DP?C"_\'GP?#T_\'ˆR__!
M_L7_P?[#_\'KP?W"_Y_!_?3____1____T?_!_LO_P?[%_\'ˆQ__!_F_#_W_P
M_\'ˆ?\K_P?Y_Q/_!_G_&_\'F9\/_/_#_P?YWQ/_!X`?$_\'ˆ=\3_P?YWQO_!
M\E_#_S_#_S_L_\'N<\3_P?'!Y\3_P>YSQ/_![G/&_\'X/\/_/\/_/ˆS_P>YC
MQ/_!\\'WQ/_![F/$_\'N8\;_P?I_PO_!ˆQ_"_\'ˆ?ˆS_P>9GQ/_!\\7_P>9G
MQ/_!YF?*_\'Y+\+_P?Y?[/_!QD_$_\'SP=_$_\'&3\3_P<9/RO_!\R?"_\'\
M#ˆS_P<X_Q/_!\!_$_\'./\3_P<X_RO_!\S/"_\'Y)ˆS_P>R_Q/_!\\'?Q/_!
M[+_$_\'LO\K_P><QPO_!\S/L_\'\G\3_P?/%_\'\G\3_P?R?RO_!YSC"_\'G
M,>S_P?G!W\3_P?/%_\'YP=_$_\'YP=_*_\'//,+_P<\X[/_!ˆ<'/Q/_!\\7_
MP?G!S\3_P?G!S\K_OSW"_[\][/_!\\'GQ/_!\\7_P?/!Y\3_P?/!Y\+_P>N_
MQ_\_P_\_PO_!Z[?H_\'GP?/$_\'AQ?_!Y\'SQ/_!Y\'SPO_!S9_'_S_#_S_"
M_\'-F>C_P<_!\,K_P<_!\,3_P<_!\,+_C,'/QO_!_#_#_S_"_XS!S.C_O\'Y
MROˆ_P?G$_[_!ˆ<+_P=W!W\?_?\/_/\+_PMWH____T?___]'_P?[+_\'ˆQ?_!
M_OW_P?Y_RO_!_G_$_\'ˆ?_S_P?Y_Q/_!_!_$_\'ˆ?\3_P?Y__/_!_G_$_\'Y
MP<?$_\'ˆ?\3_P?Y__/_!_G_$_\'SP>?$_\'ˆ?\3_P?Y__/_!_G?$_\'GP??$
M_\'ˆ=\3_P?YW_/_!X`/$_\'GQ?_!X`/$_\'@`_S_P?Y_Q/_!Y\7_P?Y_Q/_!
M_G_\_\'ˆ?\3_P>?%_\'ˆ?\3_P?Y__/_!_G_$_\'GP</$_\'ˆ?\3_P?Y__/_!
M_G_$_\+GQ/_!_G_$_\'ˆ?_S_P?Y_Q/_!\\'GQ/_!_G_$_\'ˆ?_S_P?Y_Q/_!
Mˆ<'GQ/_!_G_$_\'ˆ?_S_P?Y]Q/_!_`_$_\'ˆ?<3_P?Y]_/ˆ``,K_@`#$_X``
M_/___]'____1____T?_!]\'?RO_!ˆ\'WQ/_!]\'?PO_!]\'?P__!S\?_P<_L
M_\'CC\K_P?'!\\3_P>./PO_!\L'-PO_!_@?"_\'WP__!]@?L_\'GG\3_P>'!
MP\3_P>23Q/_!YY_"_\'@0,+_P?!_PO_!\\/_P?!_[/_!Y9W$_\'SP>?$_\'.
M0\3_P>6=PO_!Y9/&_\'SP__!\ˆW_P<"`Q/_!\\'GQ/ˆˆP=/$_\'`@,+_P<RY
MQO_!\\/_P?/M_\'3,\3_P?/!Y\3_P<"SQ/_!TS/"_[]_QO_!\\/_P?/M_[+!
M\\3_P?/!Y\3_P?.3Q/ˆRP?/*_\'SP__!\ˆW_PO/$_\'P!\3_P?/!T\3_PO/*
M_\'SP__!\ˆW_PO/$_\'SP>?$_X!PQ/_"\\K_P?/#_\'S[?_"\\3_P?/!Y\3_
MP>,#Q/_"\\K_P?/#_\'S[?_"\\3_P?/!Y\3_P>'!\\3_PO/*_\'GP__!YˆW_
MPO/$_\'SP>?$_\'`P?/$_\+SRO_!Y\/_P>?M_\+SQ/_!\\'GQ/_!TL'SQ/_"
M\\K_P<_#_\'/[?_"\\3_P>'!P\3_L\'SQ/_"\\K_P=_#_\'?[?_!\\''RO_!
MX\'SQ/_!\\''ROˆ_P_ˆ_[?_!]\'ORO_"]\3_P??![_S____1____T?_!_V_*
M_\'ˆ?\7_;\+_P?W#_\'ˆ?\+_P?WQ_\'_)\K_P?YOQ?\GPO_!_L3_/\+_P?CQ
M_\'_)\3_P?P_Q/_!_@?%_R?"_\'ˆ?\+_P>\[PO_!ˆ<'O\/_!_SO$_\'ˆ?\3_
MP?Y_Q?\[P_\_PO_!X`'"_\'SP>?P_\'_`<3_P?Y_Q/_!]G?%_P'#_W_"_\'G
MP__![\'#\/_!X#_$_\'ˆ?\3_P?`#Q/_!X#_&_\'GP__!X!OP_\'_.\3_P?Y_
MQ/_!\\'GQ?\[QO_!Y_7_P?\YQ/_!_G_$_\'SP>?%_SG&_\'G]?_!_Q/$_\'ˆ
M?\3_P?/!Y\7_$\;_P>?U_\'_A\3_P?Y_Q/_!\`?%_X?&_\'G]?_!_X_$_\'ˆ
M?\3_P??![\7_C\;_P>?U_\'_#\3_P?Y_Q/_![;?%_P_&_\'G]?_!_DW$_\'ˆ
M?\3_P>2;Q/_!_DW&_\'/]?_!_,'EQ/_!_#_$_\'.2<3_P?S!Y<;_P<_U_\'S
MP>#*_\'>3,3_P?/!X,;_P=_U_\'_P?G*_[[!W<7_P?G&_[_U____T?___]'_
MP?]_RO_!_L;_?\+_P?WY_\'_/\K_P?Y[Q?\_PO_!_G_X_\'_/\3_P?X?Q/_!
MP`'%_S_"_\'N??C_P?\_Q?\_Q/_!ˆ\'?Q?\_PO_!P`#X_\'_/\7_/\3_P?G!
M[\7_/\+_C\'Yˆ/_!_SW%_S_$_\'SP>?%_SW"_\'?P?OX_\'``,7_/\3_P>?!
MT\3_P<``P__!]_C_P?\_Q?\_Q/_!Z\'!Q?\__/_!_S_%_S_$_\'=F\7_/_S_
MP?\_Q?\_Q/_!_)_%_S_\_\'_/\7_/\3_P?X_Q?\__/_!_S_$_\'S/\7_/\7_
M/_S_P?\_Q/_!\#_$_\'ˆ'\7_/_S_P?\_Q/_!_'_$_\'\P<?%_S_\_\'_/\K_
MP?/!X,7_/_S_P?]_ROˆ/P?G%_W_\____T?___]'_P?[1_\'ˆQ/\_PO_!_?7_
MP?Y_R__!ˆ\3_P?Y_PO_!]YW"_\'\P?W#_\'W\/_!_G_$_\'AA\3_P?`!Q/_!
M_G_"_\'P`,+_@`#"_\'OP?/P_\'ˆ?\3_P?/!S\7_P?/$_\'ˆ?\+_L\/_P?G#
M_\'OP?/P_\'ˆ>\3_P?.?Q/_!]S/$_\'ˆ>\+_P</#_\'SP__!]\'G\/_!P`'$
M_\'S/\3_P?N3Q/_!P`'"_\'CP__!Y\/_P??!Y_#_P?Y_Q/_!\G_$_\'YM\3_
MP?Y_PO_!\\/_P<_#_\'[P>?P_\'ˆ/\3_P?!_Q/_!_<'GQ/_!_C_"_\'CP_ˆ_
MP__!ˆ\'/\/_!_+_$_\'R/\3_P?S!S\3_P?R_PO_!P\?_P?W!S_#_P?R?Q/_!
M\Q_$_\'ˆ'\3_P?R?POˆ3Q__!_)_P_\'YP=_$_\'SC\7_/\3_P?G!W\+_P=/'
M_\'ˆ/_#_P?G!S\3_P?/!Q\3_P?X?Q/_!ˆ<'/PO_!Y\?_P?X_\/_!\\'GQ/_!
M\\'GQ/_!_,''Q/_!\\'GPO_!Y\?_P?R?\/_!Y\'QQ/_!X<'#Q/_!ˆ<'CQ/_!
MY\'QPO_!S\?_P?G!Q_#_P<_!ˆ,K_P?/!\,3_P<_!ˆ,+_P=_'_\'GP>#P_[_!
M_<K_P>_!ˆ<3_O\']POˆ_Q_ˆ?P?GP____T?___]'_P?[+_\+OQ/_!_L/_P?[#
M_\'OQ__!Y\/_P?[$_\'OY/_!_G_*_\'FP>?$_\'ˆ?\+_P?Y_PO_!Y\?_P?/#
M_\'ˆ?\+_P>`'Y/_!_G_$_\'AQ?_![F?$_\'ˆ?\+_P?Y_PO_!Y\?_P?7#_\'ˆ
M=\/_P<_D_\'>>\3_P?/%_\'.9\3_P=Y[PO_!_G_"_\'GQ_ˆ`P__!X`/#_\'/
MY/_!P`'$_\'SQ?_!UF?$_\'``<+_P?Y_PO_!Y\/_P?Y_PO_!ˆ</_P?YEP__!
MSˆ3_P<YSQ/_!\\7_LF?$_\'.<\+_P?Y_PO_!Y\/_P?Y_PO_!\L/_@`##_\'/
MY/_!SG/$_\'SQ?ˆ&9\3_P<YSPO_!_G_"_\'GP__!_G_"_\'B?\+_P?YGP__!
MSˆ3_P<YSQ/_!\\7_P>YGQ/_!SG/"_\'ˆ?\+_P>?#_\'ˆ?\+_P>##_\'@!\/_
MP<_D_\'.<\3_P?/%_\'>9\3_P<YSPO_!_G_"_\'GP__!_G_"_\'!P__!_F_H
M_\'``\3_P?/%_X8CQ/_!P`/"_\'ˆ?\+_P>?#_\'ˆ?\+_P=)_PO_!_G_H_\'>
M=\3_P?/%_\'<B\3_P=YWPO_!_G_"_\'GP__!_G_"_[+M_\'ˆ?\3_P?/%_\'X
MB\3_P?Y_PO_!_G_"_\'GQ__!\ˆW_P?Y_Q/_!\\'WQ/_!Y<'9Q/_!_G_"_\'ˆ
M?\+_P>?'_\'S[?_!_G_$_\'@!\3_B9G$_\'ˆ?\+_P?Y_PO_!S\?_P?/M_\'ˆ
M?\K_P=N\Q/_!_G_"_\'ˆ?\+_P=_'_\'S[?_!_LO_P?=]Q/_!_L/_P?[#_[_'
M_\'W[?___]'____1_\S_P>]WS?_!_<+_P>_!_?#_S/_!YS/(_\'7P_ˆ``,+_
MP>``P__!]ˆS_S/_!Y@',_\'ˆ?\+_P>?#_\'``ˆS_QO_!Q\'QQ/_!YS/,_\'\
MP__!Y\/_P?Y_[/_&_\+CQ/ˆ!,\S_P?W#_\'GP__!_G_L_\;_P>G!T\3_P>0`
MS/_!ˆ</_P>?#_\'ˆ?ˆS_P?_!ˆ\3_P>G!T\3_P<;!_<7_P?O*_\'GP__!_G_L
M_\'``<3_P>S!T\3_P<(`Q/_!P`'#_\'OQO_!Y\/_P?Y_[/_&_\'LL\3_P<!)
MR?ˆ?QO_!Y\/_P?Y_[/_&_\'NL\3_I`'(_\'ˆ/\;_P>?#_\'ˆ?ˆS_QO_![C/$
M_Z9)R/_!ˆ'_&_\'GP__!_G_L_\;_P>]SQ/_!Y@'(_\'AQ__!Y\/_P?Y_[/_&
M_\'O<\3_P>=SR/ˆ'Q__!S\/_P?Y_[/_&_\''P>'$_\'G.<C_P=_'_\'/P__!
M_GOL_\S_P>9\T/_!W\/_@`'L_\S_P>W!_=#_O_'____1____T?_!_\'WRO_!
M_L;_P??"_[O#_\'WP__!_<3_P=_'_\'WP__!_>#_P>`#RO_!_'_$_\'@`\+_
M@</_P>/#_\'\P_ˆ`#\/_P>_"_\'@`\/_P?S@_\'_P>?$_\+CQ/_!_,;_P>?"
M_\'SP__!Y\'[PO_!ˆ<'OP_ˆ?PO_!X`?#_\'GP__!_.#_P?_!Y\3_P?'!]\3_
MP?G!_<7_P>?"_\'SP__!X`'"_\'X!\/_G\/_P<_#_\'/P__!_.#_P>_!Y\3_
MP?3!]\3_P?``Q/_![\'GPO_!\\/_P>_!\\+_P?G!S\/_G\/_G\/_G\/_P?S@
M_\'@!\3_P?3!]\3_P>?!ˆ<3_P>`'PO_!]\/_P=_!\\+_P?/!S\/_G\/_/\/_
M/\/_P?S@_\'GP>_$_\'V=\3_P<[!ˆ<3_P>?![\+_C\/_O\'OPO_!YY_#_Y_"
M_\'ˆ?\?_P?S@_\'GP?O$_\'V=\3_OGG$_\'GP?O"_\'WQ__!WY_#_Y_"_\'\
MR/_!_.#_P>`!Q/_!]S?$_\'\N<3_P>`!PO_!\\C_/\/_G\+_P?G(_\'\X/_!
M[\'SQ/_!]S?$_\'YF<3_P>_!\\+_P?/(_S_#_Y_"_\'SR/_!_.#_P?_!\\3_
MP?>7Q/_!\`G%_\'SPOˆ3Q__!_G_#_Y_"_\'GP?W'_\'\X/_!_\'SQ/_!]Y?$
M_\'[P=G%_\'SPO_!Q\?_P?S$_Y_"_\'/P?W'_\'\X/_!_\'SQ/_!]\''Q?_!
Mˆ<7_P?/*_\'YQ/ˆ=PO_!S\'\Q__!_.#_P?\CQ/_"X\7_P=/%_R/*_\'SQ/_!
MS<+_P<``Q__!_.#_P?ˆ'R__!X\7_A\K_P>?$_\'@PO_!X`''_\'@X/_!_\'/
MR__!]\7_P<_*_Y_$_\'QR__!ˆ>#____1____T?_!_LO_P?[%_\'ˆQ/\_PO_!
Mˆ\/_P?/$_[_L_\'ˆ/\K_P?X_Q/_!_C_"_\'ˆ?\+_P?'#_\'SP??#_Y_L_\'ˆ
M?\3_P?P?Q/_!_G_$_\'ˆ?\+_P?P_PO_!\\/_P>`#P_ˆ?P__!ˆˆC_P?Y_Q/_!
Mˆ<'/Q/_!_G_$_\'ˆ?\+_P?F?PO_!Y\/_P>?$_Y_#_\'YZ/_!_G_$_\'SP>?$
M_\'ˆ?\3_P?Y_PO_!\\''PO_!S\/_P<_$_Y_#_\'SZ/_!_G_$_\'GP?/$_\'ˆ
M?\3_P?Y_PO_!Y\'QPOˆGP_ˆ_Q/ˆ?P__!YˆC_P?R_Q/_!Y\'SQ/_!_+_$_\'\
MO\+_P<_!ˆ,+_P>?(_Y_#_\'OZ/_!_+_$_\'GP?/$_\'\O\3_P?R_POˆ_P?W"
M_\'GR/ˆ?P__!]ˆC_P?R?Q/_!Y\'SQ/_!_)_$_\'\G\;_P>?(_\'/P__!\ˆC_
MP?G!W\3_P>?!\\3_P?G!W\3_P?G!W\;_P>?(_\'/P__!ˆ>C_P?G!S\3_P>?!
M\\3_P?G!S\3_P?G!S\;_P>?(_\'/P__!ˆˆC_P?/!Y\3_P?/!Y\3_P?/!Y\3_
MP?/!Y\;_P>?(_\'G[/_!Y\'CQ/_!ˆ<'/Q/_!Y\'CQ/_!Y\'CQO_!Y\C_P>/L
M_\'OP?'$_\'\'\3_P>_!\<3_P>_!\<;_P>?(_\'Q[/_!W\'XRO_!W\'XQ/_!
MW\'XQO_!Y\C_P?CL_[_!_<K_O\']Q/ˆ_P?W&_\'GR/_!_>S____1____T?_!
M_LO_P?[%_\'ˆP__![\C_O\/_O\+_P?O'_\'ˆX?_!_W_+_W_%_W_"_\'GR/ˆ?
MP_ˆ/PO_!ˆ</_P=_#_\'\?ˆ#_P?\_Q/_!X!_%_S_%_S_"_\'GR/ˆ?P_ˆ?PO_!
Mˆ</_P<_#_\'\X?_!ˆS_$_\'QP<_$_\'[/\3_P?L_PO_!Y\C_G\/_G<+_P?G#
M_\'/P__!ˆ<']X/_!ˆ9?$_\'SP>?$_\'YE\3_P?F7PO_!X\C_G</_G,+_P?G#
M_\'-P__!\`#@_\'YN\3_P?/!Y\3_P?F[Q/_!ˆ;O"_\'!Q__!_`##_YG"_\'Y
MP?W"_\'`P__!Y\'YX/_!Z<'YQ/_!\\'GQ/_!Z<'YQ/_!Z<'YPOˆ$R/ˆ?P_ˆ3
MPOˆ``,+_P<_#_\'/P?G@_\'IP?G$_\'QP<_$_\'IP?G$_\'IP?G"_Z?#_\'ˆ
MQ/ˆ?P_ˆ'PO_!ˆ</_P<_#_[_!ˆ>#_P<G!_,3_P?`?Q/_!R<'\Q/_!R<'\PO_!
MY\/_P?Y_P_ˆ?P_ˆ/PO_!ˆ</_P<_$_\'YX/_!R<'\Q/_!\\7_P<G!_,3_P<G!
M_,+_P>?#_\'Z?\/_P<_#_Y_"_\'YP__!R,3_P?G@_XG!]<3_P?/%_XG!]<3_
MB<'UPO_!Y\/_P?@OP__!S\/_G\+_P?G#_X/$_\'YX/_!V<'WQ/_!\\7_P=G!
M]\3_P=G!]\+_P>?#_\'R!\/_P>?#_Y_"_\'YP__!W\3_P?G@_\'YP?/$_\'S
MQ?_!ˆ<'SQ/_!ˆ<'SPO_!Y\/_P>83P__!YL/_G<+_P?G!W\?_P?G@_\'X`\3_
MP>'%_\'X`\3_P?@#PO_!Y\/_P>Y3P__!\L/_G<+_P?D_Q__!Tˆ#_P?P'RO_!
M_`?$_\'\!\+_P>?#_\'<=\/_P?##_X#"_\'PR/_!Xˆ#_UO_![\/_P?[$_\'Y
MP__!P<+_P?O(_\'WX/___]'____1_\'_P??+_\'WQ?_!]\+_P>?#_\'ˆP__!
M_O'_P?\#Q/_!_!_%_P/%_P/"_\'GP__!_C_"_\'ˆ?_#_P?`?Q/_!ˆ<'/Q/_!
M\!_$_\'P'\+_P>?#_\'ˆ=\+_P?YW\/_!_S_$_\'SP>?%_S_%_S_"_\'GP__!
MP`/"_\'@`_#_P?\WQ/_!Y\'SQ?\WQ?\WPOˆ!P__!_G_"_\'ˆ?\/_/ˆS_P>`#
MQ/_!Y\'SQ/_!X`/$_\'@`\+_P>;#_\'ˆ?\+_P?Y_PO_!]S?L_\'_/\3_P>?!
M\\7_/\7_/\+_P>7#_\'ˆ;\+_P?Y[PO_!\`/L_\'_/\3_P>?!\\7_/\7_/\+_
MP>/#_\'@!\+_P<`!PO_!\S_L_\'_/<3_P>?!\\7_/<7_/<+_P>/#_\'ˆ?\+_
MP?Y_PO_!\S_L_X``Q/_!Y\'SQ/ˆ``,3_@`#"_\''P__!_G_"_\'\P__!\SWL
M_\'_/\3_P?/!X\7_/\7_/\+_A\/_P?Y_PO_!_,/_@`#L_\'_/\3_P?G!Q\7_
M/\7_/\+_A\?_P?G$_S_L_\'_/\3_P?P/Q?\_Q?\_POˆGQ__!\\3_/ˆS_P?\_
MQ?_!S\7_/\7_/\+_P>?'_\'GQ/\_[/_!_#_%_\'GQ/_!_#_$_\'\/\+_P<?'
M_\'/Q/\_[/_!_W_+_W_%_W_"_\'OQ_ˆ_Q/\_[/___]'____1____T?___]'_
MP>_!ˆ\3_P>`?Q/_![\'[Q/_![\'[_/_!X`'$_\'QP<_$_\'@`<3_P>`!_/_!
MY\'SQ/_!\\'GQ/_!Y\'SQ/_!Y\'S_/_!Y\'SQ/_!\\'GQ/_!Y\'SQ/_!Y\'S
M_/_!Y\'SQ/_!\\'GQ/_!Y\'SQ/_!Y\'S_/_!Y\'SQ/_!\<'/Q/_!Y\'SQ/_!
MY\'S_/_!Y\'SQ/_!\!_$_\'GP?/$_\'GP?/\_\'GP?/$_\'S/\3_P>?!\\3_
MP>?!\_S_P>?!\\3_P?.?Q/_!Y\'SQ/_!Y\'S_/_!Y\'SQ/_!\\'/Q/_!Y\'S
MQ/_!Y\'S_/_!X`/$_\'SP>?$_\'@`\3_P>`#_/_!Y\'SQ/_!X<'#Q/_!Y\'S
MQ/_!Y\'S_/_![\'WRO_![\'WQ/_![\'W_/___]'____1____T?_!Y\'[RO_!
M[\']Q/_!Y\'[P__!]\K_P=_!ˆˆS_P>`!RO_!X\'<Q/_!X`'"_\'``\/_P?W&
M_\'``>S_P>?!\\3_P?@7Q/_!X`S$_\'GP?/#_\'GP_ˆ`QO_!SˆW_P>?!\\3_
MP?/!Y\3_P<O!W,3_P>?!\\/_P>?#_\'YPO_!ˆ\/_P<_(_Y_D_\'GP?/$_\'G
MP??$_\'*P=3$_\'GP?/#_\'GP__!ˆ<+_P?G#_\'/R/ˆ?Y/_!Y\'SQ/_!Y\7_
MP=I4Q/_!Y\'SP__!Y\/_P?G"_\'\P__!S\/_P?@'P_ˆ;Y/_!X`/$_\'GQ?ˆ*
M5,3_P>`#PO_!P`?#_\'YPO_!_G_"_\'/P__!ˆ<3_@>3_P>?!ˆ\3_P?!_Q/_!
MRE3$_\'GP?O#_\'OP__!ˆ</_/\+_P<_#_\'X#\/_Gˆ3_P>?%_\'ˆ#\3_P<I4
MQ/_!Y\C_P?G#_S_"_\'/P__!ˆ<3_Gˆ3_P>?&_\''Q/_!RE3$_\'GR/_!ˆ</_
M/\+_P<_#_\'X#\/_G>3_P>?&_\'GQ/_!RE3$_\'GR/_!X</_/\+_P<_#_\'Y
MQ/ˆ`Y/_!Y\7_P>_!Y\3_P<I<Q/_!Y\C_P?O#_S_"_\'/P__!ˆ<3_Gˆ3_P>_%
M_\'CP<_$_\',O,3_P>_,_S_"_\'/P__!ˆ<3_Gˆ3_P<_%_\'H'\3_P<F<Q/_!
MS\S_/\+_P<_!_<?_Gˆ3_P=_+_\'+P<C$_\'?R__!_#_"_\'``,?_Gˆ3_O\O_
MP=?!W<3_O\O_P?Y_PO_!W\C_Gˆ3____1____T?_!]\'?RO_!\\'GQ/_!]\'?
MPO_!ˆ<'/PO_!ˆ[_"_\'SP>?&_\'YGˆC_P?/!S\K_P?C!S\3_P?/!S\+_P?G!
MS\+_P?F?PO_!ˆ,'/QO_!ˆ9OH_\'SP<_$_\'@!\3_P?V;Q/_!\\'/PO_!ˆ<'/
MPO_!ˆ9_"_\']F\;_@`'H_\'SP<_$_\'N=\3_P<`!Q/_!\\'/PO_!X`/"_\'Y
MG\+_P<`!PO_!ˆ<'/PO_!ˆ9_H_\'SP<W$_\'ˆ?\3_P?G!S\3_P?/!S<+_P?G!
MS\+_P?F?QO_!ˆ<'/PO_!ˆ9_#_\';Y/ˆ``,3_P?Y_Q/_!ˆ<'/Q/ˆ``,+_P?G!
MS\+_P?F?QO_!ˆ<'/Q_ˆ9Y/_!\\'/Q/_!_G_$_\'YP<W$_\'SP<_"_\'YP<_"
M_\'YG\;_P?G!S<?_F>3_P?/!S\3_P?Y_Q/ˆ``,3_P?/!S\+_P?G!S\+_P?F?
MQOˆ``,;_P?Z9?ˆ/_P?/!S\3_P?Y_Q/_!ˆ<'/Q/_!\\'/PO_!ˆ<'-PO_!ˆ9_&
M_\'YP<_&_\'\F3_C_\'SP<_$_\'ˆ?\3_P?G!S\3_P?/!S\+_@`#"_\'YG\;_
MP?G!S\;_P?R9/ˆ/_P?/!S\3_P?Y_Q/_!ˆ<'/Q/_!\\'/QO_!ˆ9_&_\'YP<_&
M_\'\&#_C_\'SP<_$_\'ˆ?\3_P?G!S\3_P?/!S\;_P?F?QO_!ˆ<'/Q_ˆ9Y/_!
M\\'/Q/_!_G_$_\'[P<_$_\'SP<_&_\'YG\;_P?O!S\?_F>3_P?/!S\3_P?P_
MQ/_!\\'/Q/_!\\'/QO_!ˆ9W&_\'SP<_'_YGD_\'P#\K_P??!S\3_P?`/QOˆ`
M`,;_P??!S\?_F>3_P?/!S\K_P>_!W\3_P?/!S\[_P>_!W\?_F>3_Z__!V>3_
M___1_\S_P?>____#_\'_?\K_P?,=Q?]_QO_!W\3_P=_P_\'_/\3_P>'!X\3_
MP>)`Q?\_QO_!S\3_P<_P_\'_/\3_P?/!]\3_P>3!S,7_/\+_G\'\PO_!S\3_
MP<_P_\'_/\3_P?/!]\3_PLS%_S_"_Y_!_,+_P<_$_\'/?ˆ__P?\_Q/_!\\'W
MQ/_!Q,',Q?\_POˆ?P?S"_\'/P__!ˆ\'/?ˆ__P=\]Q/_!\\'WQ/ˆDP<S$_\'?
M/<+_G\'\PO_!S\/_P?/!SS_O_\'//,3_P?/!]\3_P>3!S,3_P<\\POˆ?P?S"
M_\'/P__!\\'//ˆ__P<\\Q/_!\\'WQ/_!Y,',Q/_!SSS"_Y_!_,+_P<_#_\'S
MP<\_[__!SSS$_\'SP??$_\'DP<3$_\'//,+_G\'\PO_!S\/_P?``/ˆ__P<\\
MQ/_!\\'WQ/_!Y,'(Q/_!SSS"_Y_!_,+_P<_#_\'P`#_O_\'//,3_P?/!Y\3_
MP>0-Q/_!SSS"_Y_!_,+_P<_!_</_P<_P_\'//,3_P?G!S\3_P>1/Q/_!SSS"
M_Y_!_,+_P<_!_</_P<_P_\'``,3_P?P?Q/_!Y<'/Q/_!P`#"_Y_!_,+_P<_!
M_,/_G_#_P<_!_,K_P>?!S\3_P<_!_,+_@`#"_\'``,/_/_#_P=_!_<K_P>?!
MW\3_P=_!_<+_G\'\PO_!X`'"_\'ˆ?_#_WO_!_/'____1_\'ˆR__!]\']Q/_!
M_LO_P=_Q_\'ˆ?\K_P>(`Q/_!_G_*_\'/\?_!_G_$_\'#P>/$_\'GF<3_P?Y_
MRO_!S_'_P?S%_\'GP??$_\'/F<3_P?S$_\'GPO_!Y\/_P<_Q_\'\P?W$_\'G
MP??$_\';F<3_P?S!_</_P<_"_\'GP__!S_'_@`#$_\'GP>_$_[&9Q/ˆ``,/_
MG\+_P>?#_\'/\?_!ˆ<'/Q/_!\\'OQ/ˆ!F<3_P?G!S\+_P?X_PO_!Y\/_P<_Q
M_\'YP<_$_\'SP=_$_\'F`<3_P?G!S\+_P?S#_\'GP__!S\/_P>_M_\'SG\3_
MP?/!W\3_P<ˆ9Q/_!\Y_"_\'SP__!Y\/_P<_#_\'D[?_!\9_$_\'YO\3_@YG$
M_\'QG\+_P<`/PO_!Y\/_P<_#_\'F<ˆS_P?P_Q/_!ˆ;_$_Yˆ9Q/_!_#_"_\'O
MP__!Y\/_P<_#_\'G)ˆS_P?X?Q/_!_7_$_\'YF<3_P?X?QO_!Y\/_P<_#_\'G
MCˆS_P?R/Q/_!_'_$_\'CN<3_P?R/QO_!Y\/_P<[#_\'FCˆS_P?G!X\3_P?[%
M_X>YQ/_!ˆ<'CQO_!Y\'WPO_!P</_P>4C[/_!Y\'QRO_!SWG$_\'GP?'&_\'@
M`\+_A\/_P<+!\.S_G\'[RO_!_@#$_Y_!ˆ\;_P>?#_\'?P__![\'][/___]'_
M___1_\S_P?W)_\'?P?W"_\+W]/_!W\'[RO_!_G_$_\'?P?O"_\'``,+_P?`#
M]/_!P`'$_\'"$<3_P?X_Q/_!P`'"_\'/P?G"_\'SP>?T_\'.<\3_P><[Q?]]
MQ/_!SG/"_\'/P?G"_\'SP>?T_\'.<\3_P><[Q/ˆ``,3_P<YSPO_!S\'YPO_!
M\\'G]/_!SG/$_\'G.\3_P?S%_\'.<\+_P<_!ˆ<+_P?/!Y_3_P<YSQ/_!\Q?$
M_\'\P??$_\'.<\+_P<_!ˆ<+_P?/!Y_3_P<`#Q/_!\I?$_\'\`\3_P<`#PO_!
MS\'YPO_!\\'G]/_!SG/$_\'RE\3_P?S!Y\3_P<YSPO_!S\'YPO_!\\'G]/_!
MSG/$_\'RE\3_P?S!Y\3_P<YSPO_!S\'YPO_!\\'G]/_!SG/$_\'RE\3_P?G!
MY\3_P<YSPO_!S\'YPO_"Y_3_P<YSQ/_!ˆL'7Q/_!ˆ<'GQ/_!SG/"_\'/P?G"
M_\+G]/_!P`/$_\'YP<_$_\'SP>?$_\'``\+_P<_!ˆ<+_P<`!]/_!S\'SQ/_!
Mˆ\'?Q/_!\\'GQ/_!S\'SPO_!S\'YPO_![\'G]/_!W\'WRO_!YX_$_\'?P??"
M_\'``</_C_3_S/_"W\C_P=_!ˆ\/_P=_T____T?___]'_P?Y_RO_!_G_$_\'ˆ
M?\;_P?SU_\'ˆ?\K_P?Y_Q/_!_G_"_\'ˆP__!_C_&_\'?[?_!_G_$_\'#P>/$
M_\'ˆ?\3_P?Y_PO_!_G_#_SW&_\'/[?_!_G_$_\'GP??$_\'ˆ?\3_P?Y_PO_!
M_G_"_X``QO_!YˆW_P?Y_Q/_!X\'GQ/_!_G_$_\'ˆ?\+_P?Y_RO_![ˆW_P?Q_
MQ/_!\<'/Q/_!_'_$_\'\?\+_P?YORO_!]ˆW_P?X_Q/_!ˆ)_$_\'ˆ/\3_P?X_
MPO_!_@?*_X/M_\'ˆ'\3_P?P_Q/_!_A_$_\'ˆ'\+_P?Y_RO_!YˆW_P?Y'Q/_!
M_G_$_\'ˆ1\3_P?Y'PO_!_G_*_\'G[?_!_F/$_\'ˆ?\3_P?YCQ/_!_F/"_\'ˆ
M?\;_P?Y_PO_!YˆW_P?YQQ/_!_G_$_\'ˆ<<3_P?YQPO_!_G_'_Q_"_\'G[?_!
M_GO$_\'ˆ?\3_P?Y[Q/_!_GO"_\'ˆ?\?_O\+_P>?M_\'ˆ?\3_P?Y_Q/_!_G_$
M_\'ˆ?\+_P?Y_QO_!_</_P>/M_\'ˆ?\3_P?P_Q/_!_G_$_\'ˆ?\+_P?Y_QO_!
M_C_"_\')[?_!_G_*_\'ˆ?\3_P?Y_R_\?POˆ<`.S_P?Y_RO_!_G_$_\'ˆ?\O_
@O\+_P=X![/___]'____1____T?___]'____1____T?\`
`
end

