Tim is setting up to talk to some energetic yet drowsy people about learning from the book market.
In the meantime, I check out Microsoft’s Channel 9 where you can listen in to engineers as you could listen in to pilots in the old days. Also of course the blogs of MSDN.
Tim is talkng about how BookScan (sorta like SoundScan) tells fairly accurately the sales of their top sellers by surveying top book sellers (amazon and borders are 16% of market each in US). Tim also uses some open source software that looks a lot like smartmoney’s map of the market to map various categories. He also crawls Amazon in smoe way that I missed but we’ll see that live in a few minutes,
computer book sales are off 15% so far this year from 2003 and the seasonal track looks nearly identical. why? no new tech solutions hit the market (no significant ones that required significant new books).
Why doesn’t Google do more with Zietgeiss (sp)? asks Tim. Stock prices of tech match nearly book sales trends on those topics. Java books tell about Sun stock for example. Not a leading indicator but a matching indicator.
This stage says Tim is the period of cleaning up on an ongoing basis the ontologies. the ontologies need to be clean and repositioned constantly (?) now.
Note that Java sales are dropping while C# is gaining and PHP is climbing sharper yet. Python is more or less flat after a little bleep up. PERL is flat despite no new developments in PERL in a long while (this is actually good for PERL. This is when looking at the language dimension.
msql and sqlserver are cleaning Oracle’s clock. postgress well very small.
applications servers chart is so small and with so few books that even one book changes the chart. basically not a good trend measure.
java vs .net .net rose above java in 4th Q of 2003 and remains higher. Note that intro books sells better as the market matures. Much discussion on this chart with lots of Java appologists
Market shift from reference to tips and tricks and intros. why? references are better done on the net.
microsoft leads open source despite OS upturn. apple spikes are each new release (as you might expect). who are these buyers? CS students, professionals, whatever. Note the number of CS majors worldwide are shrinking could this be a contributing cause?
Shapes of the curves for $$$ of sales looks a lot like Lotka’s curves.Photoshop books make money! Mac OS X or windows XP does about half. Java and C/C++ have identical curves! C# mostly sells only 4 or 5 profitable books.
but programming language sales show an 70 – 30 split between propriety — if java is considered open source. without java on either side they are equal.
In ad words on Google, Java rules based on charges.
See also Tim’s Treemap
What other sources of data might be used? Want ads? Blog entries? Free docs online? rural vs urban? price sensitivity. idiots are cheaper then dummies.