Thursday, August 18, 2011

YAPC::EU 2011 - day 3

The last day started with Jesse Vincent's "Perl 5.16 and Beyond". I was expecting the same talk as in Asheville (YAPC::NA), but it wasn't. He explained his plans for a traditional and minimal Perl. The traditional Perl would be like today (with all the core modules). The minimal one would just contain enough to get CPAN started. That is a great change! Also, some functionality (UNIX process functions, sockets, formats) could get extracted into modules. This would lead to a leaner Perl C source. I absolutely like this decision!

Next was Steven Jenkins (from Venda) with "Modern Perl - getting there from here". The abstract sounded better than the talk actually was. He talked about some problematic areas in the Venda code base and what they did about it. Unfortunately it contained almost no general advice.

After that, I attended Thomas Klausner's talk about text fixtures. He talked about file system and database fixtures and refactoring tests with own modules. For database fixtures he recommended transactions (and rollback).

Before the lunch break Matt S. Trout talked about Data::Query. He started with the history of DBIx::Class and later talked about Data::Query and the future integration into DBIC. There is already a 'dq' branch at SQL::Abstract. He mentioned is still using a private Class::DBI fork. And I wrote down this quote: "Monkey patching is completely fair, if the line starts with 'local'." I like it! :)

In "I <3 my community" Mark Keating talked about last years London Perl Workshop and mentioned/thanked members of the Perl community.

Next was Clinton Gormley with "Terms of endearment - the ElasticSearch query language explained". He had a nice demo application we could try on his Laptop (also on GitHub). He compared (and explained) filters and queries.

Last talk before the lightning talks was Abigail with "The Business Aware Programmer". I did not attend this talk in Asheville, but I knew from a blog post that this talk was controversial. So I was looking forward to it (and indeed, Abigail could not finish in time because of upcoming discussions). I wrote down one quote from the audience: "Your crappy code makes smart people not want to work for", which sums up the disagreement nicely. If I find the tuits I will write a whole blog post about it, as some of his points are valid (but we developers want to believe otherwise).

The lightning talks included, Net::DRI, White Camel Awards and Acme::ASM.

Matt was closing the conference with his "State of the Velociraptor" talk. This time it was almost the same talk as in Asheville. A few attendees brought the WebGUI octopus with them and threw it at mst again. Even Larry landed a shot. :)

Thanks to the organizers for a fantastic conference. Next years YAPC::EU will be in Frankfurt/Main, Germany.

Wednesday, August 17, 2011

YAPC::EU 2011 - day 2

The second day started with Damian Conway's keynote "(Re)Developing in Perl 6" where he showed the Perl 6 versions of some of his modules. Usually the new version was one third of the size of the Perl 5 version. His message was: port your CPAN modules now to Perl 6.

Next was Valentin Guillois with "DBIx::Class::Tree::Mobius" about tree representations in SQL databases. He explained materialized paths, nested sets and nested intervals. His module implements the Moebius encoding, which is based on continued fractions. Unfortunately he could not finish his talk, so here are some materials:
I skipped the next two slots, because I was not particularly interested in any of the talks, and used the time to reharse my presentation (which was scheduled for later this day).

After the lunch I was attending Peter Rabbitson's "DBIx::Class guts 1.1 (or how SQL sausage is made)". I was looking forward to this talk, because I switched from Rose::DB::Object to DBIx::Class (and I still want to write a post about my reasons for this) and the initial learning curve is quite steep. Peter explained the four abstractions of DBIC. It made sense to me, but I have to do some homework and study it further.

Now it was "showtime" for me. I presented "Sanitizing HTML 5 with Perl 5". After all, it didn't went so well as in Asheville (YAPC::NA), but it was okay. There were a lot more attendees (thanks to the weak spot I was in - no heavy competition) and the participation (remarks/questions) was higher. I have doubt about the naming: HTML5::Sanitizer. The "HTML5" might be a bit misleading, so I will release it to CPAN as HTML::Sanitizer in the next weeks (and blog about it).