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Mar 11 06:36:30 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA11834; Thu, 11 Mar 93 06:36:28 CST
Message-Id: <9303111236.AA11834@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7635; Thu, 11 Mar 93 07:31:02 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU
 (LMail V1.1d/1.7f) with BSMTP id 3441; Thu, 11 Mar 1993 07:31:00 -0500
Date:         Thu, 11 Mar 1993 11:21:56 GMT
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Tutorial: ChangJIe and WubiZiXing
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Upon popular request (about 30 people a day), I have put the tutorial for
WuBiZiXing and ChnagJie in

ftp site: ifcss.org
under the directory: /incoming
filename: input.zip

It includes these files:
1. README
2. XUEDAZI, introduction
3. CHANGJIE, ChangJie tutorial GB format
4. CJB.PCX, ChangJie Radical Table PCX format
5. WUBI, WuBiZiXing tutorial GB format
6. WBB.PCX, WuBiZiXing radical table PCX format

I suggest you print the Radical Tables so that you can refer as you learn.

Send in your comment.

Thank you.

P.S. Could anyone suggest which diretory is most appropriate to put this kind
of files ?

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Mar 12 16:56:51 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA14363; Fri, 12 Mar 93 16:54:39 CST
Message-Id: <9303122254.AA14363@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 5499; Fri, 12 Mar 93 17:49:00 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU
 (LMail V1.1d/1.7f) with BSMTP id 1317; Fri, 12 Mar 1993 17:48:56 -0500
Date:         Fri, 12 Mar 1993 17:38:47 -0500
Reply-To: Wang Wei <14744w19@CPS201.CPS.CMICH.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Wang Wei <14744w19@CPS201.CPS.CMICH.EDU>
Subject:      Chinese coding, input, etc.
Comments: To: ccnet-l <ccnet-l%uga.bitnet@cmuvm.csv.cmich.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Hi, fellow netters,

This is Wang Wei again. Here I'd like to post some of what  I
think about  Chinese  computing. A  couple  of weeks  ago,  I
commented on Mr  Xu's article on  Chinese software on  MS-DOS
using some  offending  words. I  got  flames from  Mr  Xu  of
course. I'd apologise here. But as a matter of fact, I  still
have some disagreement with what Mr Xu has stated in  several
of his articles, which has nothing to do with my personality.

As Mr Xu said, Chinese computing  started back from 40  years
ago. If the argument  could stand, we  could then extend  the
history much further back to Fu2Xi1Ba1Gua4 thousands of years
ago. That was close to binary coding and related to  Chinese,
too. Right? But I see no point in there. I suppose there must
be at least a  computer involved when  we talk about  Chinese
computing.

Actually I'm not thrilled to write these "old craps" and post
them to the net. And actually  I do not have any  "influence"
as a poor  student. I  was and  still am  stating the  truth,
which may not be so cheerful to us.

OK. I apologise to  whom ever I offended.  Let's get down  to
business. It's a pity that Mr Xu doesn't reply my request for
his first article. I  wanted to base my  writing on his  one,
since I don't  have much material  handy. Everything  written
down here are based on my personal experiences before I  came
out of China  in mid 1991  and the postings  appear on  CCNET
recently. Comments, flames, whatever are always welcome.
_____________________________________________________________

To my personal knowledge,  Chinese computing started back  in
mid or  late 1970's,  or maybe  a  little earlier,  on  mini-
computers. But  since it  was  difficult to  display  Chinese
characters lacking graphical display, practical or commercial
Chinese hardware and  software add-on  modules appear  around
1980 when micro-computers based on 8080/Z80, and then IBM  PC
and clones gradually becoming more and more available. Before
CCDOS got  so popular  on the  market,  there was  already  a
LianXiang Chinese video  card and some  sort of  modification
package for micros based on S-100 bus and 8080/Z80 CPUs  with
dumb-terminals. It was nothing comparing with today's Chinese
"systems" but pretty cool back then.

I would  say CCDOS  is quite  a landmark  in the  history  of
Chinese  computing  in   the  main  land   China.  Its   main
contribution is  that it  made  Chinese computing  so  easily
available (Although  maybe the  lack of  software  protection
contributed to its popularity,  too). The evolution of  CCDOS
and the  emergence of  many its  clones  show that  CCDOS  is
probably going to survive the market competition.


1. Chinese character coding

Besides the modified version of GB  coding scheme (or we  can
call it GB11 comparing to GB10 below) used in CCDOS and other
similar systems,  there are  other coding  schemes like  GB10
(Hi-bit of  hi-byte  set, of  lo-bayte  reset) on  the  Stone
typewriters,  EBCDIC   "spot-filling"  on   IBM   mainframes,
leading-escape-code,  escape-code-quoted,   and  A9I   coding
scheme for  some not-remembered  systems :-).  Many  systems,
such as AT&T  Unix, IBM  5550, and  WANG MVS  (Thanks to  Guo
Jin), use their property coding scheme to represent  Chinese.
I believe  there are  some more.  And now  there's the  zW/Hz
coding scheme, which is "pure 2-byte  GB" code as claimed  by
Mr Wei Yagui in his article in a recent issue of HXWZ.  Among
these GB11 is definitely the most popularly used one in  main
land China, maybe just like Big5 in Taiwan.

As I mentioned  in my comments  to Mr Xu's  article, using  a
specific  coding  scheme  does  not  automatically  make  the
software "compatible" to the  others. And as  in the case  of
Beijing Library system, and  other situations, where odd  (or
maybe even dead)  characters are often  used, GB  is way  not
sufficient. As I know, the Beijing Municipal Public  Security
Bureau had problems coping with the strange characters people
used in their  names. That  also happens  a lot  in the  news
reports of Xinhua News Agency. So many people kind of  modify
the GB  table to  fit their  own needs.  Many have  used  the
unoccupied rows to put the extra characters when they have to
deal with both simplified and traditional characters and  the
conversion between  them.  Many  others have  used  in  these
undefined rows to store user-defined characters (I guess this
is also what CCDOS does).

I believe  the  end users  like  the secretaries  in  China's
offices would never care  how the characters are  represented
inside the computers. What they see  is how they input  their
texts, how they arrange things on a page, and how they  print
them out nicely.

Sometimes the using of a coding  scheme is restricted by  the
hardware and software environment. For example, when  writing
a software  to support  Chinese in  Email,  zW/Hz is  a  good
choice for most of us. But  for AT&T, they'd rather stick  to
their EUC (?: Extended Unix Code or whatever).

I personally  see UNICODE/ISO  10646 as  very promising.  But
there's a  danger  that the  GB2312  is going  to  appear  in
UNICODE untouched (I  read an article  about UNICODE on  this
net months ago).  Although adopting  a new  coding system  is
definitely costly, but  GB2312 is  definitely not  sufficient
either.


2. CCDOS and other MSDOS add-ons

CCDOS and most  of its clones  are add-on  modules to  MSDOS.
Most likely such an add-on module is divided into three  sub-
modules: input,  display, and  print. Most  of these  add-ons
achieve what Mr Xu referred as "BIOS-level compatibility"  to
MSDOS, which means if a software running on top of the add-on
module use only DOS  and BIOS function  calls to do  keyboard
input, screen  display, and  printing, it  will have  Chinese
processing capability.

Some software  add-on packages  have  the ability  to  handle
direct video memory mapping, which Mr Xu referred as "DMA  or
register level"  compatibility.  But DMA  and  registers  are
terms used  to talk  about hardware.  I don't  consider  them
suitable here. And  the technique such  packages use are  not
any more "secrets".  The simplest  way to  do it  is using  a
dual-buffer video  card, like  Hercules mono  or VGA,  and  a
timer  interrupt  (or  better  memory  fault  exception  with
286/386/486 in protected mode? I don't know). You can "cheat"
on BIOS and the user software to use one of the buffers while
setup the hardware to use the other one. The timer  interrupt
can be used to update the screen. (I haven't done this, but I
guess that's the way it works.)

Pure software packages normally  work as TSR  (Terminate-and-
Stay-Resident) programs in  DOS. They are  started after  DOS
booted. Then they change the  interrupt vectors, INT 10H  for
video display, INT 16H (or even INT 09H) for keyboard  input,
and INT 17H for printer, and then stay in memory. So  there's
another point here, when  you need to  write a program  which
must work with CCDOS and handle  INT 16H itself, you need  to
run your program  AFTER the  CCDOS stuff  rather than  before
them. You may take over INT 16H in your program, and route an
unwanted call  back  to the  old  INT  16H in  your  INT  16H
handler. It may not work totally flawlessly. But you may  not
get any control  of INT 16H  at all the  other way  as Mr  Xu
stated in his first article.

There are many hardware  supported MSDOS add-on packages  now
in the main land China market. But the most remarkable one is
the LianXiang Chinese video card  designed by Mr Ni  GuangNan
and his colleagues in the Research Institute of Computational
Technology, Academia Sinica, manufactured and distributed  by
the  Legend  Group  headquatered  in  ZhongGuanCun,  Beijing.
Another good  product back  in about  1984  was the  New  Era
Chinese video card from XiaoFeng.

The LianXiang card are remarkable  not only for its  hardware
design given the IC supply condition  in China, but also  for
its supporting software system.  The LianXiang card  realized
real register-level  compatibility  for both  text  mode  and
graphic mode  display at  the first  place.  So the  INT  10H
interrupt is  not changed  as with  CCDOS. The  card has  1MB
built-on RAM, which is  used to store  the 16x16 bit  display
font, input method data, and part of the 24x24 print font. So
it can  support both  simplified and  traditional  characters
easily. The user can switch between a number of input methods
in the fly, and  install any number of  input methods with  a
couple of DOS command lines. Since  the CPU is released  from
updating the  screen  and  the buffer  memory  is  relatively
large, its response to keyboard input and screen paging is as
fast as  non-Chinese systems.  Printing is  also much  faster
than other Chinese systems. All  these also made it  possible
to have disk-less Chinese workstations for PC networks.


3. Input methods

I don't agree with Mr Xu's judgement on PinYin input  method,
in which I mean that most of the Chinese key-in work is  done
by the professionals using  things like WuBiZiXing,  CangJie,
DaZhong, Eten,  etc.. PinYin  for them  is just  a  fallback.
PinYin may be  the most popular  one among  us the  so-called
"intellectuals"  who  can't  even  spell  the  right   PinYin
ourselves (Sorry for this comment ;-).  But how much work  do
we do? Several pages a week? That's just one or two hours for
a professional.

I'd     rather      agree      with     Mr      Jack      Sun
(GR4486%SIUCVMB.BITNET@uga.cc.uga.edu),   WBZX/CangJie    and
PinYin target to different groups of users.

I have a  question here:  How many of  us on  this net  could
input the  last  character in  GB  table easily  with  PinYin
(QuWei: 8794)? Huh? Just try it yourself.

That's one  point of  using the  writing-based input  methods
like  WBZX.  Most  of  the  professionals  would  "spell"   a
character out using WBZX or  whatever without knowing how  to
pronounce it.  And  besides, there's  no  point for  them  to
"think" about the text they type  while typing. And even  for
the people who  "think" while typing  (I remember reading  an
article in People's Daily about a Chinese writer who has a PC
and writes his novels on it), they remember things  gradually
while using a method and become spontaneous when typing.

Comparing to so many writing-based input methods available on
the market,  PinYin  has many  weak  points: there  are  many
homophones, even using word/phrase based input packages. This
makes PinYin extremely powerless when input  single-character
words. Many  have  praised the  sentence-based  PinYin  input
methods on this net. I don't know how many of them have  ever
tried to use  them to  input a Tang  Dynasty poem.  I bet  it
would make you  sweat out.  I know  that this  is an  extreme
example. But  the fact  is, there  are  many words  in  daily
spoken Chinese that are  single-character ones, such as:  Wo3
Ai4 Ni3 (I  love you),  Wo3 Xiang3  Qing3 Ni3  Bang1 Wo3  Ge4
Mang2 (I  want  to  ask  you  for  a  favour).  Every  single
character is an independent word in these two sentences. So I
may conclude that even the best "sentence-based" PinYin input
method  is  only  good  for  written  Chinese,  things   like
"Hong2tou2wen2jian4", not  good  enough for  the  writers  to
write novels with.

Another  point  is,  when  you  don't  know  how  to  read  a
character, you  have no  way to  input  it with  PinYin.  And
further more, there  are so many  people in  China who  can't
spell PinYin correctly (as some one on this net pointed out a
few days ago, there's  no 'n consonant,  there's only 'ng  in
Shanghai dialect).

But I still view PinYin as  the most promising input  method.
Not only because the younger  generation in China are  taught
with PinYin in  primary schools, not  only because there  are
just about 430  syllables (or roughly  1300 counting all  the
tones), but  also  because  it's  the  only  way  to  achieve
something  with  speech  recognition  (There  has  been  some
significant things  done  by Tsinghua  University  and  China
National Electronic Devices  Corp.). Besides  I also  believe
there're much  space  to  improve the  various  PinYin  input
methods we  are having  now (given  the example  provided  by
Haizi Xu  <USERSP45%UBCMTSG.BITNET@uga.cc.uga.edu> on  how  a
new word is added to XiaLiBaRen's dictionary: there are  more
than 10 steps  involved). The simplicity  of a  writing-based
input method may have  some score here:  it's so simple,  you
got nothing to complain about. :-)

In this specific case, I guess  the new word adding  approach
in PJS (a PinYin and phrase based input package developed  by
a Chinese  governmental research  institute and  somebody  in
Singapore) is a somewhat better one. Upon receiving an  input
string for some none-existent phrase, it'll try to break  the
string into single syllables and start to display the choices
for each of them. As  you finishing selecting characters  for
the phrase, it's added  into the dictionary  and sent to  the
application program which  is waiting for  input at the  same
time. I guess that's also what Mr Ed Lai's CiHui on Macintosh
does, but I haven't tried it, so no further words.

As to me personally, a perfect input package better have  two
input methods at  the same time,  by which I  mean I can  use
them both without switching back and  forth. I, as the  user,
can choose one  as my primary  input method,  and can  easily
fall  back  to  the  secondary  one  when  I  encounter   any
difficulty with the  primary one.  Things like  adding a  new
phrase into  the dictionary  should be  easy to  do.  There's
better some intelligence to  accept my not-so-correct  PinYin
input (as for 'n and 'ng,  c' and ch', etc.). Actually I  see
the problem with Chinese  input not any  more with the  input
methods themselves (given the fact  we have over 500  hundred
of them in main land China  alone), rather with the  software
packages that implement them (as Bo HU  <hubo@IFT.FHG.DBP.DE>
pointed out  in his  article dated  8 Mar.).  The  difference
between using a better package and a not-as-good one is  like
that between using MSDOS and MS Windows on the same PC.

To me, LianXiang is not a very good thing in a input  package
if it takes much effort to  integrate it. I agree with  Haizi
Xu again  on  this.  I  add  only  one  con  about  it:  it's
physically too tiring  for the eyes  when you  really use  it
cause you must scan the prompt line(s) to find what you want.

I also agree  with Haizi Xu  on separating  phrases on  their
scopes of usage. It's a great idea. But to implement it there
must be a  lot of  computational linguistics  involved. Am  I
right about  that? I  guess that  might also  be a  trend  of
future  Chinese  input  packages.  There  might  be   Chinese
"spelling-checker" or "grammar-checker" sort of things out on
the market in the near future.


4. Fonts

Just a few words. Mr Xu stated in his article that there  was
a law  prohibiting  the  existence  of  non-standard  Chinese
character coding schemes. I guess my friend Guo Jin is  right
about  this.  There  might  be  such  an  action  about  font
patterns. There might be  a reason for them  to do so. But  I
see it as merely  a commercial issue:  it's just because  the
government (or some specific  government agencies) has  spent
so much time and money to develop these font patterns (By the
way, I agree that they are really well done), they don't like
them to be wasted. They want some money back as compensation.
But I think such a law as unnecessary: (1) They can't ban all
the non-standard fonts, the others  may simple label them  as
"artistic"; (2) The  sense of sophistication  of fonts is  an
artistic term rather than a scientific one, which may develop
in time; (3)  A standard  should help  the manufacturers  and
vendors to develop better systems, rather than kick them  out
of the market.
_____________________________________________________________

Phew! I finally finished this  lengthy "old crap". I  promise
you I won't  do it again.  I tremblingly hope  this helps  my
"personality development".

Again, thanks for your patience  with this. I prefer  emailed
comments if it not interesting to anyone else.


Wang Wei

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Mar  3 21:26:06 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA05792; Wed, 3 Mar 93 21:26:00 CST
Message-Id: <9303040326.AA05792@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 0412; Wed, 03 Mar 93 22:21:17 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU
 (LMail V1.1d/1.7f) with BSMTP id 9124; Wed, 3 Mar 1993 22:21:14 -0500
Date:         Thu, 4 Mar 1993 11:23:10 +0800
Reply-To: guojin@ISS.NUS.SG
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: GUO Jin <guojin@ISS.NUS.SG>
Subject:      Pinyin, Bopomofo and MPS II
Comments: To: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>,
          Wang Wei <14744w19@CPS201.CPS.CMICH.EDU>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Just an info.  the Education Dept. of Taiwan also published in 80' a
"pinyin" named _Mandarin Phonetic Symbols II_ (MPS II).  They treat
Bopomofo as I (the first).  It's much similar to Pinyin adopted in
mainland China.  And, for an English educated poeple, MPS II is MUCH
BETTER than Pinyin.  For example:

        Pinyin          MPS II
        ======          ======
        j               ji
        x               shi
        q               chi

Pls also note that:  the ONLY difference among Bopomofo, Pinyin and
MPS II is thet way of using what symbol to represent a given
pronounciation.  The Mandarin phonological structure is independent
of any representation schemes.

And, among them, only Pinyin has been accepted in late 70' as an ISO
standard as the only mean to transcribe Chinese prop names (person,
place, organization, etc.).  Or you can say: Pinyin is the only
INTERNATIONAL STANDARD for ...

GuoJin

From @PUCC.PRINCETON.EDU:OWNER-CCNET-L@UGA.CC.UGA.EDU Fri Mar 19 01:36:30 1993
Return-Path: <@PUCC.PRINCETON.EDU:OWNER-CCNET-L@UGA.CC.UGA.EDU>
Received: from pucc.Princeton.EDU by ifcss.org (4.1/IFCSS-Mailer)
	id AA21123; Fri, 19 Mar 93 01:36:26 CST
Message-Id: <9303190736.AA21123@ifcss.org>
Received: from PUCC.PRINCETON.EDU by pucc.Princeton.EDU (IBM VM SMTP V2R2)
   with BSMTP id 6894; Fri, 19 Mar 93 02:34:20 EST
Received: from PUCC.BITNET (NJE origin LISTSERV@PUCC) by PUCC.PRINCETON.EDU (LMail V1.1d/1.7f) with BSMTP id 6830; Fri, 19 Mar 1993 02:34:20 -0500
Date:         Fri, 19 Mar 1993 02:32:36 -0500
Reply-To: Tan E Ming - guojin <emtan@ISS.NUS.SG>
Sender: Chinese Computing Network <CCNET-L%UGA.BITNET@pucc.Princeton.EDU>
From: Tan E Ming - guojin <emtan@ISS.NUS.SG>
Subject:      We are together
Comments: To: ccnet-l%uga.BITNET@nusvm.nus.sg
To: Multiple recipients of list CCNET-L <CCNET-L%UGA.BITNET@pucc.Princeton.EDU>
Status: RO

In fact what the man hunt fever should begin when there is a need for the public
to know that in Malaysia and Singapore also, there are chinese computing
research going on at these nations and some of them can change cause dramatic
changes to the current Chinese Information Processing. One of the activists in
field is Dr Chuah Chong-Cheng. He's a Chinese electronic dictionary designer
, input method designer/commentator and the lecturer himself. In March,1988 ,
Dr Foo Meng How of Kelantan, West Malaysia, has invited
Prof.Wang Yong Ming to the state to evaluate the WuBiZiXing "super-fast learning
" method devised by Dr Foo himself. And the inventor of the WBZX himself,
Prof.Wang has acknowledged that the method is far more effective than the
one he had suggested in his manual. Thus if anybody from Malaysia or Singapore
want to learn WBZX input method at a faster pace and effectively, this is the
opportunity. In fact, the WuBiZiXing group ( called "WBZX Yan2 Jiu1 Xiao3 Zu3 ),
led by Mr Gun It Gon and Dr Foo Meng How as the group adviser, has published
several booklet and learning manual of the input method and they all are
quite affordable. I think the most crucial point that they or the group has
done is, they have made all the materials/books on the input method a non-
technical "treasure".

It is a sarcasm if an input method devised and nobody is interested in knowing
it. At least in this context, Dr Foo has done his part in making the stroke-
based input method not so as hard to learn as what some might have claimed it
to be.

The fact that the stroke-based method so unpopular among the non-professional
user is partly due to the fact that there are not many of us who care about
making it easy to learn or use. Though the Pinyin based input method has a more
bright future, but if we haven't done anything to improve the status quo of
the stroke-based input methods, we also shouldn't say that they are of no
importance at all.

I wish to thank liangsheng (L.S.Ng1@southampton.ac.uk) for his moral support
for me to cantact local intelectuals to discuss and promote the Chinese
computing in the region.

At least, we Malaysian and Singaporean Chinese are not "sleeping" here and
just waiting for the perish or survival of the Hanzi and Chinese culture.

E.M.Tan

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Mar 24 12:14:57 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA10227; Wed, 24 Mar 93 12:14:42 CST
Message-Id: <9303241814.AA10227@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 3123; Wed, 24 Mar 93 13:11:14 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 0794; Wed, 24 Mar 1993 13:11:12 -0500
Date:         Wed, 24 Mar 1993 17:35:22 GMT
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Re: Radicals
Comments: To: TAN@CWU.BITNET
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <no.id>; from "george chang" at Mar 23, 93 3:07 pm
Status: RO

Hi G.Chang:

NJSTAR provides special input method for bushou. It uses V in Pinyin mode to
input special characters, as V is not normally used in Putonghua.

However if you want to include some undocumented bushou e.g. for creating a
Cangjie or Wubizixing table, you may have to export conventional characters to
a PCX file, then edit it with Paintbrush etc. This is also good if the receiver
does not have a Chinese Wordprocessor YET.

liangsheng
L.S.Ng1@mail.soton.ac.uk

P.S. Although most people in CCNET will surely have one Chinese WP, it would
be a good idea to promote Chinese Computing by sending mails in graphic format
to friends, which I did.

> Hi, net:
>
> Does anybody know any software that can bring radicals (bu4 shou3) to
> the screen?
>
> Thanks for teaching.
>
> G. Chang

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Mar 25 17:43:02 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA28844; Thu, 25 Mar 93 17:42:50 CST
Message-Id: <9303252342.AA28844@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 1031; Thu, 25 Mar 93 18:39:08 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 5510; Thu, 25 Mar 1993 18:39:06 -0500
Date:         Thu, 25 Mar 1993 15:20:52 GMT
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Re: Cangjie input method
Comments: To: Ruixian Bao <bao@cc.helsinki.fi>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <199303241859.AA27009@kruuna.helsinki.fi>; from "Ruixian Bao" at
              Mar 24, 93 8:59 pm
Status: RO

Ruixian wrote:
> I, however, have found out it that there are some Cangjie
> codes are not accepted when they are applied in Jianti Cangjie of KC. So
> a question to my mind: is the Cangjie coding a universal one or variable
> with different Cangjie progammes? Just like to know more info about the
> method.
>

I have noticed this problem for some time, but decided not to discuss it to
avoid confusion. I think now is the time to clarify, hopefully.

Cangjie was invented by Mr. Zhu Bangfu in Taiwan many years ago. It was
intended originally to code Fantizi. At least 5 Generations of Cangjie coding
had evolved ever since, and all these have been development for Fanti only.
The basic principle of Cangjie has not changed much, except there have been one
or two radicals added to Cangjie table. And in earlier Generations, there
wasn't a special method to deal with repeated code, but in 5th Generation, less
commonly used codes are preceded with X to distinguish from its commoner
counterpart, e.g. Song (A dynasty) as XJD, Wei (Not yet) as JD. I think Eten
is a good example of older generation of Cangjie, where X wasn't use for
repeated code.

As for Jianti Cangjie, some one in the net said that it was invented by people
outside Taiwan. And the coding of Jianti Cangjie is absurdly misleading. It was
blamed that the software developer did not follow certain Cangjie principles.
This version of JianCang survived in many software e.g. NJStar. To check if
your copy of Cangjie is this one I mention, check if HK is for Tian1 (sky),
which according to Fanti Cangjie should be MK.

