Mike Hendrickson from O'Reilly has published a survey of the technical book market. As when viewing the TIOBE site, I think it is wrong to extrapolate any broad conclusions about the development languages market from the data you see here. For instance, lots of developers are still writing batch files, it just so happens that the book-market for that syntax is small. Changes in the C/C++ market are slow, which means that developers stick with tried and true books for those very popular languages rather than rushing out to buy new texts. Nevertheless, it is instructive to at least view this data, especially if you happen to be interested in technical books or trends in the broader developer landscape.
Well known languages that have been growing quickly include ActionScript, which is up 53%, Python, up 31%, and PowerShell which has come out of nowhere to claim 1% of the market. Some well known languages that have been losing market share are Visual Basic, which is off by 34%, vbscript, which is off by 30%, and perl, which is off by 23%. Ruby has 5% of the market, and has grown at a rate of 10%. SQL is up by 5%, VBA is off by 11%. Some of these numbers may simply reflect fluctuations due to release cycles, but others may represent trends in the publishing market.
There are some interesting details in the figures found in this article. My old favorite language, Delphi, sold a grand total of 126 units in 2007. F# sold 698 units, but I bet we will see that number grow quite a bit in 2008. The overall market is dominated by languages that run on some kind of virtual machine, with C/C++ being the only native language that is still selling broadly in the book market. Oh yes, I should also add that sales in the technical book market as a whole are off by about 1%. The web is a huge, and very hungry, machine.
Mike Hendrickson says that the information in his article is drawn "from Bookscan's weekly top 3,000 titles sold. Bookscan measures actual cash register sales in bookstores. In other words, if you buy a book it gets recorded in this data."
Remember, you should not draw too many conclusions based on this data. For book authors and publishers, this information has obvious value. For the rest of us, it is just a tantalizing peak into a market that is very difficult to understand. Only one thing is certain: the actual usage numbers for these languages is probably quite different from what you see in this article.