Next was Marcus Ramberg with "Mojolicious - A New Hope". I like Mojo and I don't like it at the same time. :) Marcus talked about Mojo::Client and Mojolicious::Lite. While I like the non-Mojolicious modules (like Mojo::Client), I'm not convinced of Mojolicious for bigger apps. All the examples are too simple. I would like to see an application with lots of controllers. But back to Marcus' talk: It was a nice introduction to the Mojo(licious) module family.

For the next talk the organizers ignored the "I will attend this talk" feature of ACT. Smylers "SSH Can Do That? Tips for Working More Productively with Remote Servers" was totally overbooked. For me the talk had familiar and new material, for others it was all common stuff. Years ago, I was attending a similar talk at the Chemnitz Linuxtage, which contained more stuff. So while it was nicely done, it could have been better.

Today's lightning talks included:
  • architecture of
  • 14th German Perl Workshop in Erlangen (3. - 5. March 2012)
  • "If you are not on IRC you do not exist" (where I disagree)
  • "Don't optimize for speed" (Paul Johnson) - was really nice
  • Ingy released Larry Wall's presentation software "Stump"
After that we were transfered to the social event, which was at Lido Restaurant. It seemed like we had too few seats, but we made the best out of it and had a good evening.

YAPC::EU 2011 - day 1

Actually, it's already day 3, but I want to keep the chronological order... :)

The main conference started with Larry's keynote. He showed some pictures of Riga and talked about (post)modern Perl.

I attended Carl Maesak's "How not to screw up your business application" next, which was about CQRS (command-query responsibility separation). The basic message was "it is okay to use more than one model" (a normalized one for writing and an denormalized/optimized one for reading).

After that Mallory von Achterberg talked about "HTML5: What it is, what it isn't, and should you use it?". Unfortunately, she could not show all the slides. Eye-opening was the counter until the HTML 5 spec is official, which showed over 3 years left.

Next was Carl Maesak again, with "Making data dance". He talked about the dancing links algorithm from Donald Knuth.

After the lunch break (we got free food at the venue), I attended Jonathon Worthington's "Rakudo Evolved: speed, feedback and hackability", which was about the Rakudo nom branch.

Chisel Wright's "Mostly lazy DBIx::Class testing" showed Test::DBIx::Class::Schema. He also uses Test::Aggregate to avoid the slow schema for every test file.

Next was Zefram with "why time is difficult". He talked about different time systems leading to UTC (coordinated universal time). He must be very interested in this topic as he could answer a lot of questions from the audience.

I missed the next slot, so next were the lightning talks. I just want to mention a few:
  • Leon Timmermann "Reinventing Build.PL" (throw away Module::Build)
  • Peter Rabbitson "Way Too Far Down The Memory Lane" (Sub::Name in pure Perl)
  • Dave Cross "The Perl Community - A Modest Proposal" (Perl Masons and the first rule of Perl Club - funny)
After the lightning talks was the auction. It was not as funny as in last YAPC's, when there were more fun items and competition between the auctioneers. The day ended with dinner in Riga's old city.

Sunday, August 14, 2011

YAPC::EU 2011 - day 0

Mmh, my blog seems to be mostly an YAPC blog this year...

Yesterday I arrived in Riga and took a round through the city centre. In the evening I met the first Perlmongers. Today was "Speaker Training" with Damian Conway and pre conferece meeting.

The speaker training was very interesting, too bad Damian could not give his tutorial last year (I booked it). My talks this year would have been so much better. My talk "Sanitizing HTML 5 with Perl 5" is scheduled for Tuesday - too short for me, to rework everything and apply all of Damian's wisdom.

I made notes on paper and I'm too tired now to type them here. I will do that when YAPC is over. So you can read them, if you are interested.

Tomorrow starts the "real" conference...