The correction done to Jianti Cangjie is, again, by Mr. Zhu himself. This
'authorised' version of Cangjie was implemented, as far as I can recalled, in
a package called Ju4 Zhen1, where Cangjie is the only input method!! (Anybody
heard of it?) The coding is, of course, consistent with FanCang.

However, in the market place, the 'non-authorised' version of Jianti Cangjie
flourished, rasing much confusion. The solution I will offer, is 'DON'T USE
IT'. At the sametime, I have reproduced a new Jianti Cangjie Coding Table by
following as closely as possible to Ju4 Zhen1 JianCang, which I believe is the
authorised version. With this table, I made my customized input method using
NJStar. Of course, it can be used for other packages which allows customized
input method.

Whether or not Ju4 Zhen1 is the authorised Cangjie is not an important issue
after all. What is important is that the coding for Jiantizi should be
CONSISTENT with the coding for Fantizi. I first learned Cangjie using Fanti
with Eten. The only package that provided Cangjie in Jianti and still
consistent with Fanti (i.e. 'Original' Cangjie) is Ju4 Zhen1. However Ju4 Zhen1
itself had many problems. So I decided to 'port' the input method to other
'platforms' by reproducing the input method table - I listed all 7000 Chinese
Characters in Guobiao,  and typed the corresponding Cangjie code one by one
for a few days and nights.

Lastly, I would like to offer any software developers or end-users my own
Cangjie table, if you like, for your own use.

Thank you,
liangsheng
P.S. A note for software developer: Developing a input function using Cangjie
is very much easier than Pinyin because each character is coded UNIQUELY.

Other issues concerning Cangjie are welcome!!

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Mar 26 11:39:25 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA13641; Fri, 26 Mar 93 11:39:14 CST
Message-Id: <9303261739.AA13641@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7739; Fri, 26 Mar 93 12:35:58 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 4006; Fri, 26 Mar 1993 12:35:57 -0500
Date:         Fri, 26 Mar 1993 09:33:25 PST
Reply-To: Yongxin Yan <yxyan@SIGNUM.MV.US.ADOBE.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yongxin Yan <yxyan@SIGNUM.MV.US.ADOBE.COM>
Subject:      Re[2]: Cangjie input method
Comments: To: CCNET-L@uga.cc.uga.edu, L.S.Ng1@SOUTHAMPTON.AC.UK
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

As some of you know, I am not a great fan of WuBiZiXing or CangJie, or any
Xing2Ma3. But I do agree that these input methods have their place and should be
supported in a responsible software. In XLBR 3.0, you can have the bian1ma3 of
your choice. The method works this way, you provide a bianma table. One line for
each Chinese character, followed by its ascii coding string. At present, the
coding string cannot exceed 10 letters, and should contain letters only, i.e.,
no symbols or numerals. This covers WuBi and CangJie. The whole file should also
contain a one line header give the coding a Chinese name. In the system setup
file, you select that bianma file as your bianma input. You activate bianma
input by pressing F10 (or any other key or key-combination you put this function
on) or set bianma as default input.

XLBR 3.0 has been out there for a while, I wonder if anyone has used this
feature. If some one used it, I would like to get some feedback.


Yan Yong-Xin

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Mon Mar 29 19:37:04 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA05097; Mon, 29 Mar 93 19:36:49 CST
Message-Id: <9303300136.AA05097@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 0510; Mon, 29 Mar 93 20:32:53 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 3385; Mon, 29 Mar 1993 20:32:51 -0500
Date:         Mon, 29 Mar 1993 20:36:25 -0500
Reply-To: D-P Deng <deng@BUNNY.RHIC.BNL.GOV>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: D-P Deng <deng@BUNNY.RHIC.BNL.GOV>
Subject:      Re: problems with pinyin
Comments: To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  Your message of "Mon,
              29 Mar 93 15:38:00 EST." <9303292043.AA02190@bunny.rhic.bnl.gov>
Status: RO

Though I have difficulties with pinyin, pinyin input is still my exclusive
input method ;-). You see, I may have trouble with "z" and "zh", but I know
it's either "z" or "zh", so it's not too bad to find the right character,
except wasting a bit of time. A pinyin input should be FUZZIER AND
CONFIGURABLE to tune for personal taste. Fuzziness inevitably introduces a
larger set of choices of characters. To cut down the size of the set, a
phrase (cizu) based pinyin input can do the job well. I have played with the
method for a while, my weakness in speaking good putonghua has reduced
considerably.

I am an end user, I'm not going to argue which method is better, but to give
my experience (which probably was brought up before) as input to input method
developers.

1) input with and without a han-zi text next to you.
Without a han-zi text next to me, the sounds of characters pop up in the mind
first. It's kind of nature to type in pinyin. In addition, I don't want to
type in characters I don't know how to pronounce. That makes pin-yin even
more appealing.

With a han-zi text to follow, it's a different story. First, you look at the
characters, the sounds come to your mind after you see the shape of the
characters, it's kind slow to type in pin-yin. You know you are in trouble
when you see something you don't know how to pronounce it ;-).

2) input pin-yin first, then convert into han-zi.
It is probably just a personal preference, I found it fast and less
distracting.

3) cizu or not
It's ok to use cizu to write modern han-yu article, but cizu doesn't help much
when you want to type in gu-wen since gu-wen is mostly single character
based.

It seems to me that we need both a phonetic input method (like pin-yin) and
shape input method (cang-ji).

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Apr  1 10:45:53 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA13359; Thu, 1 Apr 93 10:45:47 CST
Message-Id: <9304011645.AA13359@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7546; Thu, 01 Apr 93 11:42:15 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 1694; Thu, 1 Apr 1993 11:42:11 -0500
Date:         Thu, 1 Apr 1993 16:36:25 BST
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Cangjie Table : XinCang
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Dear CCnet,

XinCang - Cangjie Table for Jiantizi by Liang Shing Ng @1993
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^*************
(Note square brackets [] denote Cangjie code)

This table is produced by Liang Shing Ng, based on Cangjie of
Fantizi and Cangjie of Jiantizi. Reference packages were Eten for
FanCang and JuZen for JianCang.

DISCLAIMER
-----------
NJINPUTX.DIC AND REGEN.BAS WILL BE KNOWN AS 'THE PACKAGE'
HEREAFTER.
SOME BUGS IN THE CANGJIE TABLE/REGEN.BAS SHOULD BE EXPECTED.
I DISCLAIM WARRANTY UPON ANY DAMAGE/LOSSES IN ANY FORM CAUSED BY
USER USING THIS PACKAGE.
-------------------------------------------------------------------

Differences from previous Cangjie
**********************************
I have made a few minor refinements to the Cangjie codes of the
above mentioned packages. I hope it is acceptable and I would like
to hear your feedback.

Here I list all the refinements made:

1. To solve for repeated code, X was used AFTER the normal code.
   e.g. wei4 (not yet) [JD] , sung4 (a dynasty) [JDX]
   In case there is a second repeated code use [XX] and so on for
successive repeated code. But this is rare.

2. Cang2yong4fu4he2zi4 ( Commonly used characters ) now include Ye4
(Page) whose CJ code used to be [MBC] in fancang and [MBO] in
jiancang, now it is [MO] when used as parts but [MBO] when alone.

3. Ge (elder brother) is now [MRDR] instead of [MRNR]

Reason for making such refinements:

1. User never realise if a character has a repeated code. If X
prefixes the code, user often have to retype. With X suffixes the
code, he/she will realise it when the code is type, by looking at
the characters displayed at the bottom of the screen, and it just
requires an extra X. Note that the feature of displaying character
before space-bar is pressed is not found in both Eten and JuZen.
NJSTAR has it and so does other good word-processor.

Limitation exists when the repeated code itself is five characters
long. In this case, user have to use Digit to select from a list.
However such list comprise normally 2 characters only with the
commoner character in the first place. User need not type 1 as
pressing space bar will suffice. Note that there are less than five
of such character-code out of nearly seven thousand GB character.

2. Conventionally, cangyong fuhezi includes Gui3 (ghost) [HI] and
[OG]. In fact Ye4 (page) [MBO] has a very high frequency and often
acts as bu4shou3.

3. This is more reasonable than [MRNR].

-------------------------------------------------------------------

The format of the Cangjie Table NJINPUTX.DIC and regenerating the
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
table in ASCII/GB format
************************
I formatted the Cangjie table using NJSTAR INPUTDIC. I uuencoded it
for posting. When you have received and uudecoded it, it is a
binary file of 27200 bytes long. This can be used directly for
NJSTAR. I used this format because it is the most compact one.
PKZIP can't achieve smaller size.

I also include a QBASIC program REGEN.BAS here to REGENERATE the
CangJie Table in ASCII/GB format. The regenerated table lists the
Cangjie code as lowercase A-Z in the first column and the
corresponding character in GB in the second column. Due to some
programming constraint, shorter code will have trailing ` (asc
value 96). You can replace it with white-space easily using any
word processor.
-------------------------------------------------------------------
BUGS REPORTING
--------------
As you are able to regenerate the Cangjie table, so if you find any
bugs in it, don't hesitate to correct it yourself and inform me if
you so wishes.
-------------------------------------------------------------------

The uuencoded canjie table will be posted seperately to those who
have made a request. Be warned, its size is about 40Kbytes.

I include REGEN.BAS here.

-----REGEN.BAS-----
DECLARE FUNCTION i$ (j&)
OPEN "njinputx.dic" FOR BINARY AS #1
OPEN "njinputx.gb" FOR OUTPUT AS #2

FOR h = &HB0 TO &HF7
FOR l = &HA1 TO &HFE
p = ((h - &HB0) * 94 + (l - &HA1)) * 4 + &H6D
GET #1, p, t&
PRINT #2, i$(t&) + " " + CHR$(h) + CHR$(l)
NEXT
NEXT

CLOSE

FUNCTION i$ (j&)
a& = j& \ 14348907: r& = j& MOD 14348907
b& = r& \ 531441: r& = r& MOD 531441
c& = r& \ 19683: r& = r& MOD 19683
d& = r& \ 729: r& = r& MOD 729
e& = r& \ 27
i$ = CHR$(a& + 96) + CHR$(b& + 96) + CHR$(c& + 96) + CHR$(d& + 96)
+ CHR$(e& + 96)
END FUNCTION

----END OF REGEN.BAS--------

Thank you
Liangshing Ng
31-3-93

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Apr  7 13:47:27 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA05002; Wed, 7 Apr 93 13:47:07 CDT
Message-Id: <9304071847.AA05002@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 4694; Wed, 07 Apr 93 14:41:36 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 5853; Wed, 7 Apr 1993 14:41:34 -0400
Date:         Wed, 7 Apr 1993 11:43:40 -0700
Reply-To: Ed Lai <lai@APPLE.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ed Lai <lai@APPLE.COM>
Subject:      Re: Input Method
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

This is an old discussion, but I was not around at that time.

Haizi Xu wrote

>Why can't we accomplish the same procedure in a more
>convenient way: you input the PinYin of your word but
>only find the machine displays a "wrong" word; you can
>simply move the cursor back to the word; the machine
>prompt all available words for the same PinYin; if the
>word you want is in the list, simply select the word
>and the machine will replace the "wrong" word with the
>right word for you; if the word is not in the list,
>press a function key, say F2 to add it to the
>dictionary;

Cihui on the Mac works likes this to some extent. When the
word/phrase choices are shown, all the possible first
characters are also shown. If the phrase is not in the
dictionary then you have to choose it from the characters,
followed by the choice of the second character in the
phrase etc, but after that the phrase would be entered into
the dictionary automatically (if that is your preference
setting).

Yuntao wrote

>It's said "If we can dream it, we can do it." Here is my dream:
> ........
>When I use Pinyin, I hope it is smart. So smart that I don't have
>to type in all the vowels. Instead, all I have to do is to type the
>first consonant of each character in a word, "CiHui". For example,
>if I want to input "Zhong1 Wen2", (chinese), I would type "zhw", and
>all the corresponding words that falls in this category will prompt
>up, then I will type the number of the one I want. If there two many
>choices, I would continue to type in the vowel of the last character
>in the word without having to start over again. In the above example,
>I will continue to type "e", ("zhwe", four key stokes now) or further
>still, "en" ("zhwen"). Hopefully, this will illiminate most choices,
>most probably leaving only one or two. If the character has no consonant,
>one would just type the first letter of the vowel, e.g., "An1 Quan2",
>one >would type "aq", or "aqu", "aqua"......
>If I could not find the word I wanted, I would go back to normal pinyin.

Again Cihui already works like this to some extent.

for zhong1wen2, all of the following are possible

zhong1wen2
zhongwen         ignore the tone
zhong1wen       ingore the tone in wen
zhongwen2       ignore the tone in zhong
z.w.      first character start with z, second start with w
zh.w.    first character start with zh, second start with w
zh.we.  first character start with zh, second start with we
z..        first character start with z, second can start with anything.

Of course they don't have the same result because other pharse other
than zhong1wen2 fits the critera z.w.

>choices, I would continue to type in the vowel of the last character
>in the word without having to start over again. In the above example,
>I will continue to type "e", ("zhwe", four key stokes now) or further
>still, "en" ("zhwen"). Hopefully, this will illiminate most choices,

Again Cihui works like this, before you hit the conversion key,
Cihui assumes that you have not finished typing, but nonetheless
it display the options based on what you have typed in so far.

So you type z
   Cihui displayes all characters and phrases that starts with z
You type h next
   Cihui displays all characters and phrases that start with zh
You type . next
   Cihui displays all characters and phrases that start with zh
You type w next
   Cihui displays all phrases where first char start with zh
    and second char start with w
You type e next
   Cihui displays all phrases where first char start with zh
    and second char start with we
You type n next
   Cihui displays all phrases where first char start with zh
    and second char start with wen

You can stop the process and pick the phrase anytime you
have type enough to get the phrase you want in the choice
box.

>one should be able to input a word with the average
>of three or four key strokes. Which
>is even faster than Wu5 Bi3 Zi4 Xing2.

If one is willing to use ShuangPin, then it takes an
average of 2 stroke (3 to include tone) in each
character even before abbreviation is used.

I think what this proves is that we spend too much effort
on inventing new input method and not enough on
implementing input method to its full potential. Cihui is
in the right direction, but a lot more can be done. And
CangJie is an example of an input method where there
is almost zero effort to implement it well, it could
have been make a lot easier and faster to use.

/* Disclaimer: All statments and opinions expressed are my own */
/* Edmund K. Lai                                               */
/* Apple Computer, MS37-AS                                     */
/* 20525 Mariani Ave,                                          */
/* Cupertino, CA 95014                                         */
/* (408)974-6272                                               */
zW@h9cOi

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Apr  8 07:40:29 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA13380; Thu, 8 Apr 93 07:40:23 CDT
Message-Id: <9304081240.AA13380@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7167; Thu, 08 Apr 93 08:36:15 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 0018; Thu, 8 Apr 1993 08:36:13 -0400
Date:         Thu, 8 Apr 1993 08:36:07 -0400
Reply-To: hubo@IFT.FHG.DBP.DE
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Bo HU <hubo@IFT.FHG.DBP.DE>
Organization: FHG
Subject:      Re: Input Method
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Also on PC there is a system called Munich System which can even more than
described in the forwarded messages below. Essentially, one can input a whole
sentence or several sentences in pinyin (for example: 512 characters), without
looking at the screen at all and than drive the cursor back to the words which
were not correctly interpreted by the system. The system marks all words of
which it is not sure. Adding, deleting, listing words and changing their
priorities are state of the art. The user can also strike a key to translate a
word back to pinyin for editing (typically "can" <-> "chang").

This system was initially developed by Prof. Heinzl (Technische Universitaet
Muenchen), Dr. Zhang Shuying ("Jiaotong DaXue"[?], Shanghai), Dr. Ding Tianhuai
(QingHua Daxue, Beijing), Dr. A. Lachner (resent address: Universitaet Bern)
and some students (Begin of the project was dated as early as 1979). I
re-implemented this for DOS systems (useable in combination with the most
character based DOS-programms) two years ago. A demo of another implementation
of this method is ftp-able under the name EasyWin (a stand-alone DOS
wordprocessor) from ifccs.org or ftp.lrz-muenchen.de (Europe).

Since GB-coded, stable Windows-systems are now available, it seems to be
interessting now to implement Munich System for Windows (MSW)...

------- Ed Lai wrote:

Date: 07 Apr 93 20:42:11+0200
From: Ed Lai <lai@APPLE.com>
Sender: Chinese Computing Network <CCNET-L@UGA.BITNET.DBP.de>
Reply-To: Ed Lai <lai@APPLE.com>
Subject: Re: Input Method
To: Multiple recipients of list CCNET-L <CCNET-L@UGA.BITNET.DBP.de>

This is an old discussion, but I was not around at that time.

Haizi Xu wrote

>Why can't we accomplish the same procedure in a more
>convenient way: you input the PinYin of your word but
>only find the machine displays a "wrong" word; you can
>simply move the cursor back to the word; the machine
>prompt all available words for the same PinYin; if the
>word you want is in the list, simply select the word
>and the machine will replace the "wrong" word with the
>right word for you; if the word is not in the list,
>press a function key, say F2 to add it to the
>dictionary;

Cihui on the Mac works likes this to some extent. When the
word/phrase choices are shown, all the possible first
characters are also shown. If the phrase is not in the
dictionary then you have to choose it from the characters,
followed by the choice of the second character in the
phrase etc, but after that the phrase would be entered into
the dictionary automatically (if that is your preference
setting).

Yuntao wrote

>It's said "If we can dream it, we can do it." Here is my dream:
> ........
>When I use Pinyin, I hope it is smart. So smart that I don't have
>to type in all the vowels. Instead, all I have to do is to type the
>first consonant of each character in a word, "CiHui". For example,
>if I want to input "Zhong1 Wen2", (chinese), I would type "zhw", and
>all the corresponding words that falls in this category will prompt
>up, then I will type the number of the one I want. If there two many
>choices, I would continue to type in the vowel of the last character
>in the word without having to start over again. In the above example,
>I will continue to type "e", ("zhwe", four key stokes now) or further
>still, "en" ("zhwen"). Hopefully, this will illiminate most choices,
>most probably leaving only one or two. If the character has no consonant,
>one would just type the first letter of the vowel, e.g., "An1 Quan2",
>one >would type "aq", or "aqu", "aqua"......
>If I could not find the word I wanted, I would go back to normal pinyin.

Again Cihui already works like this to some extent.

for zhong1wen2, all of the following are possible

zhong1wen2
zhongwen         ignore the tone
zhong1wen       ingore the tone in wen
zhongwen2       ignore the tone in zhong
z.w.      first character start with z, second start with w
zh.w.    first character start with zh, second start with w
zh.we.  first character start with zh, second start with we
z..        first character start with z, second can start with anything.

Of course they don't have the same result because other pharse other
than zhong1wen2 fits the critera z.w.

>choices, I would continue to type in the vowel of the last character
>in the word without having to start over again. In the above example,
>I will continue to type "e", ("zhwe", four key stokes now) or further
>still, "en" ("zhwen"). Hopefully, this will illiminate most choices,

Again Cihui works like this, before you hit the conversion key,
Cihui assumes that you have not finished typing, but nonetheless
it display the options based on what you have typed in so far.

So you type z
   Cihui displayes all characters and phrases that starts with z
You type h next
   Cihui displays all characters and phrases that start with zh
You type . next
   Cihui displays all characters and phrases that start with zh
You type w next
   Cihui displays all phrases where first char start with zh
    and second char start with w
You type e next
   Cihui displays all phrases where first char start with zh
    and second char start with we
You type n next
   Cihui displays all phrases where first char start with zh
    and second char start with wen

You can stop the process and pick the phrase anytime you
have type enough to get the phrase you want in the choice
box.

>one should be able to input a word with the average
>of three or four key strokes. Which
>is even faster than Wu5 Bi3 Zi4 Xing2.

If one is willing to use ShuangPin, then it takes an
average of 2 stroke (3 to include tone) in each
character even before abbreviation is used.

I think what this proves is that we spend too much effort
on inventing new input method and not enough on
implementing input method to its full potential. Cihui is
in the right direction, but a lot more can be done. And
CangJie is an example of an input method where there
is almost zero effort to implement it well, it could
have been make a lot easier and faster to use.

/* Disclaimer: All statments and opinions expressed are my own */
/* Edmund K. Lai                                               */
/* Apple Computer, MS37-AS                                     */
/* 20525 Mariani Ave,                                          */
/* Cupertino, CA 95014                                         */
/* (408)974-6272                                               */
zW@h9cOi


------- End of Forwarded Message

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu May  6 06:18:34 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA18083; Thu, 6 May 93 06:18:26 CDT
Message-Id: <9305061118.AA18083@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 3791; Thu, 06 May 93 07:17:20 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 7827; Thu, 6 May 1993 07:17:19 -0400
Date:         Thu, 6 May 1993 11:55:56 BST
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Re: Hanzi Cangjie strokes table. (fwd)
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Forwarded message:
>From lsn Thu May  6 11:51:15 1993
Date: Thu, 6 May 93 11:51:15 BST
To: deng@gov.bnl.rhic.bunny (D-P Deng)
Subject: Re: Hanzi Cangjie strokes table.
In-Reply-To: <9305051546.AA13924@bunny.rhic.bnl.gov>; from "D-P Deng" at May 5,
 93 11:46 am
X-Mailer: ELM [version 2.3 PL11]

>
> Hi.
>
> Do you have the table between han-zi and cang-jie strokes? I'd like to have
a
> copy if you do. Thanks.
>
> Dong-Ping Deng
>

I believe Mr Deng and other user as well are asking for a list of Chinese
Character and Cangjie code.

I have in my hand a GB Chinese Charater - Cangjie Code Table. The size of the
file is about 60Kbytes. I created this table one character by one character
myself. It includes all GB Chinese Characters which number to about 7000. It
is based on simplified Chinese Characters. I believe there has not been a
similar work as yet.

Additionally, I have written a tutorial on Cangjie Input method (published as
wb-cj-input.zip at ifcss.org/software/data). It consists of the rules of
Cangjie coding. I veiw this tutorial as the technical specification to the
product - the Cangjie Code Table and the table as the product to the
specification.

It is of use to those Chinese Wordprocessor that support external or user-
defined input method such as Xia Li Ba Ren and NJStar.

I am still not sure if it is legally healthy to publish it. (Although some
Chinese novels have been put in some ftp site)

Is it permissible to upload this table to ifcss.org or can anyone suggest a ftp
site?

Thanks
liangsheng wu

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu May  6 12:24:27 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA24983; Thu, 6 May 93 12:24:19 CDT
Message-Id: <9305061724.AA24983@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7985; Thu, 06 May 93 13:23:13 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 2770; Thu, 6 May 1993 13:23:11 -0400
Date:         Thu, 6 May 1993 17:27:36 BST
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Re: Cangjie Tutorial
Comments: To: Xiao Deng <xiao@scooter.amc.dec.com>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <9305061604.AA21147@scooter.amc.dec.com>; from "Xiao Deng" at May
              6, 93 9:04 am
Status: RO

>
> Hi,
>
> I am looking for a Cangjie input method tutorial in English version.
> Can you tell me where I can get it?
>
>
> Thank you
>

I have no idea. But I think it is not applicable since if one doesn't know
Chinese in the first place why should he use Cangjie?

....unless you need to train your French Secretary......

(just joking :-) please forget and forgive what I say)

Now seriously,

Some one once told me that her American secretary can type Cangjie though she
doesn't know much Chinese. (Are you reading this mail, please share your
experience again). It is due to the fact that you just need to recognize the
shape and stroke of a Chinese character and able to classify it then you can
use Cangjie to type Chinese. You don't need to know its pronunciation or
meaning.

Perhaps all of us needs some practice on input. The sentence oriented pinyin
input is a dream faraway.... I think to expect something like this is called
Hao4 Gao1 Wu4 Yuan3....

Anyway, will anyone like to translate the tutorial into English or French or
....? Thanks a lot in advance :-)

Cheers
liangsheng
Wubizixing and Cangjie Tutorial

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Mon May 17 09:09:19 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA00317; Mon, 17 May 93 09:09:16 CDT
Message-Id: <9305171409.AA00317@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9634; Mon, 17 May 93 10:08:20 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 7769; Mon, 17 May 1993 10:08:18 -0400
Date:         Mon, 17 May 1993 14:04:32 BST
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      A Chinese Keyboard Question
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

A Chinese Keyboard Question
---------------------------

A new keyboard layout for a Chinese Script had been designed and used for
sometime. It was found to have the following frequency distribution:

Key            Relative Frequency
m              4669
i              3632
h              3470
o              3346
l              2476
n              2453
y              2207
k              2092
b              1913
p              1840
d              1803
r              1788
a              1662
s              1497
g              1488
q              1481
u              1381
t              1319
v              1304
j              1224
e              1112
f              1041
w              906
c              840
x              188
z              2

'a' to 'z' denote the usual position of a standard QWERTY keyboard.
The numbers associated with each key is the relative frequency.

Problem: It can be seen that the layout was not ergonomically
optimized:
1) The left-hand and right-hand workload were not balanced.
2) The frequently-hit keys did not fall into the most comfortable
places.

A solution is therefore wanted to remap or redesign the keyboard
layout so that it has a optimized ergonomic design.

A notation is suggested when answering:

m->J
y->S
....
where lowercase letters denote the original positions and the
uppercase letter denote the position in the new design.

(This question was taken from a analysis on Chinese typewriting. It
had been simplified by taking away some unrelated technical and
linguistic details for the purpose of this question. The aim is to
present you a purely mathematical/design problem. Please feel free
to post to any newsgroup/mailing-list/interest group and reply to
Liangsheng Wu <L.S.Ng1@mail.soton.ac.uk>. This question was
originally put up to The Chinese Computing Network List CCNET-
L@UGA.BITNET)

(A more comprehensive report on the analysis will be posted to
CCNET-L soon.)

Thanks for your attention.

Sincerely
Liangsheng Wu
Southampton UK


From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue May 18 05:19:38 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA12777; Tue, 18 May 93 05:19:33 CDT
Message-Id: <9305181019.AA12777@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 5587; Tue, 18 May 93 06:18:49 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 1894; Tue, 18 May 1993 06:18:48 -0400
Date:         Tue, 18 May 1993 10:45:40 BST
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      A Chinese Keyboard Question (rephrased)
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

A Chinese Keyboard Question
---------------------------

A keyboard layout for a Chinese script has been designed and in used for some
time. It was found to have the following frequency distribution:

Key            Relative Frequency
m              4669
i              3632
h              3470
o              3346
l              2476
n              2453
y              2207
k              2092
b              1913
p              1840
d              1803
r              1788
a              1662
s              1497
g              1488
q              1481
u              1381
t              1319
v              1304
j              1224
e              1112
f              1041
w              906
c              840
x              188
z              2

'a' to 'z' denote the usual position of a standard QWERTY keyboard.
The numbers associated with each key is the relative frequency.

Problem: It can be seen that the layout was not ergonomically
optimized:
i) The left-hand and right-hand workload were not balanced.
ii) The frequently-hit keys did not fall into the most comfortable
places.
iii) plus other constraints which you can think of

A solution is therefore wanted to remap or redesign the keyboard
layout so that it has a optimized ergonomic design.

To rephrase the problem more formally, let's define some notation:

  lower case letters {a,..,z} denote the Chinese Alphabets
  upper case letters {A,..,Z} denote the usual position on a QWERTY Keyboard.

In the old design (as above), {a,..,z} are mapped to {A,..Z} such that
a->A, b->B, ...., c->Z
(read: the Chinese alphabet 'a' maps to position 'A' on the QWERTY keyboard...)

The new design is required such that {a,..z} -> {A,...,Z} such that
i) the most frequently hit key should fall on the most strategic place
ii) the left-hand and right-hand workload should be balanced
iii) plus other possible considerations
e.g.
m->J, i->F,...., z->Y
(read: the Chinese alphabet 'm' now maps to postion 'J' on a QWERTY
keyboard...)

Please send your comment and answer to Liangsheng Wu <L.S.Ng1@mail.soton.ac.uk>


Thanks for your attention.

Sincerely
Liangsheng Wu
Southampton UK

(This question was taken from a analysis on Chinese typewriting. It
had been simplified by taking away some unrelated technical and linguistic
details so as to present you a purely mathematical/design problem. Please feel
free to post to any newsgroup/mailing-list/interest group and reply to
Liangsheng Wu <L.S.Ng1@mail.soton.ac.uk>. This question was originally put up
to The Chinese Computing Network List CCNET-L@UGA.BITNET)

(A more comprehensive report on the analysis will be posted to
CCNET-L soon.)

Note: The word 'alphabet' means the elements in a script which can be any
script or the elements in any kind of representation e.g. In Digital Signal
Processing, the 'alphabets' are a High voltage and a Low voltage; any piece of
digital information is composed of a stream of High and Low voltages.

From CCNET-L@uga.cc.uga.edu Mon Jan 11 13:36:28 1993
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA28248; Mon, 11 Jan 93 13:36:13 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 1179; Mon, 11 Jan 93 13:34:03 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 4395; Mon, 11 Jan 93 13:34:01 EST
Date:         Sun, 10 Jan 1993 17:19:29 PST
Reply-To: Qing Zhang <qzhang@INDE.BC.CA>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Qing Zhang <qzhang@INDE.BC.CA>
Subject:      A new friendly HZ input method for cxterm
Comments: To: CCNET-L@UGA.UGA.EDU
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Hi, everybody,

If you are using cxterm and chinese, and are not satisfied with the input
method, you might be interested in this new HZ input method.

