use Perl

Tuesday began with Abigail introducing us to Lexical::Attributes. He started by taking a look at Perl's handling of OO, and some of the learning experiences and misunderstandings he had when beginning with Perl OO. There was some interesting stuff, but it was commented on IRC 'when is he going to get to Lexical::Attributes?' He did eventually, by way of OO::Closures, Fly Weight Patterns, Inside Out Objects and Class::Std. Abigail also noted that unfortunately Lexical::Attributes is currently slow due to it using Filter::Simple::FILTER_ONLY. Hopefully this can be addressed in the future.

Dave Rolsky's talk was rather full, and seeing as I previously investigated his Params::Validate module, I didn't expect to learn anything new, so dropped in on Autrijus' Perl 6 talk. Caught part of the testing section and how to become a commiter ... it's really hard, just "commit stuff!" I wanted to see Kevin Scaldeferri's talk, so on hearing the applause from Dave Rolsky's talk I headed to the next room.

The talk, 'How to server a billion pages a day with Perl', is based on experiences at Yahoo!. Performance matters, so techniques looked at in the talk were caching, partioning and distributed computation. The latter aids cleaner design and makes more efficient use of memory and CPU. But it does require more servers :) Walt Mankowski was next up for a security talk. Set theory begins the talk, and Walt notes that Perl doesn't have sets .. but we can make them. Looking at big O notation to show how quickly the look ups can be, Walt even got to have a dig at Damien, as it was done in constant time .. without Quantum::Superpositions ;) 9 months of data and 80MB of logs, managed to find 120,000 non-spam and 14 confirmed bad addresses in 3 minutes on a 1.5GHz machine.

After lunch Steve Purkis looked at address parsing in Perl, as performed at Multimap. So what does an address look like, well they're all different or at least there are a sizeable variety. Also you need to bear in mind countries like Belgium, which have 2 formats depending on the language. Most people don't write addresses to spec either. What if an English person writes a Polish address, they're likely to use the English names. A very interesting talk and I look forward to investigating this further. Daniel Yacob then talked about regular expressions in non-Latin-1 languages. It got quite complex with the all the language parsing involved, never mind the regex parsing. Not sure whether it'll be taken up, but interesting research nonetheless. Then it was my turn. A talk about phrasebooks and specifically introducing Data::Phrasebook. I think the talk went well, as I certainly had a couple of people asking questions afterwards and again on the boat trip later in the evening.

After the break, I sat and listened to some talks based around Class::DBI. I don't use it myself, but I am intrigued by it. I keep feeling I should get into it, but at the moment I get by with the phrasebook Design Pattern. I still like to hear what everyone is getting up to with it though. Casey West started with a Beyond the Basics of Class::DBI talk. And a very good one. Casey is an excellent speaker and I'm rather envious as even faced with difficult questions he manages to gentle wafted them away. Dan Friedman was up next talking about Class::DBI::DataMigration. It sounded like a very interesting module, and may well be worth investigating further in the coming months.

The evening was taken up with the boat cruise. The trip was a 5 hour circuit around the harbour with a bite to eat along the way. It was a chance to converse with some of the attendees I hadn't previously had the opportunity to meet, as well as some of the London.pm'ers who I'd seemed to miss for most of the conference so far. Mark Stosberg got to display his talents as a juggler, although I wish I'd know it was Mark while I was on the boat. It would have been nice to have a chat about a few things. This is the problem with being a newbie to a conference, you sometimes don't even know who some of the notorious/famous people are until either you're introduced, or you see them speak. Paris in 2003 started to change this by taking everyone's photo and pinning them all on a board. It made life a lot easier if there was someone specific you wanted to meet.

Later on we had the auction. I was too busy talking and missed a couple of the books I would have liked to have bid on. The big auction item was the first page of Damien's new book, Perl Best Practices. I have plans to buy it myself, but there was no way I could compete with the winning bid of CND 1500. Wow! We've had some amazing bids at YAPC::Europe before now, but for one person that is probably the highest bid I've ever seen. I was surprised that this was the only unusual bid though. There is often a great deal of tomfoolery going on in Europe, but the North Americans don't seem to like that kind of thing. No idea why, maybe it was just the setting, which made it a little awkward for an auction. It has to be said that Uri did a grand job though, and was a perfect opposite to both Greg McCarroll and Marty Pauley.

The evening ended with a trip to a pub by a few of us. Perl, politics, music and beer carried us into the night.