QC draft

[Guylhem Aznar <guylhem@oeil.qc.ca>]

Here's Alessandro work on QC.

I think we could use it as "QC manifesto"

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

There are two versions: the first includes the rationales behind any
paragraph and the second is just the rules (for people with little time
to read).


	Proposal for a Quality Control Group

Teams of the group are volunteers  and being part of the group doesn't
reequire   devoting  any  predefined   amount  of   work  to   the  QC
needs. People interested in being  part of the QC group must withstand
approval of  the LDP  leader, upon presentation  of a resume  or other
self-supporting evicence.  (Rationale: the quality-control  group must
be quality-controlled as well; we  must prevent unfair people to enter
the  QC group  in  order to  damage  the LDP.  Also,  we must  prevent
incompetent people  to judge  LDP authors, even  if in good  faith, as
that would damage the LDP anyways).

The QC people  is subscribed to a closed mailing  list.  Every new LDP
document is forwarded to the list  in source form, maybe by having the
list itself as  a subscriber of ldp-submit. When a  QC member wants to
review a document, the member must express this intention by sending a
"lock" message to the list, to inform other reviwers that the document
is  already taken.   In case  of multiple  replies, the  first  one is
effective (according the  ID of the message within  the mailing list),
unless the reviewers agree  differently via private email. (Rationale:
the list  must be  kept low-traffic: we  don't need another  place for
discussion; on the other hand  all reviewers should have direct access
to reviewable material without human intervention).

Submitted documents appear immediately in a "beta" area in the LDP ftp
site, maybe  only in  source form (sgml  or whatever is  agreed upon).
The QC group should approve the new document within one week; if no QC
approval arrives  within a week, the  document is moved  to the normal
ftp  areas  anyway,  like  any  "normal"  document.   (Rationale:  the
information must be  made availabe as soon as  possible to the general
public, and  the QC  group cannot  be allowed to  slow release  of the
documentation.   The "beta"  area allows  information-eager  people to
access to the  latest and greatest independent of the  QC week and the
possible  extra delay  before  the  document is  moved  to its  proper
place. Compiling  and mirroring all  the output formats for  those few
people would be an unneedeed waste of network resources).

The reviewer should sort any  problem directly with the author, who is
prepared to be be responsive during  the QC week.  If an agreement can
not be reached,  the document is either reviewed  by another QC member
(releasing  the lock  on the  QC list  and allowing  anothere  week of
"beta" status) or  the whole question is brought to  the LDP leader or
HOWTO  maintainer.  (Rationale:   personal  criticism  should  not  be
performed   in  public  and   the  issues   can  usually   sorted  out
friendly. Sometimes, however, the author  and reviewer may not come to
an   agreement  for   a  variety   of  reasons,   including  character
differences; in this case a second chance must be allowed).

The LDP leader and HOWTO  maintainer are allowed to refuse a document,
if it was considered unacceptable by at least two QC people. This does
not prevent the author for releasing the document through other means,
it just  won't be  distributed as part  of the  LDP and using  the LDP
resources. (Rationale:  a negative  judgement doesn't deny  anyone the
right to speak,  and this must be stated clearly  before we are called
fascists or anything similar).

Documents that passed QC review will  carry the QC note just after the
title. A document that gets no  QC review within the allowed time will
be part of the LDP without the  QC mark. The QC approval will refer to
a specific  version of  the document, and  an unreviewed  release will
still  carry   its  QC  approval   for  the  last  version   that  was
approved. (Rationale: the readership  must benefit from QC, thus being
warned whenever  a document has been  reviewed and when  it didn't. On
the other hand,  the QC group can feel less compelled  to review a new
revision of a  document that was already QC'd  but the readership must
know  about   that  and  adapt   their  confidence  in   the  document
accordingly).

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

	Proposal for a Quality Control Group

Teams of the group are volunteers  and being part of the group doesn't
reequire   devoting  any  predefined   amount  of   work  to   the  QC
needs. People interested in being  part of the QC group must withstand
approval of  the LDP  leader, upon presentation  of a resume  or other
self-supporting evicence.

The QC people  is subscribed to a closed mailing  list.  Every new LDP
document is forwarded to the list  in source form, maybe by having the
list itself as  a subscriber of ldp-submit. When a  QC member wants to
review a document, the member must express this intention by sending a
"lock" message to the list, to inform other reviwers that the document
is  already taken.   In case  of multiple  replies, the  first  one is
effective (according the  ID of the message within  the mailing list),
unless the reviewers agree  differently via private email.

Submitted documents appear immediately in a "beta" area in the LDP ftp
site, maybe  only in  source form (sgml  or whatever is  agreed upon).
The QC group should approve the new document within one week; if no QC
approval arrives  within a week, the  document is moved  to the normal
ftp  areas  anyway,  like  any  "normal"  document.

The reviewer should sort any  problem directly with the author, who is
prepared to be be responsive during  the QC week.  If an agreement can
not be reached,  the document is either reviewed  by another QC member
(releasing  the lock  on the  QC list  and allowing  anothere  week of
"beta" status) or  the whole question is brought to  the LDP leader or
HOWTO  maintainer.

The LDP leader and HOWTO  maintainer are allowed to refuse a document,
if it was considered unacceptable by at least two QC people. This does
not prevent the author for releasing the document through other means,
it just  won't be  distributed as part  of the  LDP and using  the LDP
resources.

Documents that passed QC review will  carry the QC note just after the
title. A document that gets no  QC review within the allowed time will
be part of the LDP without the  QC mark. The QC approval will refer to
a specific  version of  the document, and  an unreviewed  release will
still  carry   its  QC  approval   for  the  last  version   that  was
approved.



-- 
Guylhem Aznar, Linux Documentation Project leader: http://www.linuxdoc.org
               Clef PGP/PGP key:    http://oeil.qc.ca/~guylhem
               Chez moi/At home:    guylhem \@/ oeil.qc.ca
               Anywhere/Partout:    guylhem-pager \@/ oeil.qc.ca

PGP signature