What is new?  The new feature is that you don't have to spell out PinYin
EXACTly in order to find words.  Sound great to thoes who are struggling hard
with PinYin?  Also, it support about 25,000 phrases, and allows you to define
your own phrases and symbols, all in a consistent style.

I call this method "MEPY" which stands for "Main Element PinYin" or in chinese
    \        ---      |   \  /      \
  -----     -----   --|- -----    -----
    |        | |      |   | |      \ /
  -----      | |      |/  | |     =====
    |       /  |    --|  -|-|-    |---|
  -----    /   L_     |   | |     |___|

and I had integrated it into cxterm X11 R5 in GB mode.  The main idea is that
it uses the main element of PinYin to identify character or phrases.  In the
following discussion, I use "HZ" for single Chinese Char, and "CW" for Chinese
Word or phrases.

1.      It support both HZ and CW mode. You will be switched to CW mode
        automatically by continuing typing in PinYin.

2.      The great thing about MEPY is its capability to tolerate wrong spelling
        of pinyin when you are looking for HZs or CWs. As many people had suffer
ed
        trying to get correct pinyin in order to find what they are looking for,
        MEPY does not require you to always type in correct pinyin, and it will
        tolerate many common mistakes, such as the confusion between "c" and "ch
",
         "r" and "l", "in" and "ing", "o" and "ou" and "uo", and so on.

        However, MEPY achieve this capability without sacrifying anything, if
        your pinyin is good, you will find MEPY is easy and convenient to use,
        if your pinyin is poor, you will be happy to find that you can still
        get the words you looking for, instead of frustrating by constant
        trying and error and trying.

        For example, if you are looking for a single HZ "zhang" and you typed
        in "zan", you can use a single key '>' or '<' to switch to "zang", "zhan
",
        and "zhang", each time, the corresponding pinyin will be prompted. But
        if you know the correct pinyin, you don't loose anything. Is this nicer
        than having to use Backspace to delete and try again?

3.      The main strength of MEPY is its ability to capture CW even though you
        don't type in correct Pinyin.  For example, if you want to find
        ChangJiang (long river) and you type in "canjan", you can still get
        what you want. The reason is that "changjiang" and "canjan" share the
        same ME Pinyin. This feature has two benefits: the first one is the
        ability to tolerate your wrong input; the second is to allow you to
        purposely use this feature to save some key strikings.

        Even great is that if you happened to type in "zhangjiang" to look for
        "ChangJiang" (long river),  it will still get the word.  Remember
        "Chang" (means long) has another pronounciation "Zhang".  There are
        many HZs which have more than one pronounciation, but under specific
        CWs, they have to be pronounced in one way or the other, but not both.
        In many occassions, people tend to make mistakes to get the correct
        prounciation, you can't blame people for these kinds of mistakes, you
        have to blame HZ.  In MEPY, it can tolerate these kinds of mistakes,
        therefore ease the user to find the CW they want.

4.      You can define your own CWs in a specific file, and define various
        symbols ( like punctuations, arithmetic symbol, numerical symbols )
        in another file. When MEPY is loaded, it will read these two files.
        Once loaded, you can use them in the same way as HZ and CW, the only
        difference is that, for user defined CW, you have to start with 'u',
        and for all symbols, you have to start with 'v', then you can input
        the key ( consisting of zero to four chars) you defined in the file
        to search the CW and symbols you want. With 4 char key, you can almost
        define everything you need, like name, technical term, commonly used
        phrases, even Greek characters, Japanese Hiragana, and so on. The
        advantage is that you don't need to switch input method to get them.

5.      Dynamically adjust the frequently used HZ and CW to make them easy to sp
ot.
        For CWs, it is simple, I just put the most recently selected CW at the
        first position. But for HZs, I decide not to do it that way, because
        many HZ entries have dozens of HZ to select, constant position change
        makes them randomly ordered, the result is that you have to examine each
        of them closely. Instead, I decided that each HZ entry has a fixed
        order, arranged from simple strokes to complex strokes (not very strickl
y),
        and those HZ which looks similar are clustered together to make them eas
y
        to find. Of course, those rarely used HZs are put at last, arranged in
        the same principle. In addition, I created some hot HZ caches to put
        those recently selected HZs, each starts with a vowel Pinyin letter.  So
        when you type in 'c', you will find all the recently selected HZ start
        with 'c'. The capacity of each cache depends on how many HZ you can
        displayed each time in the input area. The advantage is that you can
        easily find the frequently used HZ, at the same time, the order of
        regular HZ is not disturbed.

6.      MEPY data structure is very compact. With more than 7000 HZs and 25,000
        CWs, the data structure takes about 260K bytes momery. The whole module
        is totally portable, and ready to integrate into any word processor
        package. For example, I just spend a few hours to integrate it into
        cxterm.

I decided to put in into the public domain for every body to try it.  For
the moment, I only integrated into cxterm for X11 R5.0, there should be no
problem with X11 R4.0, but I haven't try it. Before putting it into public
domain, I would like to have some one to try it first to see if there is any
problem in interating.  If there is already a cxterm on your machine, and
you have the source code, all you need to do is to add two more source file
into the Imakefile, replace one source file, compile and run it.

Send me a mail if you want to try it.

The only "obligation" to use it is that you "have to" report any bugs to me
or places for improvement. For example, if you find some commonly used CWs are
missing, record them and send me a copy, I will put them into it in later
release, because a thorough data check is so time consuming that I can't afford
to do it myself.

#### NOTE ####

I do claim the copyright of this "MEPY" HZ input method and the related
source code and data.  You may not use this input method, souce code and
data for commercial purpose without written perssion from the author.

Qing Zhang,

qzhang@rivendell.uvic.ca
qzhang@inde.bc.ca

###############################################################################

From CCNET-L@uga.cc.uga.edu Tue Jan 26 17:54:26 1993
Received: from uga.cc.uga.edu by sirius.ctr.columbia.edu (5.67/5.10) with SMTP id AA01439; Tue, 26 Jan 93 17:53:48 -0500
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2605; Tue, 26 Jan 93 17:51:47 EST
Received: from UGA.BITNET by UGA.CC.UGA.EDU (Mailer R2.08 PTF008) with BSMTP id
 6188; Tue, 26 Jan 93 17:51:45 EST
Date:         Tue, 26 Jan 1993 12:55:50 PST
Reply-To: Qing Zhang <qzhang@INDE.BC.CA>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Qing Zhang <qzhang@INDE.BC.CA>
Subject:      MEPY for X11R4 is ready
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Hi, everybody,

After sending out the message about MEPY for cxterm X11R5 a few weeks ago,
quite a number of people have tried that, and the response are very good.

========================================================================
Enclosed in the user manual of MEPY.



~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                USER MANUAL OF MEPY HANZI INPUT METHOD

                   Copyright (C) 1993 by Qing Zhang
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Terminology

MEPY --- Main Element PinYin.
HZ   --- single chinese character.
CW   --- Chinese word or phrase.
UCW  --- user defined CW.


I. Definition of MEPY

In the current release, I defined the MEPY as follows:

Vowels:
        MEPY            non-MEPY
        a               ia, ai
        e               ie, ei
        o               ou, uo
        u               uu, iu, ui
        an              ang, ian, iang
        en              eng
        in              ing
        ong             iong
        ao              iao
        i               /* when not used with other vowels */

        some vowels may be decomposed into two parts, each has a MEPY
        corresponding to it. Like "uang", its MEPYs are "u" and "an".

Consonants:
        MEPY            non-MEPY
        z               zh
        c               ch
        s               sh
        l               r

        all other consonants are defined as MEPY.

The users can have their own preference to define MEPY, such as take
"ai", "ei", "ui", "uo" as MEPY.  There is a utility to automatically
convert user's preference into data, but this utility is not released
yet, because I am afraid some un-cautious user may do more damage than
good if not understanding the syntax clearly when change the preference.

II. Chinese Word (phrase) Input

For typing CW (Chinese Word or phrase), you just need to type in the
MEPYs of the first two HZs (Hanzi).  The advantages of using MEPY are
twofold, on the one hand, it is much easy to spell the MEPY than to
spell the complete Pinyin, on the other hand, MEPY are usually one or
two letters shorter than the non-MEPY, thus save some key striking.

EX: If you want word "ChangJiang" (means Long River), it is not difficult
to recognize the element "c", "an", "j", but some people may be uncertain
whether to use "c" or "ch", "an" or "ang".  In most PY input method,
if you use "c" in the place of "ch", you will never get the correct
word.  But in MEPY, it will tolerate your possible wrong spelling, and
give you the correct word.  In this example, if you type in "chanjian",
or "canjan" or "cangjang", or other combinations, it will always show
you the correct word you are looking for. The reason is that, all these
inputs share the same MEPYs, that is, "c", "an", "j", "an".  In fact,
you can make use of this feature to save some key strikings, and purposely
use "canjan" to look for "changjiang".

May be you will worry if MEPY will result in too many selection for each
entry.  Statistical results based on 28000 CWs show that, on average,
the number of selection for each entry is about 1.9 CWs, and 85% CWs are
in the entries with not more than 5 CWs to select, that means, they can
be displayed in a single screen in input area.

Moreover, dynamically adjusted CW list will put those frequently used CWs
at the beginning, making them easier to spot.  Also, for CWs with more
than two HZs, you can typing one more key to further reduce the selection
range, this additional key is the first PinYin letter of the third HZ.
This is useful when typing CWs like Sunday, Monday, ... Saturday. As you
know, in chinese, the first two HZs of these CWs are the same.

III.  Single Hanzi Input

Though the strength of MEPY is in CWs input, but in single HZ input, it
also has some nice feature not available in other PinYin method.  Because
of MEPY, we can link those similar sounded HZ entries into a circular
list, if you can't find the HZ you want, you can switch to the next entry
with one key stroke: '>' or '<' (upper case).

EX:  If you are looking for "zhang", but you typed in "zan", you can use
'>' or '<' to look for other entries with the same MEPY. That is, you can
switch to "zang", "zhan", and "zhang" with '>' or '<'.  Each time it
switches, the correct PinYin corresponding to the new entry will be
displayed to help you to remember.  So if you know the correct PinYin,
you don't loose anything, but if you don't, MEPY provides an easy way to
help you search. Is this nicer than having to use backspace to delete
input and try other choice again?

Note, in HZ input, the key '.' and ',' (lower case) are different to '>'
and '<' (upper case).  the key '.' and ',' are used to select right screen
and left screen of the same entry, while the key '>' and '<' are used to
select the next entry and previous entry in the same MEPY list (Different
entry has different PinYin).

When you type in a single consonant element, like 'c', 'm', you will find
some frequently used HZs which start with 'c' or 'm'.  I call these lists
Hot HZ Caches, one for each consonant MEPY. When you select a
HZ, this HZ will be put into this hot HZ cache, so next time, you just
need to type in one key to find this HZ.  The HZs in these caches will
be replaced if they are not used for a number of selections.  I decided
not to dynamically adjust the normal HZ lists because this will make the
lists randomly ordered. Also I decided not to use tone because I don't want
to impose more restriction on the users.  Instead, the normal HZ lists
are regularly ordered, arranging from simple strokes to complex strokes.
Those which look similar are ofter clustered together to make them easier
to spot.  The un-commonly used HZs are alwayes put at the last.  You will
be supprised to find that this arrangement make HZ selection much easier
than other methods, because you can roughly scan the display and know if
the HZ is in the display.

IV.  Symbol Input

All symbols input starts with the 'v' followed by a key which consists of
zero to four letters.  The key is defined in the text file "symbol.txt"
and can be edited by the users.  Read the comments in "symbol.txt" to
find out how to define symbols in your own way, but for most users, they
should be happy with the current definition.

Symbol input are basically the same as HZ input, except you have to
start with 'v'. Try it and you will like it.  One thing special is that
you can use some punctuation key as input. For example, if you type in
"v:", you will get a ':' in HZ; if you type in "v[", you will get all
different shapes of left brackets, and type in "v]" to get all right
brackets.  So you don't need to remember these key at all.  Try other
key like "va", "vm", and see what will be displayed.

If you are using Japanese, you can define "vjka", "vjki", "vjku", "vjke",
"vjko" for Japanese "Ka", "Ki", "Ku", "ke", "ko", and so on.  If you are
not using them, you can comment out them completely, or put them into a
single key entry "vj".

Note, each key in "symbol.txt" can only be defined once, but you can
define as many symbols as you want under that key.  For multiple
definitions of the same key, only the first key is valid.

V. User Defined Word (Phrase)

All user defined Word start with 'u' followed by a key which consists of
one to four letters.  The keys related to UCWs are defined in text file
"usr_cw.txt".  Read the comments in file "usr_cw.txt" to find out how to
define UCW of your own.

User defined CW could be anything you like, such as names, organizations,
technical terms, addresses, commonly used expressions, and so on.  These
stuffs are usually specific to the user, and vary from person to person.
With MEPY, everyone can define his own set of UCWs, and these UCWs are
used in much the same as other CWs, except this is one to one or one to N
mapping from key to the related UCW, and you have to remember these keys
youself. If you have lots of UCWs, then it is better you define the keys
in a consistent way. For example, you always use the first PinYin letter of
each HZ as the key, up to four letters.  For technical terms, you can
also use the relevant English as the key.

Note like symbols, each key in "usr_cw.txt" can be defined for more than
one time, and when typing the key, all UCWs related to that key will be
displayed, just like the build-in CWs.


==========================================================================
                Rules of Thumbs in Using MEPY
==========================================================================

0. If you can use CW, use it, unless it is not defined.  For commonly
used phrases, define them as UCWs.  For commonly used symbols, assign
a specific key to each of them.  For those frequently used single HZs,
like numbers, adverbs, conjunctions, most probably they can be found in
hot HZ caches.

1. The first consonant PinYin letter is the key to the hot HZ cache which
is updated every time you select a HZ starting with that letter.

2. Continue typing, you will get HZs. The symbols '>' and '<' in the input
area indicate there are more in the right and/or in the left. Use '.' and
',' to search more in the right or left hands.  Also you can use '>' and
'<' to search other entries with similar sounded PinYin.

3. You can continue typing in to get a CW. For some CWs with 3 HZs or more,
you can type in one more key after the first two HZ Pinyin to search more
specifically.  You can also search right and left in the same way as HZ.
The only thing to notice is that sometime, you may need a <'> to separate
two HZ, such as "xi'an" is for a city name, but "xian" is a single HZ; also
"ke'neng" is for the CW meaning "possibility", otherwise, it may be
interpreted as "Ken Eng".  The keys are defined in file "usr_cw.txt".

4. To find various symbols, type in 'v', followed by some punctuation keys
or English letters, not exceeding four.  To find user defined CWs, type
in 'u', followed by some English letters, not exceeding four.  The keys
are defined in file "symbol.txt".

5. When you find the HZs or CWs, you can use number keys "0" to "9" to
make the selection, or use space bar to select the one at the very
beginning (same as "0").  If you want to keep the input after selection,
you can use <shift> + "0" to "9" to make the selection, in that way, you
can continue selecting the same HZ or CW without retyping.

Warning:  You may find some commonly used phrases missing in MEPY. That
is true, I just don't have time to thoroughly check the correctness and
completeness of all data.  If you find any commonly used CWs missing,
record them, and send a copy to me. I will put them into MEPY in the next
release.

Enjoy and have fun, you have got the most friendly HZ input method you
have ever met.

Qing Zhang

qzhang@rivendell.uvic.ca
qzhang@inde.bc.ca

Burnaby, B.C. Canada

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue May 18 13:28:27 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA17845; Tue, 18 May 93 13:28:24 CDT
Message-Id: <9305181828.AA17845@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7937; Tue, 18 May 93 14:27:35 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 0546; Tue, 18 May 1993 14:27:33 -0400
Date:         Tue, 18 May 1993 11:25:00 PST
Reply-To: TAN%CWU.BITNET@uga.cc.uga.edu
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: george chang <TAN%CWU.BITNET@uga.cc.uga.edu>
Subject:      Re: Chinese keyboard layout
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I read a book on the history of typewriter a few months ago.  I remember
the layout of keyboard is determined by not only the letter frequency, but
also letter combinations, such 'er', 'or', etc.  Logically, Chinese keyboard
layout needs to take letter combinations into account.

The book also says the current layout of English keyboard is far from the
optimum.  A new layout, which is designed by a professor of University of
Washington, might be introduced and enforced at an authoritative conference
as early as 2005.

============================================================================

From:   IN%"jiang@IFCSS.ORG"  "Yuan Jiang" 18-MAY-1993 10:31:18.80

I wonder what input method you based your statistic on.
Key-strike frequency depends on the input method.
For PinYin input methods, "z" is used frequently but
"v" is never used.

There are many DOS and MS-Win utilites that let you to
remap your keyboard.  But you need to know the frequency
statistics or your input method to design your remapping.

From L.S.Ng1@southampton.ac.uk Tue May 18 13:33:59 1993
Return-Path: <L.S.Ng1@southampton.ac.uk>
Received: from sun2.nsfnet-relay.ac.uk by ifcss.org (4.1/IFCSS-Mailer)
	id AA17888; Tue, 18 May 93 13:33:57 CDT
Via: uk.ac.southampton; Tue, 18 May 1993 19:13:09 +0100
From: "L.S.Ng" <L.S.Ng1@southampton.ac.uk>
Message-Id: <28159.9305181757@mail.soton.ac.uk>
Date: Tue, 18 May 93 18:57:15 BST
To: jiang@ifcss.org (Yuan Jiang)
Subject: Re: A Chinese Keyboard Question (rephrased)
In-Reply-To: <9305181715.AA16870@ifcss.org>; from "Yuan Jiang" at May 18, 93 12:15 pm
X-Mailer: ELM [version 2.3 PL11]
Status: RO

That analysis was based on Cangjie. I purposely left out the details to make
the question simpler. But I will post my report on the analysis to CCNET soon.

Since you said keyboard remapping depends on the frequency distribution: that's
exactly what I give you (you don't really need to care which input method it
is). It should be possible to work out a design. But I do not have much
knowledge on keyboard designing so I post it to ask for solution.

Cheers
liangsheng wu

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed May 19 23:38:23 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA09022; Wed, 19 May 93 23:38:19 CDT
Message-Id: <9305200438.AA09022@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7039; Thu, 20 May 93 00:37:20 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 3020; Thu, 20 May 1993 00:37:20 -0400
Date:         Wed, 19 May 1993 21:35:00 PST
Reply-To: TAN%CWU.BITNET@uga.cc.uga.edu
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: george chang <TAN%CWU.BITNET@uga.cc.uga.edu>
Subject:      Re: Processor-assisted keyboard
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

According to an article in People's Daily, 30 to 40 hanzi can be handwritten
per minute at the fastest. typing must be faster. Furthermore, handwriting
is more easily tedious.

G. Chang
===========================================================================
From: Duh? <bo@SAVIIK.CEM.MSU.EDU>
Subject: RE: Processor-assisted keyboard...

I agree that pen-input does have quite some potential, regardless of the
language base. But there is one area that a pen can never outperform a key-
board: text entry. For English texts, one can easily outperform the swiftest
writer with a keyboard after minor amount of training. For Chinese text it may
not be so clear-cut for the time being and this is the whole point. With any
reasonably designed keyboard-based input scheme, a keyboard is still far
superior to hand-writing in text intensive tasks.

Regards

Bo

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue May 18 17:57:46 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA21448; Tue, 18 May 93 17:57:32 CDT
Message-Id: <9305182257.AA21448@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6514; Tue, 18 May 93 18:56:19 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 4709; Tue, 18 May 1993 18:56:17 -0400
Date:         Tue, 18 May 1993 18:54:33 EDT
Reply-To: Duh? <bo@SAVIIK.CEM.MSU.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Duh? <bo@SAVIIK.CEM.MSU.EDU>
Subject:      Processor-assisted keyboard...
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

First off, let me say this: I think keyboard input is inherently improper for
Chinese. Keyboard is likely to stay in face of technological advances for
languages with one-dimensional structure and very limited number of basic
elements -- even if voice recognition becomes perfect, there're still lots of
situations where you don't want to speak out. It's the **natural** thing to do.
You can do all kinds of tricks to use keyboards for inputting Chinese. You
can make it fast, or easy to learn. But you simply can't make it intuitive,
not without some revolutionary thinking.

I can't envision the natural choice for inputting Chinese. Voice recognition
for Chinese language has a big advantage over European counterparts. Every
Chinese character is of single syllabose (sp?), making the variety in Chinese
sounds much more limited. Perhaps it'll take a big role in certain types of
situations of Chinese computing someday. But I want to talk about keyboard here.

For an input mechanism to be widely accepted, it has to be intuitive. That
means it should match the way people learned the language (let's refrain from
being futuristic here...), i.e., how they write. As all of us know, current
keyboards are designed for one-dimensional structures while Chinese characters
are inherently two-dimensional (just think how much trouble you have to go
through to accept superscripts and subscripts!). For handling such a complicated
task, it makes all the sense to have a built-in processor in the Chinese key-
board, taking care of the preliminary part of assembling the keycodes -- mappingkey-combinations to a (preferrably) superset of standard ascii code. An 8086
chip sells for $5/piece and it should suffice for this task.

One proposal:
1. It has to be a superset of the "standard" keyboard. Maybe with two modes --
one Chinese and another ascii -- toggled via a switch/key. Some technical
ingenuity is needed to make this intuitive and comfortable.
2. The basic-element part contains a carefully chosen set of basic Bi3Hua4 and
Pian1Pang2. Preferrably no more than 60, including shifted (call it Ci4?).
3. The other major part is a position-directing unit. May very well be the same
layout of the current numeric keypad. That'd be equivalent to dividing a
character into a 3x3 grid. For example, type a Pian1Pang2, press 3 on the
"position keypad" to put it on the lower-right corner.
4. Of course, simplifications are possible. There are three positions for
starting a character: left-middle, top-left, or top-center. Then it proceeds
from left to right and/or top to bottom. This is where the "on-board" processor
could be of tremendous help. Hopefully, for most of the characters one doesn't
need to use the "position keypad," as it is automatically inferred by the
processor -- the user simply types Pian1Pang2 in the same order as in standard
writing. This way, most of the characters may be finished with no more than 4-5
keystrokes. Best of all, it'd be quite intuitive as it matches the way one
writes. (Of course, training/practice is need to be familliar with the layout
-- the same for English speakers using typewriters.) Lian2Xiang3 can be added
on top of that, of course. (The difference between Fan2 and Jian3 would be
tough to handle, though. {-(  )
5. As for multiple choice from Lian2Xiang3, there could be a little LCD display
on the keyboard displaying successive choices in interval of, say, 0.5 second.
Just press spacebar when you see the right one. Currently, it's usually handled
via pressing a number. I find that very unintuitive and disrrupts the flow.
This doesn't conflict with the number-pressing mechanism, though. If you are
familliar enough to know the choice number of the character you want, just
press the number instead of waiting for it to appear on the LCD.

Anyone interested in making one? ;-)

As a side issue, it makes all the sense for a "Chinese computer" to have
special OS as well as hardware to handle the input/display of the characters.
Again, of course, all should remain compatible with the "standard" counterparts.
There is a big price in performance to be paid as now -- for most of the
current Chinese packages, all the tasks are handled at the application level.

Naturally, such a "native Chinese computer" would be more expensive. But I'd
be surprised if there is a lack of interest in Taiwan, Hongkong, and other
Chinese communities. Assuming the mailand economy keeps going well, there'd be
a humongous interest there soon. A big sales volume would help bringing down
the price.

Bo Peng

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu May 20 00:19:25 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA09734; Thu, 20 May 93 00:19:12 CDT
Message-Id: <9305200519.AA09734@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2579; Thu, 20 May 93 01:17:58 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 4041; Thu, 20 May 1993 01:17:57 -0400
Date:         Thu, 20 May 1993 00:18:25 CDT
Reply-To: Yuan Jiang <jiang@ifcss.org>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yuan Jiang <jiang@ifcss.org>
Subject:      Re: Processor-assisted keyboard...
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I believe pen-based input would be ideal for Chinese input.
Professionals may use Changjie or WuBi for speed input,
but the majority needs something that does not require
extra learning and struggling.

But the technology is not there yet for pen input.
I saw several pen input products in Taiwan last year.
They are inprecise and slow.  Maybe 2-5 years later,
it may be different.  Am I too optimistic?

From hubo@ift.fhg.de Mon May 24 04:49:38 1993
Return-Path: <hubo@ift.fhg.de>
Received: from fhg.de by ifcss.org (4.1/IFCSS-Mailer)
	id AA00815; Mon, 24 May 93 04:49:19 CDT
Received: by fhg.de (mail-gw.fhg.de) with PRESMTP; Mon, 24 May 93 11:48:15 +0200 from FHG-GATEWAY 
Received: by fhg.de (mail-gw.fhg.de) with UUCP; Mon, 24 May 93 11:47:55 +0200 from ift 
Received: by ift.fhg.de; Mon, 24 May 93 11:33:23 +0200
Received: by hubo.ift.fhg.de; Mon, 24 May 93 11:29:34 CDT
Message-Id: <9305241629.AA04699@hubo.ift.fhg.de>
To: Yuan Jiang <jiang@ifcss.org>
Cc: Multiple recipients of list CCNET-L <CCNET-L@UGA.BITNET.DBP.de>
Subject: Re: Processor-assisted keyboard 
In-Reply-To: Your message of 21 May 93 16:52:03 +0200.
             <DFNGATE:N2pdKwTT> 
Date: Mon, 24 May 93 11:29:33 CDT
From: "HU Bo, IFT Munich, +49 89 54759129" <hubo@ift.fhg.de>
Status: RO

>...
>typist uses a pen pointing at the desired radicals.  The typist
>I talked to liked it.  I wonder why it failed in the market place.
>I'm sure the price was a major factor.  But was there any major
>technical falt that prevent it from becoming popular?
>The only thing I can thing of is the speed and distracting --
>...

Many netters have posted their arguments for a pen input method of
Chinese characters (or Kanji for Japanese). I think that it is time
to point out the main (and, like a German would say: crushing) 
disadvantages of such an input method.

1) From the theoretical point of view of physics...
.. Your arms are much more longer than your fingers. So you have to take
longer time time to move your arms than fingers. That was the main problem
for a mouse-oriented input method I have implemented experimentally three
years ago[1]. To stroke a key is **much faster** than to draw a line. That
is the same reason why the one who types using two fingers is much slower 
than who uses his all ten fingers.
2) From a theoretical point of view of computer science (data technique)...
.. To enhighten the speed, a broad bus (8 bits? 16 bits? 64 bits?) should
be used. We happen to have got ten fingers but only two arms. Three-key
combinations are often used [especially for system programmers
and the people who have many undocumented programms], how would you 
implement a three-arm combination? [san1 zhi1 shou3 !?])
3) From the cultural aspect...
.. We have many Chinese characters with many strokes, in Taiwan [I will
try to not be political] even more. For example you want to input a
character which means "trout", zun(1) in pinyin. (You have 5 characters
in GB-Chinese with the same pinyin, in BIG5 maybe a couple more.) Ok,
you have to input the first 8 or 11 strokes which means "fish" and there 
are still 63 (!) characters in GB-Chinese fitting these strokes (assuming
there is no character which happens to have the same stroke sequence but
is not fish-redicaled). So you have to input 12-15 strokes to get the
right one. It is likely that it is easy for children to learn to input
their first (very easy) characters with pen-pad. But my three year old
son is even able to start Windows tipping and saying "w,i,n,big-key" and 
then start paintbrush to paint using mouse...
4) From computer scientific aspect...
.. One could say, ok, we implement the radicals as a bottom on our
pen-pad. So you have about 200 bottoms, or 100 bottoms with shift. So
you have exactly the M-101 or M-102 keyboard which we have but only
one finger (your expensive pen) to tipp on. If you want
to reorganize some of the radicals into to some groups, you have again
to learn Cangjie (or Cangjei) or Wubi or whatelse and learn to guess or 
think like the designer of the concept. [An Apple's slogan: "You don't need
to think like me, because I am thinking like you"]
5) From economic aspect... 
.. there are about 100,000,000 keyboards or much more installed. there is
no doubt that a keyboard will always be about 5-10 times cheaper than a
pen pad.

Any more questions?


Bo



[1] I have implemented the method of Li Jinkai, or the eight-stroke method
    for mouse.

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue May 25 04:20:29 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA14694; Tue, 25 May 93 04:20:23 CDT
Message-Id: <9305250920.AA14694@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 0476; Tue, 25 May 93 05:19:06 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 0782; Tue, 25 May 1993 05:19:04 -0400
Date:         Tue, 25 May 1993 05:19:00 -0400
Reply-To: "HU Bo, IFT Munich, +49 89 54759129" <hubo@IFT.FHG.DBP.DE>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "HU Bo, IFT Munich, +49 89 54759129" <hubo@IFT.FHG.DBP.DE>
Organization: FHG
Subject:      Re: Pen-based input
Comments: To: Yuan Jiang <jiang@IFCSS.ORG>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <DFNGATE:N2Eu7ANO*>
Status: RO

I tried out one of the first pen-input boards in 1987. It was a
"true pen input" (see far below). It did have some pen-driven bottoms
to accelerate the input. The developer told me that this is the only
way to avoid the many strokes one had to write. (For example: "trout",
zun1.)
The board looked schematically like following:

****************************************
*                                      *
*  Fish    Ox     ..   ..  ..  ..      *
*                                      *
*  Wood    Water  ..   ..  ..  ..      *
*                                      *
*  Human   +---------------------------*
*          |\            |            /*
*  Grass   |  \          |          /  *
*          |    \        |        /    *
*  ..      |      \      |      /      *
*          |        \    |    /        *
*          |          \  |  /          *
*          |            \|/            *
*          |-------------*-------------*
*          |            /|\            *
*          |          /  |  \          *
*          |        /    |    \        *
*          |      /      |      \      *
*          |    /        |        \    *
*          |  /          |          \  *
*          |/            |            \*  (viewgraph made
****************************************  using keyboard)

In 1990, I met him again. His neu idea was to design a keyboard
with mini LCD displays on each key to bild a tree for input Chinese
characters...

