First of all, a round up of the talks that I attended and some of those I wish I had.
Tuesday
Keynotes
Cog spoke on how to get the most out of a YAPC - funny and useful. Larry spoke, which was nice.
DB track
I would have liked to have seen Barbie's talk on Selenium, Juerd talking about tuits, Ranguard on evolving architecture, also apparently the perl/ssh for monitoring talk was good, and dakkar spent the morning in the Parrot hackathon playing with tree processing tools.
But I stayed in the DB track, catching the end of Smyler's "When MySQL Bites" (to get MySQL to not do certain insane things you have to tell it to behave "traditionally"). I stayed for Philip Stoev's talk on using MySQL client to talk to various databases (good subject matter but didn't hold interest) and Grrrr's talk on DBIx::Perlish (fantastic subject matter, weird syntax, crazy language hackery, didn't follow as closely as I should have. Note to self, as per Grrrr and mst, can't use overloaded ops for this, because you want to be able to really use flow control, which isn't covered by that interface.)
Catalyst track
I would have liked to have seen all of the talks (Jonathan Worthington on Parrot, markov's SOAP stuff, which sounded exciting when we discussed it at lunch - at last, some Perl SOAP stuff that might actually work!
Greg's funny talk. More markov on logging, Brian McCauley on "Usenet gems" and pjcj on the use of Perl in a London bank.) But you've got to make choices, so I settled for the catalyst track:
Saw mst's "Database Haters Anonymous". After first watching the awesome power of Linux failing to speak to a projector (this sight was common over the 3 days), we had a shorter, and angrier version of the talk, which was essentially "DBs suck. Software sucks. These people rock". Great fun, a potted overview of the people and some of the technology behind DBIx::Class.
I stayed for draven's (Marcus) talk about Iusethis. Great slides, interesting project, and Adam Bartosik on a Polish team's experience with Catalyst (sounded interesting, delivery could be improved, but props as a non-native speaker).
Various
I'd wanted to see Nigel Hamilton's talk on Trexy, but somehow having a reasonable lunch hour seemed more tempting. (This was my major critique of IPW2006 too, I think a lunch hour should last more than 1 hour).
I got to meet Nigel later in the week and discuss Trexy though. Nelson Ferraz's "Adventures in Perl 6" sounded great fun, but I plumped for acme's Scaling with memcached, which was great.
Next was difficult, and I had to skip Evil's news2mail (sorry!) and miyagawa's Web Scraper to see Jonathan Worthington on Parrot, but it was my only Perl 6 related talk, and I've never seen him talk before...
Interesting, especially as I've given up on trying to read and understand p6i in the last year. I had wanted to see BooK's Net::Proxy talk (highly acclaimed by everyone I spoke to who'd been to it) but ended up staying for Richard Dice's talk which was slightly dry to begin with, but was actually fascinating. Richard looked at various data about Perl, at the point of view of management and consultancies, and gave a convincing case for the importance of the role of The Perl Foundation.
Wednesday
Keynotes
Damian spoke about positronic variables. OK, as a concept, this didn't work out as useful and intuitive as Quantum Superpositions, but a Damian talk is something to be seen. A non-programmer could attend a Damian talk and be entertained. A great combination of research, improbable connections, silly surrealism, programming bravura, photoshop and theatre.
The sponsors of the "Jobs Fair" spoke for "5 minutes" each. This overran somewhat though, meaning that there was no time for coffee before Dominus's Repair Shop and Red Flags.
Dominus
OK, I'd have like to have seen various of the other talks, especially JJ on Functional Programming and the POE hackathon, but I've read Dominus's material on Red Flags, and wanted to see him perform it as he's a great speaker.
Didn't think all the examples worked as well as in the articles, but a great and useful talk.
Moose track
I missed clkao's talks on Jifty, as I've recently been
playing with Moose. Stevan's talk was a little dry but I think communicated a lot of the advantages and excitement around Moose. Yuval's talk was inspiring in scope (his toy project MO uses a set of purely functional transforms to do very clever things to object model stuff) but was maybe a little overdetailed and I couldn't follow it very well.
Clash of the Titans
Dominus vs the Damian. How else would you schedule this? I'd seen Damian's Perl 6 update five or so years ago in Paris, decided to go for Dominus on functional parsing. Funnily enough, this is more or less the only chapter in HOP that I understand, so it was a useful refresher more than anything.
Thursday
Lightning talks
Some of the best lightning talks aren't:
BooK presented this year's French Perl Workshop with a cleverly put together video ("
Iasse! Oui nide iou!")
A German from $foo presented Win32::GuiTest using... Win32::GuiTest. Fantastic (he had to move the mouse once to get the testing tool started again... but we'll let him off).
Cog did the last talk... and it overran just a little :-)
Other stuff of note: Abigail complaining that File::Copy is broken, parsing OCR using Regexps, a comparison of image processing tools (conclusion: use Imager for processing, GD for creation), Juerd explained a simple workaround to make unicode string handling consistent.
The value of advance preparation and remembering where you put things.
If I'd prepared my slides earlier, I could have seen various intersting sounding talks about AI, the debugger, demerphq on regular expressions, Abigail on solving sudoku with regexps, and Gerard Goossen's talk about MAD and another cog talk (on how NOT to write a Perl resume...)
But first of all I was working on my slides. The great thing about having slides which are 95% complete is that they will stay 95% complete for ages until I get enough adrenalin kick to finish them.
Then I mislaid my bag. In the office with the auction goodies. Thanks to the lovely organizers for their patience and for seeking me out before I cancelled I
of my cards.
My talk
I would quite happily have seen the utterly lovely Karen Pauley, and mock's talk on scalable data collections looked intersting but... it's probably polite to attend your own talk.
This went ok - I think I had too much material and it could have done with being either funnier or more informative. Still, the slides are up here or at Skud's request on Slideshare.
Counterexamples
Saw Trelane on why Perl sucks (and what to do about it) and Marty Pauley on Perl Worst Practises, and why variables are bad. (See also: Haskell).
