The day started out with Larry Wall's keynote, which was entertaining as ever. His talk was not as focused as it was last year (but that just be my bias of the aural over the visual).
Larry (who is featured in this month's Forbes magazine), discussed Perl's complexity being a feature ("English is useful because it's a mess", "Simplicity is often the enemy of success"), and went on to apply Tao-like philosophy to a description of Perl's usefulness.
TPI News cannot be everywhere at once, so we cannot report on everything that happened. Nathan Torkington changed the topic of his "Free Tools and Perl" talk to "Porn and CPAN", a change most people seemed to think was a decided improvement. After his relatively short talk about identifying content stored in a proxy cache, Eric Eisenhart and Dustin Mollo showed a demo of a new CPAN interface they are working on at perl.com.
Helping users get the most out of CPAN was a hot topic Tuesday and Wednesday. CPAN Testers will be moving to TPI, and (hopefully), the perl.com CPAN interface and CPAN Testers and WAIT will all be integrated on some level. Some sort of quality rating system of modules on CPAN has been the topic of much discussion.
In a TPI BOF (Birds of a Feather meeting) Wednesday evening, Chip Salzenberg, TPI's executive director, discussed what TPI is doing, and what it will be doing, and how to get involved. What TPI is doing is getting volunteers to do work. What it will be doing is (in addition to this news page) a bug tracking database, increased and better communication with the media and corporations, better pointers to information for Perl users. One can get involved by becoming a member and contributing money which helps pay for infrastructure, and by volunteering time.
The MacPerl BOF saw a couple dozen people, most of them MacPerl users, and many of them interested in such things as the in-progress 5.005 port of MacPerl (with features like threading, the compiler, improved regex support, etc.), and DBI and Tk ports. No one has volunteered to do the DBI and Tk ports yet. Tk is a lot of work, and not many client libraries exist on Mac OS, which are needed for building client libraries. It was pointed out that local DBDs are not necessary if using the DBI proxy, which may ease a port of DBI for MacPerl.
The FreeBSD people, noting that Red Hat passed out free RH5.1 CD sets, gave out free copies of FreeBSD 2.2.6 4-CD set from Walnut Creek. The conference organizers changed their minds about giving free copies of The Perl Cookbook to speakers, and are giving books to speakers on request.
Jon Orwant, Tom Christiansen, Chip Salzenberg, Graham Barr, and Mike Stok had a "debate" about OOP in Perl, but very little blood was shed. Most people seemed to agree that OOP can be useful in Perl, but that it has some problems, and that it is waaaaaaaay overused by CPAN authors.
Jon Orwant also hosted a Quiz Show (modeled after College Bowl), pitting four teams of three against each other, two at a time. Of course, the team with two members of perl5-porters won the prize of Perl Resource Kits, Perl Journals, and other goodies. The event was entirely entertaining, and, if we may step outside of the shroud of objectivity for a moment, we'd like to see it as a fixture for TPCs in the future.