Bo

-------------------------------------------------------------------
Yuan Jiang <jiang@IFCSS.org>:

There are two kinds of "pen-based" input methods, which people
keep on mixing them up.  One is the true pen input, which you
use like a pen and the computer uses handwriting recognition
techique to recognize the character.  The other is using a pen
-shaped pointer, much like a mouse in principle, to point at
the radicals and strokes.

I don't argue with Hu Bo on the disadvantages of pen-shaped pointing
input device.  But it actually should not be called pen-based
input.  His argument does not apply to the true pen-based input.

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Jun  9 21:43:23 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA23579; Wed, 9 Jun 93 21:43:11 CDT
Message-Id: <9306100243.AA23579@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9642; Wed, 09 Jun 93 22:41:30 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 3582; Wed, 9 Jun 1993 22:39:48 -0400
Date:         Thu, 10 Jun 1993 10:40:37 CST
Reply-To: Shann Wei-Chang <shann@WANGWEI.MATH.NCU.EDU.TW>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Shann Wei-Chang <shann@WANGWEI.MATH.NCU.EDU.TW>
Subject:      STZY.tit
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

While I was waiting for the making of GNU C, I modified the ETZY.tit
file for the official keyboard layout defined in Taiwan (I don't agree
with it, but more and more users start accepting it; this makes me
recall a piece of "fortune", the "users" are the "programmers who
believe whatever you tell them").

If you use cxterm, you know tit file defines an input method.  You compile
this file with tit2cit to produce a cit file for cxterm.  ETZY is a phonetic
input method using the symbols BoPoMoFo, which are mapped into the standard
keyboard by ETen; that is, Bo at key b.  STZY also uses BoPoMoFo but with
a different mapping, Bo at 1.  A complete mapping is shown in the comment
block of STZY.tit.  Most of the keyboard sold in Taiwan (from PCs to
Workstations) now has this mapping printed on the keys along with English
alphabets.

Ftp to the anonymous account of dongpo.math.ncu.edu.tw (140.115.25.3),
get pub/shann/chinese/pinyin/STZY.tit.  I don't mind if anyone keeps a copy
at other ftp sites.

-Shann

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Sat Jun 12 09:13:41 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA01028; Sat, 12 Jun 93 09:13:39 CDT
Message-Id: <9306121413.AA01028@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9250; Sat, 12 Jun 93 10:13:07 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 2782; Sat, 12 Jun 1993 10:13:06 -0400
Date:         Sat, 12 Jun 1993 09:13:18 CDT
Reply-To: Yuan Jiang <jiang@ifcss.org>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Yuan Jiang <jiang@ifcss.org>
Subject:      Re: announcement of a freeware mac program
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Liedong's Mandarin is a very interesting software.  It's
like you are a big boss and have a secretary reading the
newspaper loud to you.

I saw such a software for English a few years ago.  Beside
being monotone, it read a MacWrite English text file perfectly
and even could pronouce a few 4-letter words.

It's Liedong's big idea that Chinese (Mandarin) is actually
easier in text->language conversion than English.  In another
word, Chinese may be easier when comes to making multi-media
software.

This advantage relates to the difficulty in Chinese input
and voice recognition.  It's the old problem of
multi-character mapping of PinYin->GB(Big5).
However, the difficulty is an advantage when Liedong
does the reverse conversion.

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Mon Jul  5 18:53:03 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA26361; Mon, 5 Jul 93 18:52:39 CDT
Message-Id: <9307052352.AA26361@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6602; Mon, 05 Jul 93 19:51:42 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 2871; Mon, 5 Jul 1993 19:51:39 -0400
Date:         Tue, 6 Jul 1993 00:29:00 GMT0BST-1
Reply-To: jdelacour@cix.compulink.co.uk
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: John Delacour <jdelacour@CIX.COMPULINK.CO.UK>
Subject:      <<CHINPUT>>
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

          * * * * * * *     C H I N P U T     * * * * * * *

Without any thought of computers in mind I devised, some 20 years ago
a system of representing any chinese syllable with two latin
characters.  I saw this as the natural way of writing Chinese in the
most economical fashion for native Chinese, who always graduate to
the traditional writing system that is imperative for the proper
representation of the language.

Now that the personal computer is a powerful tool within the reach of
anyone in the west that does not prefer to waste his money on more
expensive and less useful things and is also bound soon to become an
everyday object in Chinese life even on the mainland, the idea that I
had originally conceived in order to save ink can now be used to
reduce RSI (repetitive strain injury!).

I do not make any great claim to originality in my "Chinput", since
the nature of the Chinese language almost demands such a phonetic
representation.  "Chinput" is in effect a direct leap from the
earliest phonetic representation.

Now I don't want to "wield my axe outside Master Han's front door" or
even "teach my grandmother to such eggs", so those of you who are
well-versed in these matters will excuse me if I give a brief
introduction.  Qing yuanliang ba!

I am talking of Mandarin, though the same principles apply for all
the Han dialects.  Words written in capitals are PINYIN.  Every
Chinese character is pronounced as a monosyllable and that
monosyllable comprises an "initial" sound and a "final" sound.  There
is a finite number of initials and of finals and they may only be
combined in certain configurations.  This is not so in English.  In
English it would be possible to create <infinity minus a few>
monosyllables but in Chinese, although there is a syllable TUAN and a
syllable SHUANG, there is no syllable TUANG.  This does not mean that
TUANG would be unpronounceable for a Chinese, but simply that it
would not be allowed under the laws of the language.  This phenomenon
is extremely special but it is a fact and that is what we are working
with:- facts!  Before I leave this section I will simply list a
number of syllables that _could_ exist in Chinese but don't : - TUA;
SHONG, CHEI, PUN, MUANG, LIONG, DIANG etc.

Now if we neglect tones, there are fewer than 400 monosyllables in
Chinese.  The meaning of the spoken word is made clear by 1) the four
tones, 2) by combination (cizu) and 3) by context.

The old method of indicating the pronunciation of a ZI4 was to use
two familiar ZI4 and to cut (QIE) each of them in half, throwing away
the second half of the first ZI4 and the first half of the second ZI4
and joining the remainder togeth to give the sound of the new word.
For example we find the word HUA (flower) in the Kangxi dictionary
and the pronunciation is given as H-U:G-UA (exhale:gourd).
We throw away the middle and get the pronunciation.  This system of
indicatiing the pronunciation is still current in many areas of
China.

All the other phonetic transcriptions of Chinese have ignored the
essential duality of the Chinese monosyllable.  Even the National
Phonetic Alphabet added unnecessary "medials".  I will explain what I
mean by this:-  the word XIA2 is written thus in PINYIN; in the Wade
transcription it is <hsia@>, in the Yale transcription <sya2>, in
Zhuyin <.T|Y>, <shya> and in Kangxi <H-e:jIA> (what:add).  To sum
up:-

                   * * * SUNSET * * *

KANGXI                     H  [e:j]  IA  (what:add)  {wrong tone
often!}
WADE-GILES                 hsia2
YALE                       sya2
Bopomofo                   .T|Y
Zhuyin                     shya
PInYIN                     xia2
*Chinput*                  xa        (or xa2   or   x)

So let's "get back to our sheep" as the French say and stop wielding
our axes and teaching our grandmothers <g>.

Only two characters are needed to identify a Chinese monosyllable and
two the 26 characters of the Latin alphabet are adequate to define
any Chinese monosyllable, provided that two values are assigned to
each character.

Now I hear all you Macintosh people and you blind people shouting,
"BIG DEAL, we already have Shuangpin !!"  Well, yes, the idea is just
the same except that I invented my method and somebody else invented
the other method.  The principle is the same but *  C H I N P U T  *
has a number of features that make it more usable.

Before I continue, let's have a look at the  *  C H I N P U T  *
"alphabet", or, if you like the aabbcc (ABIANCUN)

                   *  C H I N P U T  *

||||||||||||||||||||||||||||||||||||||||||||||

         a-a                    n-ei
         b-ian                  o-o
         c-un                   p-ang
         d-uan                  q-ue
         e-e                    r-ou
         f-an                   s-ai
         g-uang                 t-ao
         h-en                   u-u        (uu = WU)
         i-i                    v-ui       (vv = SHUI)
         j-ing                  w-uai      (ww = ZHUAI)
         k-ua                   x-iao
         l-in                   y-ong      (yy = CHONG)
         m-ie                   z-eng

         lia

         n, iong, ia, iang, ie, an, iong follow only j,q,x,i


         -uo > o
         iu

The characters as initials have mostly the values that they would
have in pinyin.  The exceptions are v-, w- and y.  V- does not exist
in pinyin and there is no v sound in Putonghua.  This letter can
therefore be used to replace one of the initials that requires two
letters in pinyin.  I have assigned the sound sh- to this letter.
Now there is nothing w and y can do as initials that cannot equally
well be done by u- and i- (incidentally these initials were used for
words of the first tone in the Gwoyeu Romatzyh system also) and they
can therefore be used for the consonants zh- and ch-.  So all the
initial consonants can now be represented by a single character.

Now for the letters as finals.  All words in putonghua end in a
vowel, -n or -ng.
In *  C H I N P U T  * characters with ascents are used for finals
ending in the -n sound, and those with descents for finals ending in
-ng, thus :-

-d-uan                 -g-uang
-f-an                  -j-ing
-h-en                  -p-ang
-k-ua                  -q-ue
-l-in                  -y-ong                       -z-eng

Unfortunately it is not possible to avoid the obvious exceptions, but
at least z is as often as not written with a descent in handwriting.

So much for the basics.  Anyone with a knowledge of pinyin would
already be able to write almost all Chinese syllables using Chinput,
but there are further economies that must be explained.

The initials j-, q- and x- can only be followed by the vowel sounds
-i or .  The -i- in the pinyin jian, qiang, xie is not necessary
since the sound is as it were incorporated in the initial consonant.
These words are written in Chinput as jf, qp and xe.  They could also
be written jb, qp and xm but a decision has to be made.  There is a
separate final -ian (-b) because there are the words ban, bian, tan,
tian, lan, lian etc.  The only two syllable that are neglected in
this regard are liang and niang.  These are therefore the first two
syllable for which an arbitrary Chinput is needed:-

liang   lg                  niang  ng

Similarly when j, q or x are followed by a final -u or -u-- the u is
pronounced .  The only casualty here is lan.  There is only one word
with this pronunciation and modern dictionaries do not recognize this
pronunciation and substitute lian.  I give here a few examples of
words that should make clear how this principle is applied:-

                  lan    lb                  ne    nq
                  juan    jd                  que    qq
                  xun     xc                  l     l

Words that in pinyin have the final -uo have -o as final in Chinput
since there is no possibility of ambiguity here; the initial dictates
whether the transitional -u- sound is present:-

                  bo         bo                  cuo    co
                  duo        do                  fo     fo
                  guo        go                  huo    ho


So much for the Chinputting of RdanciS (single syllable words).  Now,
since the primary aim of *  C H I N P U T  * is to provide an
economical method of inputting Chinese to a computer, it is
unnecessary to use any upper case characters; by the time the input
is converted to Hanzi it will not make any difference.  This leaves
the upper case characters free for other purposes.

Anyway, that will do for the moment.  I have a mush longer file on
the subject of course, but I thought I would write this much in order
to hear what you all have to say on the matter.

Here is a short extact from LAO3SHE4  (ltve)  to illustrate how
 *  C H I N P U T  * might work


                            (Chinput Text)


RirmoS wnge zi zs zidb vp ir vilsge buty d djii.  hsvi ba zidb fpxa,
rp zamh svbb tf ba..  ju uo kf, ta vrit -d v- iiwy xjts.  uomh widt,
ir xudo rh vhjj-goml -d, mnmn ii godu -d gfqj kf vi, er bu kz ry
rh.veip rh jaro v uhii zoja, tade zopj-wy bi hfwe qplm -d cijixj, ho
ltst, ho vpgf; ta lt kf bmrh bu vcif, er id vi daja dr svwe ta ziji
zr, hovi dv ziji d ztiu bu mf, er vpgf d zilb.  ffwi, irmo d rh bb b
weip, ta ji b huht jtma, kf bmrh dr bv dyxi, ie b gdijzilb, kf ziji
ru iiho btbn.  Ta v ir vi vi wy kfyu kext wi db, er jiqt xmyuls.


                            (PINYIN Text)

Youmo zheige zi zai zidian shang you shilaige butongde dingyi.
Haishi ba zidian fangxia, rang zamen bian tan ba.  Ju wo kan, ta
shouyao de shi yizhong xintai.  Women zhidao , you xuduo ren shi
shenjing-guomin de, meimei yi guo du de ganqing kan shi, er bu keng
rong ren.  Zheyang ren jiaruo shi wenyi zuojia, tade zuopin zhong bi
hanzhe qianglie de cijixing, huo laosao, huo shanggan; ta lao kan
bieren bu shunyan, er yuan shi dajia dou suizhe ta ziji zou, huoshi
dui ziji de zaoyu bu man, er shanggan de zilian.  Fanzhi, youmo de
ren bian bu zheyang, ta ji bu huhao jiao ma, kan bie ren bu shi
dongxi, ye bu gu ying zi lian. kan ziji ru yihuo bao bei.  Ta shi you
shi shi zhong kanchu kexiao zhi dian, er jiqiao de xiechulai


                      (VERY ROUGH English Text)

Look up the word humour in the dictionary and you will find a dozen
or so definitions, so letUs forget about the dictionary and talk
about it.  As I see it, itUs an attitude of mind.  We know that there
are a lot of neurotic people around who perceive things with an
exaggerrated sensibility and canUt get on with other people.  If such
a person is a creative writer, his works are bound be ridden with
bitter sarcasm or complaints or sentimentality;  he is constantly
looking down on people and wants everyone to march to his tune; he is
not content with his lot and feels sorry for himself.  But the man of
humour is not like this; he doesnUt spend his time bemoaning his lot
and railing at the world and looking down his nose at people and he
doesnUt have any illusions about himself or feel sorry for himself or
regard himself as GodUs gift to the world.  He  seeks out what is
humorous in things and writes ............



Here follows a brief description of "Shuangpin" for comparison

@@@@@@@@@@@@@  Appendix: ShuangPin (by courtesy of Edmund
Lai)@@@@@@@@@@@@@@

          ShuangPin uses two letters to represent a sound.  This
means if           with Cihui homonyms are under control then you can
use about 2           keystrokes per Chinese character even without
abbreviation, so it is           near optimal for input.  The
drawback is that you have to           learn to remember the position
of the keyboard so it is not for the           casual user.  It is
based on PinYin.  It is fairly popular in China           and is one
of the input method in CCDOS.  The consonant is identical
to PinYin except you use V for Zh, W for Ch and Y for Sh, the pseudo
         consonants W and Y normally used in PinYin are not used.

          For vowels, the mapping is:
          A                  a
          B                  ai
          C                  ao
          D                  an
          E                  e
          F                  ang
          G                  eng
          H                  ou
          I                  i
          J                  ie
          K                  ei
          L                  ong,iong
          M                  in
          N                  en
          O                  o,uo
          P                  ing
          Q                  ia,ua
          R                  ian
          S                  uan,uan:
          T                  un,un:
          U                  u
          V                  ui,ue:
          W                  iu
          X                  iang,uang
          Y                  u:, uai
          Z                  iao

          For those without consonants (including the pseudo ones
like Y           and W), you either use O as the first letter except
the following           use E

          EL                  Yong
          EO                  O
          EQ                  Ya
          ER                  Er
          ES                  Yuan
          ET                  Yun
          EV                  Yue
          EX                  Yang
          EY                  Yu

          All of these are my reconstruction, and there may be
mistakes.

          And the following are my own additions

          EE                  Eh
          EZ                  Yo
          EB                  Yai


                       ************************************

There are improvements to be made in *  C H I N P U T  *  and these I
will make all the more quickly if I get some response and
encouragement from the world outside, so please have a think and see
if this is not the simplest and most economical way of inputting
Chinese to the computer.
I will write more on this subject when time allows.

                   John

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Jul  6 16:15:19 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA06193; Tue, 6 Jul 93 16:15:00 CDT
Message-Id: <9307062115.AA06193@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 8251; Tue, 06 Jul 93 17:13:49 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 1332; Tue, 6 Jul 1993 17:13:44 -0400
Date:         Tue, 6 Jul 1993 17:10:00 EDT
Reply-To: John Delacour <100111.3564@COMPUSERVE.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: John Delacour <100111.3564@COMPUSERVE.COM>
Subject:      Re: Zhenbing Zeng / zh ch sh in chinput
Comments: To: Chinese Computing <CCNET-L@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Memo #403453 (1680)Received: from smtpsrvr by cix.compulink.co.uk id aa18673;
6 Jul 93 9:33 GMTReceived: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP
V2R2)   with BSMTP id 7281; Tue, 06 Jul 93 05:31:42 EDTReceived: from
UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f)
with BSMTP id 4125; Tue, 6 Jul 1993 05:31:41 -0400Date:         Tue, 6 Jul
1993 11:30:35 +0200Reply-To:     Zhenbing Zeng
<zeng@mathematik.uni-bielefeld.de>Sender:       Chinese Computing Network
<CCNET-L@uga.cc.uga.edu>From:         Zhenbing Zeng
<zeng@mathematik.uni-bielefeld.de>Subject:      zh ch sh in chinputComments:
To: ccnet-l@uga.cc.uga.eduTo:           Multiple recipients of list CCNET-L
<CCNET-L@uga.cc.uga.edu>Apparently-To: jdelacourHi, friends,Shuangpin is
useful for many people. It would be better to map zh, ch and shto some vowel
keys near z, c and s, so people can remember them easier. ForQWERTY keyboard,
it may be better to assign              ch -> v; sh -> a;but there is no vowel
key near z. In German keyboad: qwertyuiopis qwertzuiop,so probably better to
assign              zh -> u.(How about to assign  zh -> <, the non-letter key
in the left of z?).There are two things one need to consider for define
mapping table: linguisticfact and psychological aspect. I remember that the
Stone 2402 keyboard is alsoa shuangpin one, it's quite good.It's also very
important to have a unified shuangpin keyboard.Yishang zhishi wo geren de
yidian kanfa. Wo xiwang nengou qidao pao-zhuan-yin-yude zuoyong. Shijian bu
gou, jiu xie zhexie ba. (Tingshuo zai shanghai you yigebaozhi, jiao "Hanyu
Pinyin Bao", you shui kanguo meiyou?)Zaijian!

Zhenbing,

I wonder if it is really important for there to be any relationship in the
mind between S & SH, Z & ZH, C & CH ?  After all there is no phonetic
similarity between the tow members of each pair and it is purely a convention
established for PINYIN.

One thing I did not mention when I posted my message introducing "Chinput" was
that I have tried to make the method as easy to learn as possible.  Thus each
letter of the "alphabet" is memorized as an existing syllable.  ff is FAN, dd
is DUAN etc.; every duplicated character is actually a possible word.  The
Chinput would be memorized as 26 _words_.  The device of assigning --N finals
to letters with an ascent and --NG finals to letters with descents further
facilitates the learning of the code.

I am thinking all the while of an input method that is suitable for use in the
widest possible sphere and not some obscure system that is accessible only to
academics and computer wizards.  I am thinking of ordinary people working in
offices or leaving school and how to train them as quickly as possible to do
Chinese wordprocessing.

I ought to say that I use a Macintosh and have recently bought the Chinese
system 6.0.3.  This provides a "cihui" and I have converted a duplicate of
this to work with my Chinput method with amazing success.  There are problems
which I have not had time to solve and I really hate working with system 6
after converting to system 7.1.  Besides, the whole thing seems to be 3 years
out of date, which is most annoying (I know many people in the West will know
exactly how frustrated I feel).  Nevertheless it works wonderfully.  In my
original posting I did not mention that in my "cihui" the tones are also
included, but as an option.

To explain briefly the way the Mac system works (this will be old news very
soon I hope!) you can toggle between Latin input and Chinese input by pressing
the spacebar.  You select the input method you want to use, which may be any
of the supplied methods such as bopomofo, zhuyin etc. OR your own method for
which you have created a dictionary (the dictionary is not very impressive,
containing only 165 pages of ZI, CI and CIZU - nevertheless it is useful for
experimentation and can be edited and expanded).

Suppose I want to print PU2TAO1JIU....well let's put that in Chinput first of
all and it is puttjo ( Ithink!! --- trust me to choose a word I haven't
learned yet!).  If I type "p", all the words beginning with "p" are displayed
in a scrollable strip.  If I next type "u", then all the ZI pronounced PU are
displayed. If I next type "2", then it is very likely that "pttt" will be
presented to me as an option.  I can then select the CI, press the enter key
and print the characters. And so it goes on by a process of elimination.  My
chinese is not very good but I guess that an input of "p.t.j <CR>" would print
the wine in Chinese characters.

You mention psychology and this is of course very important.  I am a Macintosh
man ansd I think in a Macintosh way.  I see poor children here in England
taking courses in "WORDPROCESSING" as though it were some sort of obscure
science, when in fact it is just a typewriter that can do what you want it to
do.  I think it is a great shame that people still have to learn arcane and
archaic wordprocessing methods and treat it as a science, when it is merely a
menial task that anyone can do. and that goes for Chinese wordprocessing as
well as for latin wordprocessing.

Before I got my Chinese system I created some Chinese fonts using fontographer
in order to experiment with Chinput.  I use Nisus for wordprocessing and this
has a "glossary" feature.  With this I would type my text in Chinput and then
run the glossary. whFrom @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Sun Jul 11 09:29:59 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA29054; Sun, 11 Jul 93 09:29:55 CDT
Message-Id: <9307111429.AA29054@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 1949; Sun, 11 Jul 93 10:27:52 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 2334; Sun, 11 Jul 1993 10:27:49 -0400
Date:         Sun, 11 Jul 1993 07:26:58 -0700
Reply-To: Ed Lai <lai@APPLE.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ed Lai <lai@APPLE.COM>
Subject:      comment on CHINPUT
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

John just asked me what I think of  *  C H I N P U T  *. I thought
of writing a repsonse  earlier, but any comment would be a
long one and takes time. Since now it is early morning and I cannot
sleep, so I may as well do it now. Since the reply would be of
general interest, I am going to post it rather than mail it to him.

This is an insightful article. I am particular impressed by his insight
that two stroke phonentic input is related to the QIE of traditional
Chinese phonetics.

John listed the same word under the various romanization. It is
a useful execercise. When I was working on the HyperCard dictionary
I have to write program to transform them into each other. It gives
you a very clear idea how systematic a particular transcription is.
After all every irreguality requires you to write an exception. I
especially recommanded the excercise to those who think PRC's pinyin
is the perfect transcription. My conclusion is the PinYin is probably
one of the better ones, but it certainly can be improved. However,
I don't even try to come out with a better one.

Recently I spend quite a bit of time teaching my 3 year old son the
alphabet, some of which he keep confusing and I can understand why.
I am quite sure we can come up with better letter shapes, but I am
not trying to.

Cang jie undoubtedly is the most popular input method in Taiwan. It
tries to map the symbols to the alphabets in a logical way, so ABCDE
are the stars, another group are parts of the body etc. The problem is
that when these keys are mapped into the keyboard, it is scattered
all over the place, and the positioning of the keys are not necessary
optimal for typing. Professor Chen, who heads the one of the three
colleages of Chinese University of Hong Kong and is one of the founder
on the Chinese Language computer society, did some research on
this problem. He came up with a keyboard mapping so that the Cang
jie symbols groups nicely on the keyboard, the more frequenctly used
symbols are in the middle row etc so that it is optimal for typing.
However, when I very occasionally use Cang jie, I still use the standard
mapping.

Professor Chen's work was inspired by the DWVOK(?) keyboard. Needless
to say I am now using a QWERTY keyboard.

It is unfortunate that human nature is that it is difficult for a lot of
people to agree on the same thing, people fight over minor differences.
So my philosohy now, not because I like it but I have learned after many
hard lession, is that if there is a standard that is at least semi-reasonable
and is widely accepted.

John thinks that ShuangPin has some special importance in the Macintosh
community. Ha, Ha, actually by my estimation he is the only one in the
world who use ShuangPin or its variation on the Macintosh. (Well maybe
the second one, I think someone in QingHua did some Chinese work on the
128K Mac, remember those?, using MacBasic, and being from QingHua means
it uses ShuangPin). I was so happy to hear John is using it so that after all
my efforts, at least one person is using it. None the less, ShuangPin is a
standard, more on the PC than on the Macintosh.

So I think we should stick with ShuangPin.

However I have no objection for John or anyone to use Chinput. There
are some people who thinks the ShuangPin like scheme remind them
too much of BoPoMoFo, and it is a evil reactionary plot to subvert
the world's most perfect language system, PinYin. The truth is that ShuangPin
is highly related to PinYin, and it is no more than a short hand notation
for PinYin. For computer input, it can be think of as a keyboard macro
arrangement. Everybody can have their own keyboard macro arrangment.
Any who like the DWVOK keyboard is free to do so. This is especially true
of the Macintosh. It tends to be a much more personal and individualistic
machine than the PC. Sharing a Macintosh with someone else is like sharing
your house with someone else, you would rather not or you have to
learn to live with the other party. In your own house you certainly
should do what is most comfortable for you.

When I wrote Cihui, I could have made the decision that it is strictly
for PinYin, as just about every other Pinyin program is. Then I can make
my data structure more compact, the program will be much easier to write,
and it will be in many way a better program. I decide to do it in this general
way, and the effort is wasted on most people. Still I am glad I do it this way,
John can use it for Chinput and ChaoChowese, Fung Fung Lee can use it for
Cantonese and I know someone is using it with Yale.

If one day ShuangPin becomes popular, I may write a special ShuangPin
version, I would be able to doing something that is hard for me to do now.

As a side note, John mentioned that there are sound combination that can
be made in theory but are "not be allowed under the laws of the language".
Actually that is not a law but a fact. I had a friend who died at a young
age. I always feel sad that such a talented indidual and friend can died
so young. He once wrote an article saying that Beijingese is the "National
Language" or "Common Language", but actually it does not quite live up
to the billing. It is a better choice than other dialects but it is rather
deficient in sounds and which is more prolific in other dialects. To be
a truly Natioal or Common language it should try to absorb some of
the sounds and vocabuary from the other dialects so it can really become
a language of all the people rather than imposed on all people. Actually
this is happening naturely to some extent. In recent years because of the
importance of the Hong Kong and overseas Chinese economy, Cantonese
starts to be affecting the langauge, just like Shanghaiese has been.
Purist may be horrified but this is how language evolve, and Beijingese
is a better dialect because it gone through such evolution more than
Cantonese.

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Sun Jul 11 20:17:32 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA05239; Sun, 11 Jul 93 20:17:25 CDT
Message-Id: <9307120117.AA05239@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2816; Sun, 11 Jul 93 21:16:13 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 7033; Sun, 11 Jul 1993 21:16:11 -0400
Date:         Mon, 12 Jul 1993 01:59:00 GMT0BST-1
Reply-To: jdelacour@cix.compulink.co.uk
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: John Delacour <jdelacour@CIX.COMPULINK.CO.UK>
Subject:      more about *CHINPUT*
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

To all, and particularly to Ed Lai.

I am most gratified to have such a long response to my posting on  "Chinput" from Ed Lai, who is so active in the field.

There are more philosophical and psychological points brought up in Ed's  posting than strictly technical ones, and they are all interesting.   Before I comment on these matters I would like to expand a little on my  reasons for advocating Chinput or more generally ShuangPin (I make no  fundamental distinction between the two and have _still_ not had time to  make a detailed comparison between the two -- nor, as I said in my  original posting, do I consider that "my way is best".  All I know is  that I was very pleased with my transcription eighteen years ago when  nobody in Shanghai wanted to know about it, and I'm equally pleased with  it now, when it seems noone in the whole world wants to know about it.)

I wrote in a posting the other day why I thought Chinput might be  slightly more mnemonic than ShuangPin (that 1. every reduplicated letter  represents an existing Chinese syllable and 2.  as far as possible,  letters with "descents" are used for the --ng finals and letters with  ascents for the --n finals, eg. ll=lin; gg=guang) but these are just nice  details; a few days' practice with any two-letter syllable input system  would be enough to commit either method to memory.  My argument is that,  since ShuangPin is not an established standard, then it should be  presented in the most perfect and least arbitrary form just in case one  day people decide they had rather type two letters than three and a half.

