-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mike Miller wrote:
> On Fri, 3 Oct 2008, Smith, Craig A wrote:
> 
>> Perl is great because you always have source and, thanks to its 
>> just-in-time complier, runs loops as fast as non-interpreted languages.
> 
> How does this compare with what Perl is doing:
> 
> http://en.wikipedia.org/wiki/Psyco
> 
> 
>> Slashdot is another solid example.  They note perl's speed
>> http://www.slashcode.com/faq.shtml#SlashCode1
>> and they've licensed their code under Stallman's GPL
>> http://www.slashcode.com/docs/COPYING
> 
> One problem with learning Python is that I will be less able to make use 
> of all the nice Perl code out there.  But focusing on Perl means I'll be 
> less able to use all the nice Python code out there.
> 
> It is definitely a tricky decision.
> 
> Mike

I remember Perl when it was really nothing more than an easier to use
sed and awk.  Then I remember when we had teams of guys who did nothing
but audit perl cgi people wanted to run on our webservers.  Fast forward
to today, really the only people using it are the people who already
know it.  It's gone from nearly the only game in town in certain arenas
(like dynamic content in web pages) to virtually unused in terms of new
development in that arena.

Re: whether perl 5.10 is a major release or not, see the changelog.

http://perldoc.perl.org/perldelta.html

Perl is probably the fastest interpreted language you will run in to out
there, on average it's about 10% faster than python, and significantly
faster than ruby.  You can certainly come up with code snippets that are
much faster or slower though.  In particular perl has a pretty fast
regex engine, so if you make heavy use of regexes it's attractive.

My personal experiences with perl are that large programs are hard to
debug and maintain, and it's speed advantages aren't worth dealing with
thre maintainance and debugging problems.  If python is too slow for a
task, perl isn't going to be enough of an improvement to matter most of
the time.

Case in point, I once tried to generate a report that determined the
highest number of hits by IP in any given ten minute period from 18
months of ten million hit per day weblogs with python.  It ran for a
weekend and wasn't even close to finishing on a dual xeon.  The C
program took me far longer to write, but it ran to completion in 4
hours.  I doubt rewriting in perl would've made it run acceptably fast.

My advice to people is, if you've gone through the work of learning
perl, all of it's operators, all of it's special cases and exceptions
(and there's a lot of those) all of it's commonly used paradigms, then
there's no compelling reason to change.  If you are wanting to learn
something today, python and ruby are attractive choices.

For what it's worth, I know sed and awk backwards and forwards and don't
recommend people learn those either.  There's no reason to forget I know
them, but for new tasks....there are better choices.

My favorite quote from Programming Perl possibly slightly misquoted:

"Perl has a rich set of operators that generally do what you expect,
unless you are expecting consistancy."

- --
Thanks,

Josh Paetzel

PGP: 8A48 EF36 5E9F 4EDA 5ABC 11B4 26F9 01F1 27AF AECB
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)

iD8DBQFI5h7BJvkB8SevrssRAmSGAJ9gswAbn9jdMVh2bvbiFZByXEDZWwCglBkC
HqrRr27fyHDtunjdOBmky8I=
=irjk
-----END PGP SIGNATURE-----