|
On the whole, O'Reilly books deserve the respect they get. Many are gems that I will keep for a long time. But some have been low quality, and that makes me sad. Here are some examples.
The MySQL/msql book was a loss, but the Postgres book is even worse (nice "index" -- and the plug chapter for their lame software is in poor taste). I want to hurt whoever let that piece of crap go to press.
*Programming Python* is poorly named; I expected it to be the Python analog to *Programming Perl*, but instead it's just a pile of tips on various APIs and Python tricks.
Which reminds me: the phone-book size is definitely no good. It reminds me of those terrible LARN LUNIX IN JUST 24 HOURS!@#$% books that are just reprints of poorly edited man pages. Stay concise, stay usable (if it's too big for RepCover, it's too big), stay close to the core of a technology. We don't need reprints of the man pages (the Postgres book is a big offender here).
Also, there seems to be a glut of less core-oriented and more fad-oriented books. *Mastering Algorithms with Perl* is great, but *Learning the Basics of Some App That'll be Gone in Two Years* is not. Teach the fundamentals, and the trends will come easily.
While I'm squeaking, let me request a new Emacs book! The existing one covers version 19, and we're at 21 now.
Thanks for all the great books!
|