My attitude to the English language is similarly minimalistic.  English  spelling is quite absurd and cannot be justified on any etymological  grounds at all.  Much time is wasted teaching people an archaic  orthography that is of no use to them since nobody now learns latin or  greek or is interested in the proper meaning of words.  Why can't we have  rational spelling like the Italians or half the rest of the world?  - -  but this is not to the point..

I see the question from the point of view of that blessed mass of men (ci  nei ziran baokuo nuren a!) who are not computer-literate and who wish  only to get their work done as quickly as possible and go home for tea.   If they can get a ZI4 on screen by typing two letters on their keyboard  rather than four (I can't remember if I mentioned the exact percentage  relationship between Chinput/ShuangPin and PinYin, but the reduction in  effort is very significant) then everyone will be happier, bosses will be  richer and goods will be cheaper and, as we say in England, pigs might  fly! - we all know how the computer has created a world for us where we  can be rich with only ten hours of work per week while the computers get  on with the work!

Now to PinYin --- I myself welcomed the hegemony of PinYin, having  learned my Chinese with Wade Giles and every system under the sun.   PinYin kicked over the traces, told the Jesuits where to take their  pedantic notions etc. etc. and emerged as a rational, no-nonsense  transcription.  But I speak as a red-haired devil, (ang mo gui) as my  ChowChou relations would say, and I think that PinYin is of more benefit  to the foreign learner than to the Chinese.  Like the other phonetic  transcriptions before it, it was born and weaned in the minds of liu  xuesheng - it does not respect the duality of the Chinese syllable and it  adds quite unnecessary elements to the transcription -- and who for?   Why, for the foreigners, of course!  Why should I type ZHUANG when I  could type XY and get the same result?  The reason for the ZH is that in  English (NOT in Italian, please note, nor in romanized Greek, nor in  German nor...probably not in Swedish either  +++ZhongWen please
comment!  +++) CH is pronounced as it is, and ZH is therefore chosen as the voiced  equivalent of the unvoiced CH.  In English, of course, we use J, but that  needs to be distinguished in PinYin and reserved for the more hissy  consonant (I hope no pedantic phoneticians are reading this!).  Jiu  suanle, I'm getting pedantic myself!  What I mean to say is that Chinese,  in whatever dialect, will remain an essentially monosyllabic language and  each syllable in every dialect will have a preordained duality.  Twenty  years ago I was afraid that the beauty and the history of Chinese  calligraphy might give way to the dreadful tide of romanization, and the  advent of personal computers augmented that fear, but on the contrary the  Japanese (god bless them <g>) have shown that the universal computer is  in fact the saviour of Hanzi.  All over the world (probably a very small  world) people are working out how to create outline fonts for Tang, Sung,  Ming styles -- goodness knows,
there may even be someone out there  creating a font for (Chinput tr.)  Mt Zedy (genuflect > > ) calligraphy.  We all miss him so much!  I had to mourn for the bastard for three weeks  before I decided I couldn't afford a new sleeve for my coat and stopped  wearing the black armband.  All this is off the point too!!  We must  never get political.

So let's get philosophical and psychological - It seems, Ed, that you've  decided to take a defeatist road - SHIBAIZHUYI de DAOLU!!!!  That is not  the way.  Look at the work you have done and the hours you've spent  creating something which is really of value, and you diminish your  efforts and say that the world does not recognize you.  OK, the kid's  three and needs attention (would that my kids were so young!) but the  work you ought to be doing now would make that kid a lot happier when  he's 13!  So keep working!

A very great friend of mine, a Jewish professor in Paris, once said to me  as a young man that you should regard your neighbour as a wild animal  asleep -- don't disturb him.  My friend is also now dead.  Of course he  was right.

                         _IN PARENTHESES_

                (???? is Liedong... the same as Cindy...Zheng?)

The terrible truth is that Qinshihuangdi formalised the cart-tracks and  the weights and measures, William the Bastard compiled the Domesday  Book,  Napoleon got Paris looking nice and Hitler and Mao also did the  same sort of thing.  They were all bloody murderers, but they left their  mark.  I don't think we want to see a million or ten million people die  in order to establish a standard Chinese Input protocol, but maybe we're  wrong <g>.

I'll carry on with my "Chaozhouput" but 26 letters are not enough for  that.  Ed says that PekingTalk is refined - I find it far too refined,  like French music.  I prefer the variety and directness of the peripheral  dialects (EXCEPT for Shanghai, which is the most awful thing I have ever  heard or ever hope to hear this side of hell!)

And so to bed.

                       John

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Mon Aug  9 15:02:08 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA12057; Mon, 9 Aug 93 15:02:02 CDT
Message-Id: <9308092002.AA12057@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7071; Mon, 09 Aug 93 16:00:30 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 5847; Mon, 9 Aug 1993 15:59:41 -0400
Date:         Mon, 9 Aug 1993 13:58:24 MDT
Reply-To: mleisher@NMSU.EDU
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Mark Leisher <mleisher@NMSU.EDU>
Subject:      Available: Wnn Programming Library (in C)
Comments: To: insoft-l@cis.vutbr.cz, CCNET-L@UGA.BITNET
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I've been meaning to finish and release this for a year or more, but
never quite got around to it, so here it is in all its rough-edged
glory! :-)

For those of you unfamiliar with Wnn/cWnn, basically they are (Unix)
network based conversion servers for converting
Hiragana/Katakana->Kanji and Pinyin/Zhuyin->Hanzi available in the
X11R5 distribution as part of the contributed i18n support.

This code still needs work, but it has been in use here for quite a
while.  We have only used this code on SunOS systems so I don't know
how many changes will be needed for other systems.  The documentation
is very poor at the moment, so please feel free to send questions.

Improvements, additions, and bug fixes of any kind are welcome!

You can get this code from:
crl.nmsu.edu:pub/misc/wnnpi-0.4.tar.Z

-------------------------------------------------------------------------
                    The Wnn Programming Interface

This software is based on the original version of Wnn developed by
Kyoto University Research Institute for MaFrom @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Mon Aug  9 15:02:08 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA12057; Mon, 9 Aug 93 15:02:02 CDT
Message-Id: <9308092002.AA12057@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7071; Mon, 09 Aug 93 16:00:30 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 5847; Mon, 9 Aug 1993 15:59:41 -0400
Date:         Mon, 9 Aug 1993 13:58:24 MDT
Reply-To: mleisher@NMSU.EDU
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Mark Leisher <mleisher@NMSU.EDU>
Subject:      Available: Wnn Programming Library (in C)
Comments: To: insoft-l@cis.vutbr.cz, CCNET-L@UGA.BITNET
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: R

I've been meaning to finish and release this for a year or more, but
never quite got around to it, so here it is in all its rough-edged
glory! :-)

For those of you unfamiliar with Wnn/cWnn, basically they are (Unix)
network based conversion servers for converting
Hiragana/Katakana->Kanji and Pinyin/Zhuyin->Hanzi available in the
X11R5 distribution as part of the contributed i18n support.

This code still needs work, but it has been in use here for quite a
while.  We have only used this code on SunOS systems so I don't know
how many changes will be needed for other systems.  The documentation
is very poor at the moment, so please feel free to send questions.

Improvements, additions, and bug fixes of any kind are welcome!

You can get this code from:
crl.nmsu.edu:pub/misc/wnnpi-0.4.tar.Z

-------------------------------------------------------------------------
                    The Wnn Programming Interface

This software is based on the original version of Wnn developed by
Kyoto University Research Institute for Mathematical Sciences
(KURIMS), OMRON Corporation and ASTEC Inc (see COPYRIGHT file in WnnPI
distribution).

Note00: Some familiarity with the Wnn family of network based
        conversion servers is assumed by this document.

The WnnPI library is a library of functions for working with the Wnn
family of conversion servers.  The code in WnnPI is based primarily on
the libraries that are provided with the Wnn distribution.  Most of
the WnnPI code doesn't look anything like the code in the Wnn
libraries, but all of WnnPI was written using the Wnn JS and JL
libraries as references.

For most conversion needs, WnnPI can be used instead of
libwnn.a/libcwnn.a or libjd.a, which are provided with the Wnn
distribution.

The libraries that come with the Wnn/cWnn distribution are set up to
be used by both the server and by other programs.  As a consequence,
they are necessarily larger than a library like WnnPI which is devoted
to accessing the Wnn conversion servers via client programs.  The two
main motivations for writing the WnnPI library were to gain a better
understanding of how the Wnn family of conversion servers work and to
provide an enhanced, general purpose interface to these servers for
application programming.

The main enhancements WnnPI provides that the Wnn libraries don't are:

      o  Code heavily documented in English (easier to modify)
      o  Code separated into small modules based on functionality
      o  No global variables (pretty much reentrant)
      o  Manages multiple environments per connection
      o  Many more environment related functions
      o  Easy access to results of conversions and candidates
      o  Chinese Pinyin/Zhuyin <-> Yin code converters (see Note01:)
      o  Error messages in server "language" and "local" language (see Note02:)
      o  Based primarily on the wchar_t type
      o  Much smaller than the libraries provided with Wnn/cWnn

Note01: There are Pinyin/Zhuyin <-> Yin code conversion functions that
        come with Wnn, but they are not compiled into the programming
        libraries: libcwnn.a, and alternatively libjd.a with the
        Chinese code added (as of version 4.10).

Note02: The Wnn/cWnn message catalogs are used and an English message
        catalog (translation of Wnn message catalog) is provided.
        However, the message handling code and structures are set up
        to allow 2 message strings to be specified for each message.
-----------------------------------------------------------------------------
mleisher@nmsu.edu
Mark Leisher
Computing Research Lab                    "Elegance is not optional!"
New Mexico State University                  -- Attributed to Richard O'Keefe
Box 3CRL, Las Cruces, NM  88003

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Aug 10 06:00:14 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA19279; Tue, 10 Aug 93 06:00:01 CDT
Message-Id: <9308101100.AA19279@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9608; Tue, 10 Aug 93 06:58:33 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 9354; Tue, 10 Aug 1993 06:54:21 -0400
Date:         Tue, 10 Aug 1993 18:53:35 CST
Reply-To: Chris Delucia <a1101006@CCMS.NTU.EDU.TW>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Chris Delucia <a1101006@CCMS.NTU.EDU.TW>
Subject:      Chinese WP - Taipei exhibition
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

 I went to a computer & software exhibition in Taipei this weekend
and saw some Chinese wordprocessing software I was rather impressed

        I saw a couple of Pen-based input systems. ETEN makes one, but
I saw a more impressive one (prob. @ $300 US) that worked in both DOS
and Windows. Once I got a feel for the pen, it recognized all of my
characters correctly (of course, I write like a 3rd grader), though it
took one or two seconds for each. Although one can not match the speed
of keyboard input (assuming one is skilled at the latter), it would
be less of an impediment to creative/thoughtful writing in Chinese.
The pen-input system is called the "PostMouse," it plugs in to the serial
port, and the company is A-Way Technology Co. Ltd, located in Taipei
on Hoping E. Rd. Sec.1, #10-1, 8th Fl.  Tel. is 886-2-368-6640,
FAX 886-2-362-0375, Modem 886-2-363-3666.

        As for the "intelligent" Pinyin based input systems for DOS &
Windows, Hanin (Han4 Yin1) is supposed to be very good (they also have
a MAC version), and Zhen4Han4 (who make a Chinese system very similar
to that of ETEN) has "Guo2 Yin1"  (DOS version @ $200 NT, Windows @ $500)
which they claim is more "intelligent" than Hanin. It was supposed to
recognize characters in context of the sentence. I haven't tried either,
so I really can't offer a judgment.
                                Univ. of Washington
                                via Taiwan U., Taipei
                                a1101006@ccms.ntu.edu.tw
                                        (til 8/29)

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Aug 27 18:10:19 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA03804; Fri, 27 Aug 93 18:10:10 CDT
Message-Id: <9308272310.AA03804@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 3953; Fri, 27 Aug 93 17:40:19 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 7780; Fri, 27 Aug 1993 17:40:18 -0400
Date:         Fri, 27 Aug 1993 22:26:38 +0800
Reply-To: QZHANG%HKUCC.BITNET@uga.cc.uga.edu
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: QZHANG%HKUCC.BITNET@uga.cc.uga.edu
Subject:      NEW WU-BI software
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

~{8wN;MxSQ~}:

    ~{NRSPR;LW4S9zDZ4x@45DNe1JWVPN#(;r3FN*MuBk#1#.#0#)#,~}~
~{4SMxIO5CV*SPPm6`MxSQ4SMxIO@{SCNe1JWVPNJdHkVPND#,~}  ~
~{Ub8vPB5DHm<~1HRTG05D0f1>SPAKPm6`5D8D=x!#K|?IRT6/L,6(Re~}~
~{4JWi#,R2>MJG2;1XMK3v1`<-W4L,>M?IRT=xPP6(Re4JWi5D9$Ww#,~}~
~{:M4SFAD;IOOVSP5D4JWiVPQ!VPKy6(Re5D4JWi=xPP6(Re!##(AmMb~}~
~{;9SP<FKcFw!"6/L,<|EL5HFdK|R;P)8(Vz9&D\!#V;R*Wn:sTZ=aJx~}~
~{9$WwV.G0#,V4PP#W#J#K#C#H#.#C#O#MUb8vND<~#,~} ~{>M?I=+Ky~}~
~{6(Re5D4JWi1#AtOB@4!#~} ~{?IRT<+4s5XLa8_JdHkVPND5DKY6H#,~}~
 ~{SHFd6TSZDG<8N;JdHkP!K55DMxSQ8|;aSPKy0oVz!#~} ~{5+1>HK~}~
~{8U8U<SHk5=MxBgIO@4#,2;V*UbVV7GILR5PT5D=;AwHm<~JG7q?IRT~}~
~{<0Hg:N=+Hm<~KM5=#I#F#C#S#S#.#O#R#G!#Gk8wN;MxSQV85cCT=r~}~
~{!#~}

    ~{AmMb#,2;V*DDN;MxSQSP=pI=VPNDO5M3#(#S#U#P#E#R~}  ~
~{#C#C#D#O#S#)5D8_5cUs71LeWV?b#,;r#6#45cUs5D71LeWV?b#,Ht~}~
~{SPJG7q?IRT#C#O#P#YR;7]!#~}

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Sep  2 14:30:32 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA12646; Thu, 2 Sep 93 14:30:28 CDT
Message-Id: <9309021930.AA12646@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 5782; Thu, 02 Sep 93 15:28:47 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 3288; Thu, 2 Sep 1993 15:28:45 -0400
Date:         Thu, 2 Sep 1993 14:50:49 -0400
Reply-To: Wang Wei <wang@CPS201.CPS.CMICH.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Wang Wei <wang@CPS201.CPS.CMICH.EDU>
Subject:      Re: Do wtalk support Pinyin or WuBi
Comments: To: YIGEN ZENG <Yigen.Zeng%MOP.LUTH.SE@CMUVM.CSV.CMICH.EDU>
Comments: cc: ccnet-l <ccnet-l%uga.bitnet@cmuvm.csv.cmich.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <9309021446.AA11691@cps201.cps.cmich.edu>
Status: RO

Hi, Edward,

I'm the author of wTalk and Way. Thanks for your interest in my work.
Sorry that I post this answer to CCNET-L, if anyone consider this of no
interest.

> Does any one know if WTALK (based on WAY3100) supports any kind of
> chinese input method, e.g. Pinyin, WuBiZiXin etc. ?
>
> For multi-window editing purpose, WTALK is nice.

wTalk does not support Chinese input methods, it uses them. The Chinese
input is supported by Way. In the version 3.1.00, I've included CiHui
(PinYin phrase), WuBiZiXing (as WuBi.dic), CangJie (in CangJie.dic),
QianMa (in QianMa.dic), Telegraph Code (Telecode.dic) input dictionaries
in Way. These are practically usable dictionaries. The others
(YingHan.dic and DeHan.dic) are examples to show how a user can build
his/her own input dictionaries with Way. If anyone out there have data to
build input methods like Cantonnese, or a real English-Chinese input
method, I believe that they are welcome to compile it with Way and upload
to ifcss.org or CND.org.

Although only I have developed wTalk and Tian BoGang have done ChPad
based on Way, but I believe there are a lot more things could be done
using Way, like a Chinese card program. That's why I released it as a
shareware without charge. Only wTalk requires a small registration fee of
$10.

> For viewing the GB document larger than 64 K, it is better than ChPad.
> I don't think any one will make use of ChPad to read HXWZ, since the
> maximal window length can not display the full length of chinese
> characters (40 char). It was strong that the author did not improve
> this "bug".

It's true that wTalk can handle large files. I tried to scan the CiHui
dictionary item list with it, which was about 570K. So I advertised wTalk
as can handle files larger than 640K :-) Though I never tried it, I
believe it works. of couse you need to have that kind of RAM. Also I like
features in ChPad, like the setting of text and background colors, and
snapshot taking of the text displayed in its window, which you can not do
with wTalk.

There are some other things I'd like to take this chance to explain. The
printing function in wTalk does not work for now. I'm working on it, but
it will take a while to come out. Other functions, such as the
communications stuff, also do not work yet. I was in a hurry to release
it before the new semester started, so I forgot to disable these menu
items. I apologize for any confusion caused.

Since I didn't have time to write an online help file for wTalk, you have
to try yourself to find out what it can do. Things like inserting and
deleting characters are pretty obvious. The "simple text clip" function I
mentioned in the announcement means that you can mark a text section and
then cut & paste. Drag-n-drop and command line file name are also
supported, which means that you can drag a file from the File Manager and
drop it into wTalk, or click on a file to invoke wTalk with that file
open in it. Again, of course you need to link that file's extention with
wTalk in the File Manager.

Thanks for the attention. Hope you and others like them.

Wang Wei  [MS CS, Central Michigan University
          [1532 S.Mission, Apt.209S, Mt.Pleasant, MI 48858 (517-773-1656)

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Sat Sep  4 17:18:12 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA16783; Sat, 4 Sep 93 17:18:01 CDT
Message-Id: <9309042218.AA16783@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7126; Sat, 04 Sep 93 18:17:08 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 3480; Sat, 4 Sep 1993 18:17:07 -0400
Date:         Sat, 4 Sep 1993 23:04:00 GMT0BST-1
Reply-To: jdelacour@cix.compulink.co.uk
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: John Delacour <jdelacour@CIX.COMPULINK.CO.UK>
Subject:      Thoughts
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

A few weeks frequenting this list have made some things very clear to me, and -
in the whole discussion of Chinese computing, certain things seem to stand out-
 and be worthy of discussion.

In recent weeks I consider the contributions of Christian Wittern, Ng Liangshi-
ng and of Edmund Lai to be of most significance, but I feel there is a long wa-
y to go and that at the moment "Chinese Computing" is rather a hobby than a se-
rious pursuit.

To begin with, I quote a posting from Wang Jianxiong of 25 August:

"Dear ccnetters who are not from mainland China:
"     I have a general question for all of you who are not from mainland
"China. Among about 30 Chinese character input methods, which one
"do you use most frequently?...... Chinese from mainland China use the "PinYin-
+LianXian.....Is it worth for us do to a internal survey of this "topic....I t-
hink it's worth to know what method we, ccnetters, use."

Wang's message raises an important question.  It is quite clearly absurd that -
there should exist 30 Chinese input methods and it is very important that as s-
oon as possible this number should be reduced to a minimum - otherwise creator-
s of software for Chinese will be lost in a wilderness.

I have already proposed my *chinput* (two Roman characters per Hanzi) system, -
but I do not pretend that that is the only way.  From the point of view of pho-
netic input it is unarguably the most efficient for Mandarin, but it does not -
solve all the problems; for one thing not everyone is a Peking mandarin and th-
ere are many people who can picture a character very well without knowing exac-
tly how it is pronounced in Peking.  Some reference to the shape (xing2) of th-
e character should therefore be available, in my opinion, in any computer "dic-
tionary", especially if we are to include characters that are not in common us-
age and which have no fixed pronunciation in the Peking dialect.

These are just thoughts, so please don't think that I am advocating one thing -
or another - I'm just trying to sort things out in my own mind while, perhaps,-
 sowing seeds in the minds of people more competent than I.

>From the point of view of input, it is obvious that two things are of prime im-
portance: 1) fewest keystrokes  2) Least reliance on the user's memory.  If th-
e user wants to write "window" it is obviously better that he should type xy1 -
than type chuang1, and it is even less convenient if he as to learn by memory -
that chuang1 is E4A2, or whatever, in hexadecimal.  The primary pointer to a c-
haracter _must_ be phonetic and the shorter that pointer is the better.

But there are going to be characters that we can see and not pronounce and we -
are going to have to find the character by its shape.  In my experience the be-
st system for this is the "four corner" system.  This system has had a long li-
fe and is still alive today; it is easy to learn and would be very adaptable t-
o computer input systems.  To me it is absurd to think of continuing with any -
radical/phonetic searching method in these days.

The point that Ed Lai made a few weeks ago was that Chinese Computing should n-
ot be equated with Chinese word-processing, and I am sure this is a very impor-
tant statement.  Whatever standard is finally arrived at in terms of codificat-
ion etc.  must have a long future.  The fact that we can now correspond in Chi-
nese via Internet does not mean that a solution has been found for Chinese com-
puting.  There is a lot of work yet to be done and a lot of questions yet to b-
e asked and answered about the codification of Chinese and all sorts of things-


One thing that does seem clear to me is that a great deal more discussion need-
s to take place with a view to establishing the best possible standard within -
the shortest possible time and that whatever standard is eventually establishe-
d should not be at all esoteric but easily used by ordinary people.

                    John

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Sun Sep  5 08:46:28 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA20430; Sun, 5 Sep 93 08:46:27 CDT
Message-Id: <9309051346.AA20430@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 8792; Sun, 05 Sep 93 09:45:25 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 8799; Sun, 5 Sep 1993 09:45:24 -0400
Date:         Sun, 5 Sep 1993 15:44:15 +0200
Reply-To: ingar.holst@ADMIN.UIO.NO
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ingar Holst <ingar.holst@ADMIN.UIO.NO>
Subject:      John's thoughts
Comments: To: ccnet-l@uga.bitnet
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Dear Netters,
   John writes that there must be a method for writing cc that we cannot
necessarily pronounce, albeit picture.
   There is already a system for doinf this; it is called the 214 radicals.
I have worked with Ni Hongbo for the last half year to implement this in
NJStar version 3.0. When it comes out, please take a look at it.
   In my view, only *2* input methods are required for any Chinese word
processor:

1.) Pinyin

2.) 214 radicals

Regards,

Ingar

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Sun Sep  5 11:12:53 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA21171; Sun, 5 Sep 93 11:12:51 CDT
Message-Id: <9309051612.AA21171@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9049; Sun, 05 Sep 93 12:12:02 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 1191; Sun, 5 Sep 1993 12:12:01 -0400
Date:         Sun, 5 Sep 1993 11:11:11 -0500
Reply-To: yuan tzeng <t90yuan@MP.CS.NIU.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: yuan tzeng <t90yuan@MP.CS.NIU.EDU>
Subject:      Re: John's thoughts
Comments: To: ingar.holst@ADMIN.UIO.NO
Comments: cc: ccnet-l@uga.bitnet
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

>    In my view, only *2* input methods are required for any Chinese word
> processor:
> 1.) Pinyin
> 2.) 214 radicals

   No. This is simly not ture. Bopomofo is one of the  cases. Actually
many "inteligent" Bopomofo imput method can read-time correct a string
of single Chinese characters into a meaningful phrase in case you typed
them wrong. A good example can be seen on ETen's WanSin input method
although it is far from perfect.

  PC Magazine Chinese Edition, August issue introduced a freeware
program called PH.EXE. This is also a good tool to make Bopomofo fly :-)

  Note that for many people who know nothing about English, the traditional
Bopomofo is just a such meaningful input tool. For these people, Bopomofo
is probably the only input method that they can use by nature (no need to
learn, and without extra equipments such as hand-write input board). In
such a case Pinyin doesn't make sense for these people.

  For your information, 2 great books, "Internet-Tools" (ISBN: 957-23-013-4)
& "Linux/Unix user's manual for Taiwanese" (still on-line) were completed
by the help of Bopomofo input method. Maybe you can imagine that such
an improved Bopomofo can reach a great input rate as well as correct rate.
(There are about n * 100,000 characters composed in these 2 books where
  1 < n < 20   :-)

--yuan


> Regards,
> Ingar

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Mon Sep  6 04:39:10 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA27040; Mon, 6 Sep 93 04:39:03 CDT
Message-Id: <9309060939.AA27040@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 8922; Mon, 06 Sep 93 05:37:47 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 4056; Mon, 6 Sep 1993 05:37:46 -0400
Date:         Mon, 6 Sep 1993 11:36:35 +0200
Reply-To: HU Bo <ucc02au@SUNMAIL.LRZ-MUENCHEN.DE>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: HU Bo <ucc02au@SUNMAIL.LRZ-MUENCHEN.DE>
Subject:      Input methods
Comments: To: ccnet-l <ccnet-l@uga.bitnet>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Christian W. said:
> People tend to use what the like...
> ... There are more than 30 fonts...
> ... why not give them the chance to [use the input method they like]..

You were right if each of these input methods were complete. The reality
is that most of them can only be used for a special set of Chinese characters.
Projected in the situation of the font resource, it is as if you had three
of four fonts, there were upper case letters in SOME of them, and lower case
letters in SOME OTHERS. In the FIFTH one, you had German Umlauts only.
You had then to compose a text of some 100 pages using these fonts...

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Sep  7 09:58:05 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA08473; Tue, 7 Sep 93 09:58:03 CDT
Message-Id: <9309071458.AA08473@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7218; Tue, 07 Sep 93 10:56:43 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 6018; Tue, 7 Sep 1993 10:56:41 -0400
Date:         Tue, 7 Sep 1993 22:44:37 EST
Reply-To: Curtis Smith <NTNUS239%TWNMOE10.BITNET@uga.cc.uga.edu>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Curtis Smith <NTNUS239%TWNMOE10.BITNET@uga.cc.uga.edu>
Subject:      Input methods
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Hello. I am new to this discussion, and only this evening read my first
mails. I noticed that when discussing input methods, everyone was talking
about phonetic romanizations or phonetic symbols (bopomofo). I am currently
at the National Taiwan Normal University working on my PhD, and I have
been using the hang-lieh input on E-Ten. The hang-lieh (column-row) method
is based on character strokes, but unlike the tsang-chieh method, makes
logical sence. All characters are input in no more than four key-strokes,
and overlap of input codes is very low. The lack of choosing from an end-
less list of homonyms (such as "yi"), along with the rational and intuitive
stroke-method makes this , to me, the ideal methosd for writing not only
in classical, but also colloquial Chinese.
   Please excuse me if this is an out-dated note. And also for the typoes (I
have no back-space or deletes on this terminal).

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Sep  7 22:27:41 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA16175; Tue, 7 Sep 93 22:27:39 CDT
Message-Id: <9309080327.AA16175@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2031; Tue, 07 Sep 93 23:26:46 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 4848; Tue, 7 Sep 1993 23:26:45 -0400
Date:         Wed, 8 Sep 1993 12:02:17 JST
Reply-To: Xiaodan Tang <tang@LIFEBOAT.CO.JP>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Xiaodan Tang <tang@LIFEBOAT.CO.JP>
Subject:      Input methods
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

   >Hello. I am new to this discussion, and only this evening read my first
   >mails. I noticed that when discussing input methods, everyone was talking
.......

I am alos want to say something about Chinese Input Method.

Normally, chinese word input method can be devided into 2 kinds. Input the
word by how to pronounce or by how to write. There are both benifit because
some word you just know how to pronounce but do not clearly know how to write
it, but you use Input Method like PinYin, you can input it correctly.

By the other hand, when you input a novel, you can see how the word is wrote,
but you do not know how to pronounce it, but if you use Input Method like
WBZX, you can still input it.

To say one input method is good or no good, we are interesting in two things,
How quick the word can be find? How easy can the method be remembered?
PinYin is easy to remember but slow to input. When I in Shanghai University of
Technology, I have tried Qian-Code Input Method. It is quick in inputing
Chinese Word, (1 word 2.3 key press average) but it is very diffcult for
remembering.

I am in Japan and there is only one Method to Input Japanese Kanji. (Many
methodes are submit, but just different in how to start/end the method,
how to change the word and so on). Japanese Input Kanji by their pronounce.
One can input a very long sentence, and then press CHANGE key, the method
will check the break point of sentence, change it word by word correctly.
The input method is also know the Grammatic of Japanese Language, so it
can change correctly in case the change depends on Gramma.

I have heard some one made  a system Named Cwnn(Wnn is a Japanese Kanji
Input Server), which not change Chinese word by word, but you just input
a sentence in PinYin, press a CHANGE KEY (like Ctrl-N), and then get the
chinese sentence. I have not seen it but it sounds not too bad.


