ECLM 2011

Contrary to my expectations earlier this year, ECLM 2011 happened and ~90 people turned up. Yay!

Saturday, a bunch of us went on a guided tour across Amsterdam. We had a great guide with an awesome sense of humor and this was a nice way to meet new Lispers and catching up with folks I'd met in past meetings.

Dinner was fun. One very prominent Lisper who shall remain anonymous didn't know what Quicklisp was. Everyone sitting at that particular table was shocked. :-)

Sunday started off with Nick Levine talking about learnt lessons from his failed attempt at writing a Lisp book. Midway through, he took the opportunity to rant a little bit about how there were at least 3 overlapping, incomplete websites (cliki, c-l.net, cl-user.net) and how the solution would be to, huh, create yet another one. The consensus seems to be that cliki is the one that's best suited for a starting point, but Nick complains that its focus on Free Software Lisps is too narrow. Anyway, cliki could indeed use a revamp.

Jack Harper talked about using Lisp on portable fingerprint scanners. He had great things to say about Lispworks which, by the way, had 3 people at this year's meeting.

Luke Gorrie did an extended version of the presentation on Teclo I had seen earlier this year in Zürich. This one included even more interesting diagrams, in particular some illustrating the contrast between 3G network performance before and after enabling their product.

After lunch, Dave Cooper talked about his GDL product, which includes yet another object constraint-ish system framework thingie.

Hans Hübner picked a controversial topic: code style and conventions. (A bit reminiscent of Norvig and Pitman's slides.) He picked on the veritable 80-column rule (blasphemy!) and pushed for project/company style guides. The discussion was entertaining. The general conclusion I drew from that is that while it might be useful to write the more important points down, it's even better to programmatically enforce the more important ones like tabs vs. spaces and trailing whitespace. In my experience, good programmers pick up and follow (or challenge) code style fairly quickly without needing to go through an explicit style guide.

Paul Miller demoed his company's data analysis tool written in Lispworks and talked a little bit about how it interacted and integrated with things like Excel via COM.

Lisp hero extraordinaire Xach Beane did an awesome presentation on Quicklisp detailing some of his implementation strategies, what problems Quicklisp purports to solve, its social impact on the Lisp community, and what his vision for the future is. I was particularly excited about his plan to enable hacking on random project à la clbuild. Definitely the juiciest talk in the meeting in my opinion.

Finally, there were lots of lightning talks this year:

  • Marco Antoniotti announced ELS 2012.
  • Christophe Rhodes's talked about R's lispiness and demoed his swankr project that brings R to SLIME and has nifty features like graphical presentations.
  • Erik Huelsmann announced ABCL 1.0! (There were three quite enthusiastic ABCL developers at the meeting.)
  • Pierre-Yves Baccou shared some thoughts on Zen, his X-server in ~5K lines of CL code.
  • ... and many others including a valiant attempt at subverting the 5+2 minute rule.

Drinks, dinner, then drinks again. Hanging out with Lispers is the best part of ECLM and there was plenty of that. Looking forward to the next one already!

0 komentoj: