(1) O'Reilley books used to be titled things like
"lex & yacc" where you might pick up the book
just to figure what it's about. Now there seems
to be an attempt to downplay the technical terms
e.g. there's no "mod_perl" book.

All of the wanna-be kids mindlessly, automatically
select what they want to be from the gurus.
Even nitche topics - academic subject, hardcore
subjects, hyper technical subjects - trickle
down. In some cases, it takes more than a few
years. OO was in the labs a generation before
it crept out, then it took a generation to
start using, and not until this generation
have we learned how to teach the subject so
that people get it. Functional programming
went through the same throws, but now
"Structure and Interpretation of Computer
Programs" is an ageless classic and a best
seller.

I won't ding ORA for catering to the Windows
crowd (even as I stand firmly outside of it).
As someone else mentioned, the despreation
of things like shark jumping comes from, well,
desperation - not having anything better to
fall back on. ORA's value is as a leader. If
ORA doesn't print hardcore, nitche, technical
books, then someone else will. I've been watching
MIT Press and the OpenCourseWare thing.

O'Reilly took risks on MacOSX (lots of product
releases have a buzz, but O'Reilly saw real
potential). O'Reilly took risks on Google as
a platform. And of course, back in the day, on
Perl. They've also taken risks on Python, which
has proven to be language well worthy of attention.
O'Reilly did not sit back and say "No one will
switch from Windows", "no one will drop Perl
for a younger but cleaner language", nor
"people can't buy Google, it isn't a real
product, therefore there is nothing to write
a book on".

Yes, this sort of thinking is risky. Tim described
O'Reilly's early history as having an element
of luck. Anyone can play it safe. Trying your
luck now and then, dispite flops, misdirection,
mispredicted trends, inadequate resources to
polish a book, and all of these nasties, is
what makes O'Reilly great.

I wonder why O'Reilly didn't pick up Damian
Conway's _Object Oriented Perl_?

I found this very interesting reguarding the
lack of truely hard-core books for Perl,
subjects getting a very glossed over treatment,
and outright downplaying of technical trends
so that Perl programmers don't have to worry
about them, and how closely some, but not all,
Perl programmers really resemble PHP
programmers with apathy for design:

http://perlmonks.com/index.pl?node_id=285091

-scott