Xiaodan Tang                            Tel   : (81)-3-3293-4717
OpenSystem Division                     Fax   : (81)-3-3293-4725
LIFEBOAT, Inc. Tokyo, Japan             E-mail: tang@lifeboat.co.jp

From @PUCC.PRINCETON.EDU:NTNUS239@TWNMOE10.BITNET Thu Sep  9 12:34:35 1993
Return-Path: <@PUCC.PRINCETON.EDU:NTNUS239@TWNMOE10.BITNET>
Received: from pucc.Princeton.EDU by ifcss.org (4.1/IFCSS-Mailer)
	id AA03607; Thu, 9 Sep 93 12:34:30 CDT
Message-Id: <9309091734.AA03607@ifcss.org>
Received: from PUCC.PRINCETON.EDU by pucc.Princeton.EDU (IBM VM SMTP V2R2)
   with BSMTP id 9856; Thu, 09 Sep 93 13:32:28 EDT
Received: from TWNMOE10.edu.tw (NJE origin MAILER@TWNMOE10) by PUCC.PRINCETON.EDU (LMail V1.1d/1.7f) with BSMTP id 5777; Thu, 9 Sep 1993 13:32:28 -0400
Received: by TWNMOE10 (Mailer R2.10 ptf000) id 2452;
          Fri, 10 Sep 93 00:44:05 EST
Date:         Fri, 10 Sep 93 00:01:25 EST
From: Curtis Smith <NTNUS239%TWNMOE10.BITNET@pucc.Princeton.EDU>
Subject:      Re:  Input methods
To: Yuan Jiang <jiang@ifcss.org>
In-Reply-To:  Your message of Wed, 8 Sep 93 21:36:13 CDT
Status: RO

In response to the request for more information on the column-row (hang-lieh)
input method I love so much, it was developed by E-Ten Information, the largest
Chinese system in Taiwan (as I'm sure you know). I was the standard input
method for E-Ten up untill only recently, when it was replaced by the Wang-
hsing (Forget the form) system, an auto-discriminatory BoPoMoFo system, which
now shares the lime-light with the row-column, one being RIM, and one PIM.
There are two versions, one using forty keys for input (10x4) and one thirty
keys (10x3), the later being pushed by E-Ten as a replacement. I use the former
out of habit. The keyboard is divided into ten columns numbered 1 - 10, each
representing a certain stroke (ex. 1 = horizontal stroke, 2 = pieh diagonal
stroke, ... 5 = downward hook to the right, as in the third stroke of wu3, 5,
etc.) One the keyboard, the 1 col. is the "1", "q", "a", & "z" column, etc.
The "pure," or single stroke row is the a - l, or second to bottom row. These
keys represent only that stroke. The second row is the q - p row, which repre-
sents the stroke for that column plus a "1" or "2" stroke. The third (1 - 0)
row is the col. stroke plus a "3", "4", "5", or "6" stroke, and the fourth (z-
/) row is the col. stroke plus a "7", "8", "9", or "0" stroke. The keyboard
is thus layed out so that when you type a character, you type the stroke or
two-stroke combinations of the character as it would be written. The entire
character must be entered within four strokes, so for some characters, you
do have no choice but to do some abreviations, as in the ts'ang-chieh system.
All in all, though, it is very intuitive (as much so as writing Chinese is
to you). The key-strokes are done in a 2-d array of sorts, and there are abre-
viations, (ie the 1:3 key, or the "1" key) can represent the character for
dragon, as it begins with a "1"-stroke, has a lot of stuff, and ends in a "3"-
stroke.
I'm sorry if this is not clear, but I was cut off twice for half an hour, and
only by pure luck was able to finish this letter. I you have any questions, ple
ase write.

Curtis Smith

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Sep  9 16:11:09 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA06669; Thu, 9 Sep 93 16:10:54 CDT
Message-Id: <9309092110.AA06669@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 0389; Thu, 09 Sep 93 17:09:11 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 3671; Thu, 9 Sep 1993 17:09:10 -0400
Date:         Thu, 9 Sep 1993 23:08:13 +0200
Reply-To: ingar.holst@ADMIN.UIO.NO
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ingar Holst <ingar.holst@ADMIN.UIO.NO>
Subject:      Wu input
Comments: To: ccnet-l@uga.bitnet
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Dear netters,
   As a side-remark to the ongoing debate on input methods, I might add
that I am planning to make available (long term project though) for NJSTAR
version 3.0 an input dictionary for the Wu dialect. I intend to make use
of some high ASCII characters, like the Norwegian 'oe' (sorry I can't
get it through on UNIX) in sounds like

PAPAMOE   ("very full" / "full to the rim")

I also plan to incorporate the Guoyu 'Q' in sounds like

GOQINJESE ("rubbish")

And I plan to indicate the long low flat tone by an underscore, like in

SA_NGHENIN, ALAGE NGABO SA_NGHENIN DEWALA.

("Shangainese, my grand-mother is Shanghainese")

Any thoughts from you netters on my project is welcome.

Regards,

Ingar

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Sep 10 03:52:46 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA12618; Fri, 10 Sep 93 03:52:37 CDT
Message-Id: <9309100852.AA12618@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9175; Fri, 10 Sep 93 04:51:42 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 8331; Fri, 10 Sep 1993 04:51:40 -0400
Date:         Fri, 10 Sep 1993 08:18:11 MYT
Reply-To: M H Foo <foo@MHFOO.PC.MY>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: M H Foo <foo@MHFOO.PC.MY>
Subject:      Input - we are almost there
Comments: To: CCNET <CCNET-L@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

On Fri, 10 Sep 1993 00:01:00 GMT0BST-1, "John Delacour" <jdelacour@CIX.COMPULINK.CO.UK> wrote:
> Dear Netters,

> alphabet, I  could achieve 97% success in three days in teaching a two
> roman-character phonetic input system.

In the Philippines, Indonesia and France, this would be an
achievement. But, in Malaysia, Singapore and the PRC, where PinYin
(luo2ma3 PinYin) is being taught from year 1 -> 3, phonetic inputs and
it's variations (meaning complete PinYin or simplified Pinyin viz
jian3pin1/shuang1pin1) is no longer a problem. The PRESENT practical
drawbacks of Jian3pin1 are:

1. Various systems have different keys to represent the vowels and
   consonants. xin4tong1 <> ZRM <> WPS <> Bei3da4 (CStar) <> John's
   method.
2. Complete avoidance of the use of radicals would still end up with too
   many repeats for SINGLE words.

> As to radical-based input, I'm very doubtful if it is preferable to the
> four-corner system but I hope I can start an argument on this subject,

Four-corner system is a good RBM (to input words you can't pronounce)
because:

1. The radicals are COMMON to those use in many dictionaries
2. Not stroke dependent
3. Repeats (chong2 ma3) is LOW (much lower than WBH and ONLY slightly
   inferior to cangjie, WBZX and da4 yi4)
4. The BASIC radicals to remember are few -> about 10 (WBZX = 100)

BUT there is no wide base of users of this input to provide
guidelines.

> Besides 214 is a lot of things to remember.

ZHENG4 MA3 is a good example. It is similiar to WBZX with MORE
radicals. Possibly interesting for the few deligent scholars. But for
the MAJORITY of my pupils (and probably your own CHILDREN as well),
trying to master the radicals (and the use of them) would require
exceptional qualities.

WHAT THEN IS THE SOLUTION TO RADICAL-BASED INPUT METHODS ?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
IDEALLY, a COMMON set of radicals should be taught for BOTH the
dictionary and the computer. Teach the students these radicals in the
primary schools - initially for dictionary use. Subsequent input of
words based on the SAME radicals would be effortless. The PRESENT
situation of having students learn one set of radicals for the
dictionary and a completely DIFFERENT set of radicals for the computer
tends to create confusion.

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Sep 10 05:05:58 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA12983; Fri, 10 Sep 93 05:05:47 CDT
Message-Id: <9309101005.AA12983@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 3099; Fri, 10 Sep 93 06:04:26 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 1136; Fri, 10 Sep 1993 06:04:24 -0400
Date:         Fri, 10 Sep 1993 12:03:29 +0200
Reply-To: HU Bo <ucc02au@SUNMAIL.LRZ-MUENCHEN.DE>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: HU Bo <ucc02au@SUNMAIL.LRZ-MUENCHEN.DE>
Subject:      input methods and dictionaries
Comments: To: ccnet-l <ccnet-l@uga.bitnet>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Forwarded message:
>From hubo@ift.fhg.de Fri Sep 10 11:30:42 1993
Message-Id: <9309101606.AA29774@hubo.ift.fhg.de>
To: ccnet-l@uga.cc.uga.edu, ucc02au@sun3.lrz-muenchen.de
Subject: input methods and dictionaries
Date: Fri, 10 Sep 93 11:06:26 CDT
From: "HU Bo, IFT Munich, +49 89 54759129" <hubo@ift.fhg.de>


There are many signifficant differences between the requirements
needed in using a Chinese dictionary and Chinese input in the
computer.

The most obvious one is that we are not really unhappy
if we find a character in a dictionary in one minute. But no input
method can be accepted if it enables an input rate of 60 characters
in an hour!

Another difference is that we never mind if we found that there are
several, sometimes hundreds of characters listed under the same item
(for example, Mu4(wood), 6 strokes). But it is quite inefficient to
build an input method based on it.

Talking about the 214 radical method, the main problem for me is to
input the redical in some intuitive way. I used a CCDOS with radical
input. You have to enter a number (1-214) for a radical. Which number
then, has the radical Fish?

The input rate will never higher than 100/h if you have to count the
strokes for each character! (H,P,S,H,O,H)

..

Gernerally spoken, the input method must be more intuitive, and the
search method in a dictionary must be systematic. And: last but not
least: it is a plus in a dictionary, if there are more registers
helping you to find a character. It is quite frustrating if a user
has to switch between several input methods, trying to input a single
character!

Specially spoken, I don't think WBZX is a good search method, nor
214 radical is a good input method (or somebody tells me how I can
input the radical FISH and count the strokes of the rest within
several seconds!).


Bo

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Sep 10 11:09:30 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA16326; Fri, 10 Sep 93 11:09:26 CDT
Message-Id: <9309101609.AA16326@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2442; Fri, 10 Sep 93 12:08:28 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 1501; Fri, 10 Sep 1993 12:08:25 -0400
Date:         Fri, 10 Sep 1993 23:45:17 MYT
Reply-To: M H Foo <foo@MHFOO.PC.MY>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: M H Foo <foo@MHFOO.PC.MY>
Subject:      Dictionary and Input
Comments: To: CCNET <CCNET-L@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

On Fri, 10 Sep 1993 12:03:29 +0200, "HU Bo" <ucc02au@SUNMAIL.LRZ-MUENCHEN.DE> wrote:
> Forwarded message:
> >From hubo@ift.fhg.de Fri Sep 10 11:30:42 1993
> Message-Id: <9309101606.AA29774@hubo.ift.fhg.de>
> To: ccnet-l@uga.cc.uga.edu, ucc02au@sun3.lrz-muenchen.de
> Subject: input methods and dictionaries
> Date: Fri, 10 Sep 93 11:06:26 CDT
> From: "HU Bo, IFT Munich, +49 89 54759129" <hubo@ift.fhg.de>
>
> There are many signifficant differences between the requirements
> needed in using a Chinese dictionary and Chinese input in the

Thank you for your comments.

> The most obvious one is that we are not really unhappy
> if we find a character in a dictionary in one minute. But no input
> method can be accepted if it enables an input rate of 60 characters
> in an hour!

Sorry for the ambiguity arising from my language. There IS a
similiarity in the sense that to be WELL-versed in the dictionary
usage or input one should know BOTH a phonetic (PBM) as well as a
radical method (RBM). The quest for SPEED is also similiar for both
dictionary and input.

> Specially spoken, I don't think WBZX is a good search method, nor
                      ^^^^^
A very interesting point indeed. Our experience is only limited to our
4 yrs experimentation with the TEACHING and USE of WBZX in some
Malaysian schools (teachers and students). We would like to briefly
share our experience.

There are 100 basic WBZX radicals (in fact less) to learn. It is easy
to comfortably teach a 11 - > 12 yr old the radicals within 14 days
(max) assumming his vocab level is about 1000 words and is able to
write Chinese in the correct sequence stroke. The repeat rate (chong2
ma3) is extremely low and therefore as a dictionary method is
EXCELLENT. The fact that WBZX has never been incorporated in
dictionary use probably contributed to HU BO's prejudice. We
have published a word list indexed on WBZX. To search for a word
(whose pronunciation is unknown), one needs only to know the first 2
WBZX keys and the number of words to search for would be almost equal
to that of the best radical based method currently available
(bu4shou3/bi3hua4/4-corner/etc). If the user is able to provide the
first 3 WBZX keys, the search would be reduced to about 3 words only.

Speed-wise, WBZX is more than acceptable in both dictionary and input.
But the biggest BONUS is that our students need only to master a set
of COMMON radicals within 14 days and they would be able to use the
dictionary to search and the computer to "write" - in both cases very
efficiently.

Incidently, to look for "yu2 - FISH", my students need only to search
under or input 'qgf' and there is only ONE word (yu2).

More comments please !!

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Sep 23 05:23:40 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA18547; Thu, 23 Sep 93 05:23:35 CDT
Message-Id: <9309231023.AA18547@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6543; Thu, 23 Sep 93 06:22:12 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 3087; Thu, 23 Sep 1993 06:22:10 -0400
Date:         Thu, 23 Sep 1993 10:16:19 +0200
Reply-To: HU Bo <ucc02au@SUNMAIL.LRZ-MUENCHEN.DE>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: HU Bo <ucc02au@SUNMAIL.LRZ-MUENCHEN.DE>
Subject:      Ming Bu Zheng, Yan Bu Shun
Comments: To: Chinese Network <ccnet-l@uga.bitnet>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Without a correct definition of terms, one can not talk fluently.
                                                - Confucius
C{2;U}#,QT2;K3
                                                #- ?WWS


I read every day on this net many terms, without a correct definition.
In following I am trying to list some definitions I found.
It is the second version and for discussion.

1) Pinyin       full pinyin according PR China's State Department.
2) ShuangPin    a two (sometimes also a three) key transscript of (1).
3) ShuangYin    based on (1) or (2), possible to input two syllable words.
4) LianXiang    based on (1), and after each CHARACTER input, a menu shows
                CHARACTERS which may be a second one in a WORD together
                with the first one.
5) Cizu         similar to (3) but also useable for two and more syllable words.
6) Zhengju      based on (5) but some more intelligence is implemented for
                A) finding word borders, or/and B) finding the correct one
                (or more) syllable word form a list of homophone words using
                some correlation functions between words.
7) Tone         forced tone input.
8) Optional T.  optional tone input.
9) Bopomofo     Its official name is "Guoyu Zhuyin Fuhao". Its computer
                related implementation is done using redefinition of A-Z
                and 0-9 (and something else) to Bo, Po
                Mo, Fo, etc. Often combind with (7)
10) Graphical   Encoding of characters using their GRAPHICAL structures like
                strokes, radicals and so on.
11) Radical     ONE possiblity to implement (10) using radical information.
                Typical: 216 radical.
12) Input macro An input "method" which does not cover all characters of the
                code system for which the input method is designed.
13) Word        Ci in Chinese
14) System      A meaningless word in context of description of a software
                package, since everything is named a system by its author
                nowadays.
15) C-DOS-Extension Software package enabling Chinese IO on the level of DOS.
                VRAM direct access must be supported. Often named as xx-DOS
                or alike.
16) No-use      (15) without support of VRAM direct access.
17) C-Editor    Editor supporting Chinese
18) C-WordProcessor (17) with at least some basic WP functions like character,
                paragraph and page formatting.

Generally, there are several very problematic confusions:
        (12) with "input method"
        (11) with (10)
        (3), (4), (5)
        (5) with (6)
        (7) with (8)
        (1) with (9)
        (13) with "character"
        (14) with everything meaningfull
        (15) with (16), (17), (18)
        (17) with (18)
To reduce misunderstanding and information redundance in our discussion, we
should release an "official" definition of the terms we often use on this
network.

Regards,


Bo

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Sep 23 17:57:13 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA28660; Thu, 23 Sep 93 17:57:11 CDT
Message-Id: <9309232257.AA28660@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 3538; Thu, 23 Sep 93 18:56:29 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 7760; Thu, 23 Sep 1993 18:56:28 -0400
Date:         Thu, 23 Sep 1993 23:46:00 GMT0BST-1
Reply-To: jdelacour@cix.compulink.co.uk
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: John Delacour <jdelacour@CIX.COMPULINK.CO.UK>
Subject:      Yan bu chang - Yi bu qing
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

*******  Hubo writes, zi4bu2zheng4, yan2bu4shun1
___________

I quite agree, but I think such a concise explanation as you give, Hubo,
with so many confusing cross-references, is not adequate, though I accept
the principle of what you say.

What is needed is a file giving a complete definition of the terms used,
for example:-



__________________________BoPoMoFo__________________________________

A phonetic transcription for Chinese syllables by means of which all
Mandarin syllables may be presented by means of one, two or three special
characters consisting of from one to four brush strokes and resembling
simple Chinese characters (eg. | , X , T , << )

The name bopomofo is analagous to ABC, ie. bo, po, mo, fo are the
pronunciation of the first four characters in the "alphabet"

This phonetic transcription was devised in the years following the
National Revolution (ca 1920).

The "bopomofo" consists of 21 (formerly 24) characters.

Also called     National Phonetic Alphabet
                Zhuyin fuhao

Other comments:- ..............

____________________________

I'm sure it would not be too difficult for a few people to compile a
useful "dictionary" of Chinese (computing) terms, which could reside in
ifcss.org and be added to and revised as necessary for general reference.

                    John

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Oct  6 06:01:55 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA01960; Wed, 6 Oct 93 06:01:53 CDT
Message-Id: <9310061101.AA01960@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 0621; Wed, 06 Oct 93 06:59:49 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 6170; Wed, 6 Oct 1993 06:59:48 -0400
Date:         Wed, 6 Oct 1993 19:58:37 JST
Reply-To: christian wittern <g53150@SAKURA.KUDPC.KYOTO-U.AC.JP>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: christian wittern <g53150@SAKURA.KUDPC.KYOTO-U.AC.JP>
Subject:      Liu's numbers for Chinese phonetics
Comments: To: ccnet-l@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Cc: soft-authors@ifcss.org

Dear netters,

I recently came across a scheme for numbering the Chinese syllables,
which apperently was developped by a Professor Liu. It simply iterates
over all syllables, which are possible and used, based on the zhuyin
fuhao syllablary.
The result looks something like this [I substitute pinyin]:

0001 ba1
0002 ba2
0003 ba3
0004 ba4
0005 ba5
0006 bo1
0007 bo2
0008 bo3
0009 bo4
0010 bo5
0011 bai1
....
2050 yun5
2051 yong1
2052 yong2
2053 yong3
2054 yong4
2055 yong5

A closer examination shows, that in a few cases obsolete syllables are
included and two syllables, which are actually in use have somehow been
forgotten.

Nevertheless, after some revision, I believe this system could be used for
internally represent the input-tables of Chinese software, thus laying a
layer of abstraction between the phonetics and the actual keystrokes used for
inputting these phonetics. This would easily allow even for user-modified
inputkey-tables, given everybody the shortcuts he likes and finds convenient
on his keyboard.

One could even think of similar tables developped for reconstructions of
archaic and ancient spellings and relate them to the characters, so that
old and obsolet pronounciations could be as easily used and looked up as
todays pinyin. I believe this is a better way for representing these
phonetics than building a font of its own (and thus a code of its own) for
rhyme dictionaries like qieyun or guangyun, as somebody suggested here some
while ago.

And even one step further, a standard similar to HBF, lets say HPF ;-),
could be developped to let different software use the same table for lookup
of readings.


Greetings from Kyoto,

Christian Wittern

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Oct  6 11:08:06 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA04570; Wed, 6 Oct 93 11:08:03 CDT
Message-Id: <9310061608.AA04570@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 3152; Wed, 06 Oct 93 12:07:00 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 0469; Wed, 6 Oct 1993 12:06:58 -0400
Date:         Wed, 6 Oct 1993 09:05:31 -0700
Reply-To: Ed Lai <lai@APPLE.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ed Lai <lai@APPLE.COM>
Subject:      Re: Liu's numbers for Chinese phonetics
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

I am all in favor of a fixed width transcription system independent
representation of Chinese sound, and I had in fact tried to invent them
a few times when I worked on a HyperCard Chinese dictionary. Liu's system
can be considered one of those scheme and I had looked at it during the
early phrase of the dictionary project when I took data from the CCDB
database and put it into my dicitonary and CCDB uses Liu' system.

However in recent years I basicly just use the ShuangPin as the internal
code. It is fixed width, phonentically netural, and most important of all,
since it is already a standard, we don't have to formulate a new standard
that people has to agree on then. It has advantage over Liu code that you
can read it directly with some training and it is easier to translate
between Pinyin/BoPoMoFo/WadeGile <-> ShaungPin than Pinyin/BoPoMoFo/WadeGile
<-> Liu code.

Of course there are still a few detail to be standardized, there are a
few sounds in GB code like Hmg that does not seem to be in ShuangPin so
I have to invent a code for it. ShuangPin is three byte long, we need a
two byte version (it is very easy, just take 5 bits + 5 bits + 3 bits,
so just to have to agree on how to put these bits). Also when I use it
with a prefix of "B" so that in future we can use other prefix for
Cantonese and other dialects.

/* Disclaimer: All statments and opinions expressed are my own */
/* Edmund K. Lai                                               */
/* Apple Computer, MS37-AS                                     */
/* 20525 Mariani Ave,                                          */
/* Cupertino, CA 95014                                         */
/* (408)974-6272                                               */
zW@h9cOi

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Thu Oct  7 00:06:06 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA14620; Thu, 7 Oct 93 00:06:02 CDT
Message-Id: <9310070506.AA14620@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 8369; Thu, 07 Oct 93 01:03:46 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 3346; Thu, 7 Oct 1993 01:03:45 -0400
Date:         Thu, 7 Oct 1993 00:01:00 -600
Reply-To: Stanley G Goertzen <goertzen@BADLANDS.NODAK.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Stanley G Goertzen <goertzen@BADLANDS.NODAK.EDU>
Subject:      Re Liu's numbers for Chinese phonetics
Comments: To: ccnet-l@uga.cc.uga.edu
Comments: cc: g53150@SAKURA.KUDPC.KYOTO-U.AC.JP
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Hello CCNETers,

Christian Wittern <g53150@SAKURA.KUDPC.KYOTO-U.AC.JP> wrote to
ccnet-l@uga.cc.uga.edu  on Wed, 6 Oct  1993 19:58:37 JST, regarding input
tables for Chinese software based on Professor Liu's numbering scheme:

>0001 ba1
>0002 ba2
>0003 ba3
>0004 ba4
>0005 ba5
>0006 bo1
>0007 bo2
>0008 bo3
>0009 bo4
>0010 bo5
>0011 bai1
>....
>2050 yun5
>2051 yong1
>2052 yong2
>2053 yong3
>2054 yong4
>2055 yong5
>
>...after  some  revision, I  believe  this  system  could  be  used  for
>internally represent the input-tables of Chinese software, thus laying a
>layer of abstraction  between the  phonetics and  the actual  keystrokes
>used for  inputting these phonetics.  This would easily  allow even  for
>user-modified inputkey-tables, given  everybody the  shortcuts he  likes
>and finds convenient on his keyboard.

If I understand the logic here,  ba1 is associated with 0001 and  0001 is
associated with all the Hanzi pronounced ba1, so that

ba1 -> 0001 -> all Hanzi pronounced ba1

but if I  wished to us Wade-Giles, all I  would have to do is  change the
inputkey-table so that

pa1 -> 0001

and the association XXXX -> Hanzi would remain constant.

>One  could even think of similar tables developed for reconstructions of
>archaic and ancient spellings and relate them to the characters, so that
>old and obsolete pronunciations could be as easily used and looked up as
>todays pinyin.  I believe this  is a better  way for  representing these
>phonetics than building a  font of its own (and thus  a code of its own)
>for rhyme  dictionaries like qieyun  or guangyun, as  somebody suggested
>here some while ago.

If I  understand correctly, there are a total of 2055  syllable types  in
Mandarin  Chinese, beginning with 0001, ba1, and ending with 2055, yong5.
That seems high, but tone 5 (which I'm guessing might be associated  with
entering tone words?) might be the reason for this.

By some estimates,  the Guangyun has  about 3875  syllable types and  the
number of unique archaic  syllables may be higher.  If  I wanted to input
ancient  (Guangyun) or archaic  Chinese I'd probably need  an input table
with over 4000  numbers, since many syllables which are  homophones today
were pronounced differently in the past.  For example:

Mandarin  Ancient   Archaic   Definition
--------  -------   -------   ------------------------------
shi2      zjy@p     djy@p     ten; decade.  ~{J.~}
shi2      dzhjy@k   dhjy@k    eat; attrition, eclipse. ~{J3~}
shi2      zjyEEk    djyak     rock, stone.   ~{J/~}
shi3      sji       sjy@r     arrow.  ~{J'~}
shi3      sji@      sjyaar    swine, pig.  ~{u9~}
shi3      sri       sly@g     historian, recorder; history. ~{J7~}
shi4      dzhji     dhjy@r    signify, show, indicate. ~{J>~}
shi4      zi        djy@r     look, see, regard.   ~{JS~}
shi4      zji@      djyaar    tribe, clan.   ~{JO~}
shi4      zji@      djyeg     this; to be right.  ~{JG~}
shi4      dzhri     dzhr@g    officer, gentleman, scholar.  ~{J?~}
shi4      sjy@k     sjy@k     polish, adorn, ornament. ~{JN~}

I'd need  a separate number  for each of  these words since there  are no
homophones in Archaic Chinese.  I'm not sure, though, if I'd ever want to
access  Chinese characters in  a word processor using  ancient or archaic
pronunciations.  But if I  did, I would probably revise the  sequence and
use the same order as in the Guangyun.

Still, there are over 26,000 rhyme words in the Guangyun,  more than any
current standard  (except for CCCII), so there would  still be many Hanzi
characters associated with numbers which I could not access.  There might
even be some numbers for which there are not any characters at all in the
current standards.

Maybe I didn't understand what Christian was proposing, but I don't think
it would be worthwhile to increase the numbers in the input table just so
that there would  be a unique number for each  unique ancient and archaic
syllable.   Moreover,  phonologists still disagree as to how  ancient and
archaic  Chinese   was  pronounced,  so  getting  a  consensus  might  be
impossible anyway.

Otherwise, I like the idea of allowing  the user to change from Pinyin to
Wade-Giles, or whatever input method he/she might prefer.

Regards,

Stan
goertzen@badlands.NoDak.edu

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Oct 20 09:02:10 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA06174; Wed, 20 Oct 93 09:02:09 CDT
Message-Id: <9310201402.AA06174@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 2745; Wed, 20 Oct 93 10:01:06 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 8498; Wed, 20 Oct 1993 10:01:05 -0400
Date:         Wed, 20 Oct 1993 14:16:59 BST
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Input dictionary
Comments: To: Chinese Computing Network <ccnet-l@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

>From lsn Wed Oct 20 14:16:24 1993
Date: Wed, 20 Oct 93 14:16:24 BST
To: dajunxu@EDU.USC.RCF
Subject: Re: input dictionary of NJSTAR
In-Reply-To: <no.id>; from "Dajun Xu" at Oct 19, 93 1:49 pm
X-Mailer: ELM [version 2.3 PL11]

>
>
> Anybody out there knows what format of input dictionary of NJStar or
> the files with extention '.dic'?

I do.

> I want to "customize" some of input
> definations.  Does "china-studies" archive have something to explain
> that? Thanks in advance.
>

1. First you must prepare a text file containing the definition of the
input methods for ALL characters - a dictionary text file.  Such file
can be obtain from ifcss.org:software/x-win/cxterm-dictionary.  You need
to modify the header so that I suit the format of NJStar input
dictionary.  (See INPUTDIC.DOC).  You can then 'customise' the
dictionary file before compiling by inputdic.exe.

2Then run INPUTDIC.EXE

inputdic cangjie njinputk

(extensions not required)

2. There are quite a few dictionary files in cxterm. The collection is quite
comprehensive. Have a look.

3. As for my self, I have written a utility to extract from .DIC to output
to a dictionary text file.

4. I have also reassign the keyboard for Cangjie using another utility of
mine. I have changed the Y key to X, X to Z, because Y is too frequently
used.

Anybody interested in those utilities or the modified Cangjie and
revised Cangjie (jianti Guobiao) dictionary text file, please write me
an email.

Sincerely
Liangsheng WU

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Mon Nov 22 11:20:56 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA06326; Mon, 22 Nov 93 11:20:50 CST
Message-Id: <9311221720.AA06326@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 9840; Mon, 22 Nov 93 12:18:56 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 6159; Mon, 22 Nov 1993 12:18:54 -0500
Date:         Mon, 22 Nov 1993 09:12:50 -0700
Reply-To: Rob Gimello <GIMELLO@CCIT.ARIZONA.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Rob Gimello <GIMELLO@CCIT.ARIZONA.EDU>
Subject:      Chinese Input Methods
Comments: To: CCNET-L%UGA.BITNET@VM1.NODAK.EDU
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

About one of the imput methods mentioned in Ms. Lasovick's query, viz., "Dayi."

I have not noticed much mention of this method on CCNET-L and I gather that it
is not well known in the States.  However, I was introduced to it last year in
Taipei, have been using it ever since, and recommend it highly.  In fact, some
People in Taiwan have told me it promises to become one of the most popular
systems in use there.

In general, it is a method of the same general sort as Cangjie -- i.e., it
breaks all characters into 244 basic components.

        (NOTE WELL: as is the case with Cangjie, these 244 are NOT
        quite the same as the old set of 214, or the new set of 200,
        "radicals" - although there is some overlap).

These 244 basic components are distributed among 40 keyboard keys (i.e., the 36
alpha-numeric keys (with no distinction between upper and lower case) plus
theperiod, the comma, the semi-colon and the forward-slash).

