Stories
Slash Boxes
Comments
NOTE: use Perl; is on undef hiatus. You can read content, but you can't post it. More info will be forthcoming forthcomingly.

All the Perl that's Practical to Extract and Report

use Perl Log In

Log In

[ Create a new account ]

gnat (29)

gnat
  (email not shown publicly)

Journal of gnat (29)

Saturday July 07, 2001
10:39 AM

language-dev list

[ #386 ]
There are some interesting conversations coming out of Simon's language-dev list. There's an interesting tension between the people who want to talk about language and people who want to talk about implementations. I think we're getting past that to talk about implementation stuff now, though.

There have been some pipedreams. A universal extension system, for instance. In some ways that's what SWIG is, or what Mostly C is. I know Chip Salzenberg has been thinking about this for a while--a way to wrap a C library and have it usable from Perl, Python, Ruby, whatever. That's only interesting if you want what those other languages have, or they want what we have. Me, I'm still giving hard love to Inline.

They're currently discussing hash functions. There was an interesting message from Tim Peters about Perl's old and new hash functions, comparing them to Python's. I wonder how important distribution is compared to speed--is there a level of acceptability with hash functions, whereby you don't to pay the cost of a slower perfect (or even better) distribution of keys to buckets, but instead settle for a flawed distribution that is fast to calculate? Perl is probably a good affirmative answer to this :-)

There was a passing discussion of a regex engine that could be used in Perl, Python, Ruby, etc. I'm not sure we'll go this route for perl6. My experience has been that making reuse a priority means you trade off efficiency. This may be moot anyway, said Dan Sugalski (perl6 internals guru), as we may integrate the regex engine closer with Perl by making what used to be states in the regex state diagram become opcodes in the Perl virtual machine. As Dan said, this will be a good test of whether we have a tight opcode dispatch mechanism--the C code in perl5 that implements a regex match contains its own dispatch loop for interpreting the compiled regular expression. For more on how perl5 handles regular expressions, check out Mark-Jason Dominus' Internals of Rx and Dirty Stories about the Perl Regular Expression Engine sessions at the Perl Conference/Open Source Convention.

The countdown to TPC/OScon is on. We've cancelled a couple of low-attendance tutorials (it shouldn't affect you because, according to our records, you hadn't registered for it anyway), had some people pull out (only dramatic one was a PHP guy who had shoulder surgery), and send stuff off to the printers. Over the coming weeks I'm going to get so eager and excited I'll twing when you pluck me (please don't pluck me).

This week I conquered my fear of paperwork. Well, not completely. I still can't get motivated to fill out expense sheets and all that crap. But I finally did some contracts I was supposed to have done a week ago. I'd been putting it off, putting it off, dreading it. And when finally I screwed up my courage to do it, I found it was pretty easy.

There's a moral there, of course: hand paperwork off to the editorial assistant so it doesn't sit on your desk for a week, you damn fool! :-)

Currently in heavy rotation in my CD player: Tim O'Brien's "The Crossing". Acoustic singer-songwriter stuff with a few instrumentals, based on Irish American experience. Lots of really catchy hooks. I hate that--it makes it so hard to put something else onto the record player!

Happy weekend, everyone!

--Nat