The system is very intuitive (much more so, I think, than Cangjie) and does not
take long to learn.  After only a few days of use, one will find it relatively
easy to guess which combination of keys in which order will call up any given
character, and almost all characters are called-up with only 2 or 3 key-
strokes.  I believe it's much faster than Pinyin or any of the other phonetic
systems, and of course it can be used even when you are not sure of the
pronunciation and/or tone of the character your trying to type. I don't think
it is even necessary to use a manual to learn the system; all you need is the
one-page key-board chart (and a little bit of patience).

I use this system on a Mac, but I gather that it (or some version of it) may
also be used on DOS platforms and with Windows.  I note too that it is included
among the several input options provided in the new Chinese Language Kit from
Apple (although I haven't yet tried that product, as my copy of it will not
be delivered for another few days).

One caveat: I believe Dayi is designed for use only with the standard character
set encoded in Big 5;  I doubt that there is a version of it designed for
the simplified character set encoded in Guobiao.  I would hope, however, that
some version of it -- or something like it -- will be developed for use with
all forms of Chinese, Japanese, Korean, etc. when the long-waited Unicode is
finally ready.

R. Gimello
The University of Arizona
gimello@ccit.arizona.edu

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Nov 23 00:04:57 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA19426; Tue, 23 Nov 93 00:04:43 CST
Message-Id: <9311230604.AA19426@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 7675; Tue, 23 Nov 93 01:02:46 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 0307; Tue, 23 Nov 1993 01:02:44 -0500
Date:         Mon, 22 Nov 1993 21:42:50 -0800
Reply-To: James Lin <chase@NETCOM.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: James Lin <chase@NETCOM.COM>
Subject:      Re: Chinese Input Methods
Comments: To: Rob Gimello <GIMELLO@CCIT.ARIZONA.EDU>
Comments: cc: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <199311221719.JAA07087@mail.netcom.com>
Status: RO

On Mon, 22 Nov 1993, Rob Gimello wrote:

>
> I have not noticed much mention of this method on CCNET-L and I gather that it
> is not well known in the States.  However, I was introduced to it last year in
> Taipei, have been using it ever since, and recommend it highly.  In fact, some
> People in Taiwan have told me it promises to become one of the most popular
> systems in use there.

As a matter of fact, there are at least four formal Chinese input
competitions held between mainland China and Taiwan.  Miss Ho from Taiwan
won them all with the fastest input from 129 cpm (the first competition
in Hong Kong 1989) to 171.4 cpm (Sep 18, 1993 in BeiKing).

>
> In general, it is a method of the same general sort as Cangjie -- i.e., it
> breaks all characters into 244 basic components.
>
> These 244 basic components are distributed among 40 keyboard keys (i.e., the 36
> alpha-numeric keys (with no distinction between upper and lower case) plus
> theperiod, the comma, the semi-colon and the forward-slash).

One very important principle of Dayi : it is based on writing stroke
sequence.  The maximum number of key stroke to input a Chinese character
is 4 (5 with Canghie).  The beauty of it compare to Cangjie is 1) Far
less multiple choice for characters of the same key stroke sequence. 2)
Far less characters have exceptions on composition rules.

The other good thing about it is not only Dayi implemented 13000
characters with these features mentioned above, but also it implemented
60000 frequently used Chinese phrases (from 2 characters to 7 characters)
using the same principle and it still provide very low chances to have
multiple choice.  I think they have a very good study in statistic.

>
> pronunciation and/or tone of the character your trying to type. I don't think
> it is even necessary to use a manual to learn the system; all you need is the
> one-page key-board chart (and a little bit of patience).

There is a newly revised Dayi CAI program available and it basically
provide ruling instruction, exercising games, learning tools...and many
more, which is also a very good tool to learn Dayi without a manual.

>
> I use this system on a Mac, but I gather that it (or some version of it) may
> also be used on DOS platforms and with Windows.  I note too that it is included

It is originally available in DOS environment for ETen, KC... and it has
been included in Microsoft Chinese Windows 3.1 as a built in Chinese
input method.


> One caveat: I believe Dayi is designed for use only with the standard character
> set encoded in Big 5;  I doubt that there is a version of it designed for
> the simplified character set encoded in Guobiao.

I kind of remember that one very famous Chinese DTP (HwaKuang) in
mainland China also provided Dayi as an option.  I am not sure about if
there is a commercially available package for simplified Chinese system.
Let me check with Mr. Wang in Taipei (the inventor of Dayi).


> R. Gimello
> The University of Arizona
> gimello@ccit.arizona.edu
>


James Lin

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Tue Nov 23 10:19:18 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA24387; Tue, 23 Nov 93 10:19:13 CST
Message-Id: <9311231619.AA24387@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 5110; Tue, 23 Nov 93 11:16:22 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 2835; Tue, 23 Nov 1993 11:15:14 -0500
Date:         Tue, 23 Nov 1993 14:47:08 GMT
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Guobiao DaYi
Comments: To: Chinese Computing Network <ccnet-l@uga.cc.uga.edu>
Comments: cc: "yhchen (Dr. Chen You Hor)" <yhchen@clcs.pc.my>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

The Simplified Character version for Dayi i.e. Guobiao Jianti Dayi input
table is implemented for KuoChiao in Malaysia.

The raw table can be available from Dr. Chen You Hor <chen@clcs.pc.my>
in Malaysia.

This table is built (from scratch I think) by Dr. Chen and his company
CLCS - Chinese Language Computing Society, a non-profit organisation.

I used to see it and was not interested at that time to ftp put it on
the net. You may contact him via email.

Sincerely
Liangsheng WU

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Dec  1 03:32:44 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA27897; Wed, 1 Dec 93 03:31:55 CST
Message-Id: <9312010931.AA27897@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6748; Wed, 01 Dec 93 04:29:15 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 4120; Wed, 1 Dec 1993 04:29:12 -0500
Date:         Tue, 30 Nov 1993 17:28:22 -0800
Reply-To: Don Huan <ezthundr@SFSUVAX1.SFSU.EDU>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Don Huan <ezthundr@SFSUVAX1.SFSU.EDU>
Subject:      On Chinese Input Methods (cont.)
Comments: To: John Delacour <JD@eremita.demon.co.uk>
Comments: cc: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

> One or two thoughts in response to these postings.....
> _____________________
>
> ....The best official record (to my knowledge) for International
> >> Chinese input competition is 171.4 Chinese characters per minute held by
> >> Ms. Ho Sep 18, 1992 in Peiking using Dayi.  (maybe there are newer record
> >> since then)
> >> Many many Chinese computer users could input Chinese at a rate over 100
> >> cpm right now.
> >
> >> James Li
______________________

> >        The previous post makes me think of two questions.  Records are
> >fun and I enjoy them as much as anyone, but my question is what is
> >the average person going to expect?  In other words, what is the
> >'Words Per Minute' for the average ...
> >
> > Ed Hughes
> ____________________

My thoughts exactly, Ed.  That's like saying the world's record for typing
English is something like 153 words per minute (I can't find my Guinness
Book of World's Records now to say exactly but just throwing up this number
for discussion sake) and saying that many, many people can type over a 100
wpm.  We netters know that that's a ridiculous statement, especially when
we type firsthand to participate on the Internet.  And although there are
people who can type >100 wpm (actually, I knew someone who could and her
daughter could at one time type a bit over 90 wpm) we know that these
people are not common.  Professional secretaries are employed at >65 in the
US, and office clerks at >40.  I type at about 40-45 wpm when I practice,
concentrate and have to meet deadlines, but the errors....Like Ed, what
about the average?

Citing the exceptional, placing it at the forefront and basing one's
argument on it is a faux pas.  And then to go on and say that "many, many
Chinese users can input Chinese at a rate over 100 cpms" seems overly
exaggerated and generalized.  This distorts the broader reality of the more
commonly shared problem of inputting Chinese for most people (including
Chinese, and especially for Westerners), that it is time-consuming and
cumbersome, rendering people less productive.

John offered an explanation stressing how the differences between the two
worlds in thought processes and Chinese language learning has created a
major roadblock in the quest to develop thee Chinese input method...
______________________

> The most natural input method for a "latin" person, even if he has achieved
> fluency in Chinese, is likely to be phonetically  based, but that may not
> seem natural to someone who in childhood viewed words not as concatenations
> of phonetic elements but as sequences of brush-strokes with a phonetic
> value attached.
>
> This psychological aspect of the question is in my opinion very important,
> especially as Europeans and Americans are involved in the development of
> Chinese software and input methods.  The question is whether the Westerner,
> who perhaps began learning Chinese at the age of 16 - 20 is competent to
> make judgments regarding the relative efficacy of input methods which are
> designed mainly for the use of people who began writing Chinese at the age
> of 5 or 6.
-----------------------
This is a very intelligent comment.  It's definitely a `must' consideration
for any Western developer of Chinese software and input methods.
-----------------------
> There are obviously ways of comparing various input methods in purely
> scientific terms -- eg. two keystrokes per character are better than five,
> but there are other considerations.  Two keystokes are not better than five
> if it takes three seconds to do two keystrokes and only two to do five.
> The speed of the human brain and the training of the individual's brain is
> crucial.
>
> The lady (I guess is was a lady!) who could enter 174 characters per minute
> using Dayi must have the sort of brain that thinks very graphically.  Not
> everybody thinks in that way, no matter whether they are educated in a
> Chinese environment or an alphabetic environment, so the fact that the
> "record" is held by a person using Dayi is no recommendation a priori for
> that input method.
>
>                             John
---------------------------------------
Plus the lady(?!) practiced and practiced for an international competition.
The event sounds like a sort of an "Olympic" competition for a golden
keyboard!

So theoretically, in regards to input speed and differences in Chinese
language thought processing, we consider time and the number of keystrokes
to input, time and the number of strokes to graphically input, or some
variation, right?  And whatever way is simplest for the "average" fluent
user to grasp, use and quickly input would thus probably be the method
most successful (assuming the "average fluent user" is targeted.
Well...where does the technology of handwriting recognition fit into this
discussion.  Imagine...

Handwritten chinese character recognition seems to be the logical step.
It's the most "natural" way for the average fluent user to input.  It
is not biased to cultural differences in language thought processes since
it would recognize the strokes, number of strokes and radicals.  This method
would open up an even larger user market because Chinese characters are
used by non-Mandarin speakers of other dialects, and by overseas Chinese in
other countries.  People learning Chinese as a second language could
use it and Japanese could use it with Japanese software.  (I wonder how the
Japanese are doing with this technology?).  Also, we wouldn't have to rely on
phonetic and tonal distinctions to get the character we wanted.  The pad or
screen would have a grid defining the area which the character must be
written within, much like character sheets we're familiar with used for
practice writing.  And with the software we could also customize the size of
the grid.  All the user would have to do is simply write.  We could have a
pop-up Windows menu and to edit, choose a scalable font of our liking, eg.,
Song or Grass, Simplified or Classical, etc...

To find a character, pinyin and bo-po-mo-fo could be used. Keyed in (like
now) or written in a specially designed space on the pad-screen, a
display line at the bottom/top/side of the pad-screen would give you the
set of characters from the corresponding inputted phonetic sound and tone.
All one would have to do is tap the pen on the character and that character
would occupy the square to be filled with a tap on the square.  Otherwise,
one could possibly drag the character to the square.

Of course, handwriting recognition technology would need to be further
--much further--developed, but it seems possible to do this in the
foreseeable future.  Also, people would have to write clearly and somewhat
neatly like children so that the character can be identified correctly.  I
would think that the problem that Apple's Newton, Casio's Z-7000, Sharp's
PI-7000 and the rest have in identifying handwriting would be similar
to that of identifying handwritten Chinese characters and bring slowdowns
which we originally wanted to avoid.  But with further advances, those
problems would eventually be ironed out.

Perhaps, more importantly, where "Newton" technology has failed to bring a
strong identifiable need-purpose-use for itself, the need for a better
method of Chinese character input may prove to be more compelling.  And the
many electronic-computer/software mega-corporations will not turn their
heads from such a potentially large and profitable market.  With the
question of affordability, even Apple, Casio, Toshiba and others have
made the prices of their handwriting recognizable products within reason,
so affordability of a "Chinese" version could also be brought down within
reason(?)

This discussion has gone off on a tangent from its initial query of
finding a way to protect a new idea for Chinese input.  The idea, as I
recall, stemmed from a netter who had a dream of inputting Chinese using a
new "programming skill" which he dared not share for fear of losing his
proprietary rights.  I responded by suggesting that he document everything
with time and date and get a lawyer specializing in proprietary matters and
rights if he's truly serious.  I told him to weigh the costs and benefits
of getting exclusive proprietary rights and suggested him to consider
alternatives as well, such as freeware, copyleft and shareware.  I'm not a
computer whiz, just a user...nothing technical.

Then a netter quoted me out of context (which he failed to refer, that of
protecting ideas) and misconstrued my well-intentions from my mention that
all Chinese inputting methods are relatively slow and time-consuming,
making users less productive.  Maybe the choice of modifiers was mishandled
and that "all" and "slow", should be "most" and "cumbersome".  But I did
not appreciate being put out of context nor the comparison to Mr He's 181.7
wpm on Dayi as an example.  I was going to leave it, but couldn't help
myself when I saw two netters discuss this extreme input record.

I usually make messages very brief.  I know this is lengthy.

        Don

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Wed Dec  8 12:30:42 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AB08736; Wed, 8 Dec 93 12:30:38 CST
Message-Id: <9312081830.AB08736@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6798; Wed, 08 Dec 93 13:27:13 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 6470; Wed, 8 Dec 1993 13:15:56 -0500
Date:         Wed, 8 Dec 1993 17:15:24 GMT
Reply-To: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: "L.S.Ng" <L.S.Ng1@SOUTHAMPTON.AC.UK>
Subject:      Remap Cangjie Keys
Comments: To: Chinese Computing Network <ccnet-l@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Dear Netters,

Is anybody using Cangjie to input simplified Characters?

Last year I have put a Cangjie for Guobiao dictionary up to
ifcss.org. It is in software/data and software/x-win/cxterm-dictionary.
Ever since, I have corrected some bugs found.

I have now with me two versions of Cangjie for Guobiao.

1.  Cangjie for Guobiao with some bugs removed.  I planned to put it to
ifcss.org shortly.

2.  Cangjie for Guobiao with new keyboard layout. I believe this will
make the input more comfortable if not improve the speed. The new layout
has the following changes from the original Cangjie:

Original          New layout
   Y (bu3)            X (bu3)
   X (chong2/nan2)    Z (chong2/nan2)
   Z (undefined)      Y (undefined)
other keys remained unchanged.

So you may note that the changes are minor. However it makes Cangjie a
lot more comfortable to use.

The main problem is the Y-key.  In a report which I posted to CCNET-l
before summer, I found that Y has a very high frequency.  Also, the
right hand has more workload than the left hand.  Besides, Y-key
position is awkward.  Upon prolong typing using Cangjie, my right-hand
fore-finger become very painful because Y is furthest from right
forefinger position J-key - it strecthes my muscle.

Putting Bu3 to X, firstly it makes your left hand and right hand
workload more balanced as Bu3 is one of the most frequently used key.
Secondly, it aviods stretching your forefinger.  The distance Y to J is
long but now, you will use your Wumingzhi, so the distance is from X to
S which is now very short.

Don't bother about the argument above, you will feel the ease when you
do it.


Also, I have found a way to force character selection by SPACE_BAR in
CXTERM. The trick is to add a 'undefined-key' (Z in original Cangjie, Y
in new layout) to every input sequence.

For example, if 'a1' is 'nlmnr', it is changed to 'nlmnrz' or 'nlmnry'.
After you typed 'nlmnr', CXTERM will not select 'a1' because it is
fooled to believe that the full input-sequence is 'nlmnrz'. Here you
don't press 'z'. Instead to press SPACE to select the character, because
by definition, pressing SPACE at any time will select Character 1
matching the sequence.



Anyway, from the user point of view, you now have a combination of
Choices:

The first Choice is whether to adopt the New layout. I assure you
it is much more comfortable and only take 5 minutes to get use to.

The second choice is whether to choose autoselect or forced select by
SPACE-BAR.

I will try to put all 4 possible combinations of the Cangjie for Guobiao
dictionary to ifcss.org.



Note that, those four Cangjie only apply to Guobiao Cangjie. I will also
consider modifying Big5 Cangjie later on. This will add to another
combination: GB or Big5. So you have altogether 8 possible combinations
to choose from:


Original (bu3 is Y)  or  New keyboard  (bu3 is X)
Autoselect or forced select
GB or Big5

I shall name them Cangjie.org-bu-Y.auto.GB, ...
Cangjie.new-bu-X.force.Big5 and so on.

Note that each GB Cangjie dictionary is about 70K. Whereas Big5 Cangjie
Dictionary is about 130K. So the complete collections will be 3x70+3x130
= 600K. This apply to dictionary in text format. Upon binary formatted,
it could be halved.

Last but not least, I will also try to write a utility to extract and
modify the Cangjie dictionary provided by Eten.

Thanks for your attention.



Sincerely,
Liangsheng WU

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Dec 24 01:20:28 1993
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA08745; Fri, 24 Dec 93 01:20:13 CST
Message-Id: <9312240720.AA08745@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 6691; Fri, 24 Dec 93 02:03:17 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 1894; Fri, 24 Dec 1993 02:03:15 -0500
Date:         Thu, 23 Dec 1993 08:28:00 -0500
Reply-To: Doug Schiffer <Doug.Schiffer@F513.N260.Z1.FIDONET.ORG>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Doug Schiffer <Doug.Schiffer@F513.N260.Z1.FIDONET.ORG>
Subject:      CC_CALC comment
Comments: To: CCNET-L@uga.cc.uga.edu
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

 On Tue 12-21-93,  8:41 am, John McRae wrote:

 >I'm interested in the transliteration conversion engine you mention.  It
 >would be interesting if you'd place these programs on ifcss.org (the
 >internet address is 129.107.1.155); I believe the procedure is that you
 >ftp to the site, use "anonymous" as your username and your e-mail address
 >as your password, and "put" your file into the /incoming directory.
 >Better check around for complete/correct instructions.

So far, the transliteration engine is built into CC_CALC.  I've thought
about breaking it out into a separate program, but didn't know if there'd
be any interest in a standalone converter.  :-)

In it's current incarnation, it converts from Pinyin to any of the
other systems, with semi-intelligent translation.  By semi-intelligent,
I mean it breaks apart a pinyin syllable into initial and
final portions as well as tone number.  It then converts these into
the corresponding elements of other systems.  It knows about exceptions
that arise in this process.  Thus, if presented with a Pinyin word
that it was not initially tested against, it stands a fair chance of
correctly converting it.  This stands in contrast to a table lookup
approach, which if presented with an unknown word has no chance of giving
the correct translation.

Internally, it uses one 16 bit word to code the Chinese syllable.  As
an added benefit, you can efficiently store any Chinese syllable in
only 2 bytes, and output in any of the systems listed.

 >BTW, are these programs for DOS?

Yes, I failed to mention that.


        Doug Schiffer

From @UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU Fri Jan 28 12:14:23 1994
Return-Path: <@UGA.CC.UGA.EDU:owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from uga.cc.uga.edu by ifcss.org (4.1/IFCSS-Mailer)
	id AA12861; Fri, 28 Jan 94 12:14:19 CST
Message-Id: <9401281814.AA12861@ifcss.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 5505; Fri, 28 Jan 94 13:10:31 EST
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.1d/1.7f) with BSMTP id 8632; Fri, 28 Jan 1994 12:39:55 -0500
Date:         Fri, 28 Jan 1994 09:31:35 -0800
Reply-To: James Lin <chase@NETCOM.COM>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: James Lin <chase@NETCOM.COM>
Subject:      Re: Cstart and Twinbridge?
Comments: To: Wilson <J75BC%CUNYVM.BITNET@uga.cc.uga.edu>
Comments: cc: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
In-Reply-To:  <199401281111.DAA13792@mail.netcom.com>
Status: RO

On Fri, 28 Jan 1994, Wilson wrote:

> does have a WAN XIEN method, but seem making lot of mistakes in translation.

WAN XIEN really follows HANIN which is the very first one introucing
intelligent input selection.   Maybe you can try this and see if it still
give you the same difficuties.  HANIN is available for DOS and CWIN.

> Currently, my main concern is the availibility of outline font, and a fast
> input method. Even though I am using ET 3 and Windows 3.1, I am not satisfied

There are currently over 200 different Chinese TTF or WIFE fonts
available for Chinese Windows (include several simplified TTF for CWIN).

> ZHENG HAN, which can take advantage of input method if the translate table for
> the input method is available, ZHENG HAN provide DAYI table, does any one know
> where I might get additional raw translation table?

There is a ETEN Technical Reference Manual which tell you how to get the
raw input table data out of the systm.  All input methods within ETEN
could be conducted.

Together with the previous issues, there si a product call Unionway
Chinese Input Master which could create your own input method for CWIN
and TBridge and CStar.  It comes with several tables already.

>    Currently, I am deciding which operating system to use, and because I am
> running OS2, chinese windows simply would not run, I am hoping t have a
> alternative, please let me know.

I just finished installing OS/2 on one of my machine and I will try to
use CWIN on OS/2 today.  Maybe I could help to find out this.  Did you
install CWIN on OS/2 partition ?  Does your machine only has OS/2
partition ?

James Lin

From owner-ccnet-l@UGA.CC.UGA.EDU Fri Sep  9 17:41:31 1994
Return-Path: <owner-ccnet-l@UGA.CC.UGA.EDU>
Received: from cnd.org (cnd.cnd.org) by ifcss.org (4.1/IFCSS-Mailer)
	id AA05216; Fri, 9 Sep 94 17:41:29 CDT
Received: from uga.cc.uga.edu by cnd.org (4.1/4.7)  id AA19017; Wed, 7 Sep 94 18:33:29 PDT
Message-Id: <9409080133.AA19017@cnd.org>
Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2)
   with BSMTP id 5100; Wed, 07 Sep 94 21:30:39 EDT
Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.2a/1.8a) with BSMTP id 6687; Wed, 7 Sep 1994 21:29:28 -0400
Date:         Sun, 4 Sep 1994 06:34:44 MYT
Reply-To: Ng Liang Shing <liang@HSNG.PL.MY>
Sender: Chinese Computing Network <CCNET-L@uga.cc.uga.edu>
From: Ng Liang Shing <liang@HSNG.PL.MY>
Subject:      Architecture of Chinese Character Input Output System
Comments: To: Chinese Computing Network <ccnet-l@uga.cc.uga.edu>
Comments: cc: "Dr. Chen You Hor" <yhchen@clcs.pc.my>
To: Multiple recipients of list CCNET-L <CCNET-L@uga.cc.uga.edu>
Status: RO

Architecture of Chinese character input/output systems


Chinese character input/output systems include functions/algorithms for
translating device input (e.g. keystroke on keyboard) into Chinese character
code and functions/algorithms for translating Chinese character code into
device output (e.g. character on screen).

Here I attempt not to discuss how the I/O functions/algorithms are actually
implemented. Rather, I shall describe two classes of mechanism for interfacing
the I/O functions with application programs. I shall consider the problem
of how to call a Chinese character I/O function from an application program.


Invasive vs. Safe approach

The two classes of interfacing mechanism I wish to comment on are the invasive
and safe approach. They differ greatly in many ways and have advantages and
shortcomings in respective areas.

What I call the invasive approach is generally regarded as the 'normal'
approach or 'conventional' approach by many Chinese programmers and users,
because of its popularity.

Consider an example in DOS.  A program usually calls an interrupt service for
character I/O.  By default, the interrupt services provided by DOS or BIOS
would display ASCII (American Standard Characters for Information Interchange)
in text mode.  The trick is to 'intercept' the interrupt service routine for
character I/O.  A call is made to the Chinese character I/O functions instead
of the default character I/O functions.  This can be accomplished by simply
placing the address of the Chinese character I/O in the interrupt vector
table.

Other mechanism for 'intercepting' character I/O exists in DOS. The ultimate
in the art of interception exist at the memory level. Ultimately, BIOS or DOS
interrupt services write characters into the display memory area. The PC have
different display memory area for text and graphics. The first step is to
switch the display hardware into graphics mode, but fooling the running
application so that it thinks it is in text mode. So the usual character I/O
would write characters into the text display memory area. The intercepting
system interprets the content in the text memory then translate this into
Chinese character bitmap and write it to the graphics memory. Because the
display hardware is in graphics mode, the content in the text memory area will
not be displayed, but the content in the graphics memory area (Chinese
character bitmap) will be displayed.

Similar schemes exist for intercepting keystrokes to accomplish Chinese input
methods.

The safe approach is a much simpler and safer practice of programming. Here,
as constrast to the above example where the running application is not aware
of the presence of the Chinese systems, the running application makes explicit
calls to a well defined Chinese character I/O functions.

Consider examples in DOS again, the safe approach can be implemented in
several ways. One would be calling library functions. Alternatively, such
functions may be installed as an extra interrupt service routine (ISR), not
overriding the existing ASCII (text mode) character I/O ISR.


Advantages and shortcomings

This immediately brings out issues concerning the difficulties of
implementation and the benefits of information hiding.

The safe approach is easier to implement than the invasive approach obviously,
simply because function calls does not involve explicitly manipulating system
area (overriding the interrupt table in DOS, for example).

The invasive approach enjoys (?) the benefits of information hiding. Here, the
running application need not be aware of the presence of the Chinese systems.
Therefore, in DOS and other operating systems, some (but not all) existing
programs can be ported for Chinese use at binaries level. The safe approach,
however, is not directly portable at binaries level, though it is portable at
source level.

The benefits of information hiding explains the popularity of various invasive
Chinese systems in the DOS environment. Recently, intercepting capabilities
have been extended into Microsoft Windows environment and again has become a
popular choice.

The difficulty of implementation on the other hand may explain the absence
Chinese systems in other operating systems. (Note: While interception has
succeeded in Microsoft Windows, no interception is successful in X-Windows.
CXterm does not fall the category for invasive approach.)

Systems that are safe and also provides benefit of information hiding exists.
They are in quite difficult to implement because of the sheer scale of the
work involved in rewriting the interface system.


Examples

Name of System                       Approach  Comments

Microsoft Windows (Chinese version)   safe     safe because no interception.
                                               no interception because
                                               Chinese is default.
Mac OS (Chinese version)              safe     (same as above)
CXterm (Chines Xterm)                 safe     (same as above)
Eten                                  invasive    overriding BIOS ISR and
                                                  memory  (?)
ZWDOS                                 invasive    (?)
Chinese Star (for MS-Win)             invasive    intercepting Windows APIs.

Note: in above systems, running applications need not be aware of the presence
of the Chinese system

NJSTAR (for DOS)                      safe     calls to library functions (?)

Note: NJSTAR is an application program.


The future

Intercepting system area will be increasing difficult in robust operating
systems. Modern operating systems are towards this end. Hence invasive Chinese
system may dwindle with the end of the DOS era.

Fortunately, modern operating systems are towards multiligual interface (e.g.
adopting Unicode). This may provide both easy and safe programming approach
for Chinese character I/O. However, multiligual interface might not be
universally available or would come as an option with a higher price tag.

Safe approach with simple library function call can be an intermediate
approach until multiligual interface becomes widely available. It could also
lives on after that. In short, it has the strongest ability to survive. It is
a choice for the present and the future.

--
Ng Liang Shing            |   This is an Internet node host inside
Sungai Petani, Malaysia   |   an old 286.  A few years back, this place
liang@hsng.pl.my          |   was not more than a small town.

From yuping@TECHNOA.CO.JP Thu Jun 25 01:47:58 1998
From: "Ping Yu" <yuping@TECHNOA.CO.JP>
To: <CCNET-L-request@UGA.CC.UGA.EDU>
Subject: GB Vector Font
Date: Thu, 25 Jun 1998 14:48:45 -0000
MIME-Version: 1.0
Content-Type: multipart/alternative;
              boundary="----=_NextPart_000_01BDA048.5A77C4D0"
Content-Transfer-Encoding: 7bit
Status: RO

This is a multi-part message in MIME format.

------=_NextPart_000_01BDA048.5A77C4D0
Content-Type: text/plain; charset=HZ-GB-2312
Content-Transfer-Encoding: 7bit

Dear Sir,

Please let me ask you a question about GB Vector Font.

I am transplanting a CAD software from  NEC EWS4800 Unix system
to PC Windows NT4.0(Simplified Chinese Version).

I want to input, display and output Chinese(GB) characters,
while running PIETRA on PC Windows NT4.0 (Simplified Chinese Version).

When Chinese are input on Windows NT4.0, the Chinese are only Character
Codes in fact, such as {010203}.
Then Chinese font is found according to the Character Codes,
and Chinese Characters are displayed on screen, or printed out from
printer.

It can be said that there are two kinds of fonts: Dot Font  and Vector
Font.
There is only Dot font in Windows NT4.0.  Is it right ?
But I want to use Vector Font in PIETRA, in order to zoom in/out the
Chinese Characters as a graphics.

Now I am looking for GB Vector Font(datafile) and its using method.
Would you tell me how to get  them please ?

Thank you very much.

Best Regards,

Ping Yu            yuping@technoa.co.jp
TECHNOA INC.
1-50 Akanabe Hishino,Gifu 500-8268,Japan
Tel 058-2731445         Fax 058-2739562

