Friday, October 05, 2007

Brief summary from JAOO 2007

Last year was my first time at JAOO, and I ended up doing a short buzzword-based review of the overall conference. This year I had slightly different feelings about it, but I think that keeping the same format will still be valuable.

  • TDD: Lots of attention on the topic. Robert C. Martin held the opening keynote and a dedicated tutorial. An interesting – and controversial – issue popped out during James R. Coplien speech about "frictions" between a sound architecture and a full TDD approach.
  • Web 2.0: whatever it means, there was a lot of hype about it. AJAX, interactivity, Ruby, Grails, and in general the possibilities of development where the world is your stage are definitely interesting, and we're still far from seeing the final destination. Really hot topic.
  • Ruby: one year after, what looked a promising language really hit the mainstream. Most of the attendants were still on Java, but many of those were having a Ruby project going on. And they all were happy. There are still a few na├»ve elements, but they don't change the overall picture of a really successful language.
  • Erlang: the language has been designed for parallel processing, and has truly brilliant ideas in it. Google bases its computational power on MapReduce, but also a lot of Social Software (more than I imagined) is putting strong effort on parallel computation as well.
  • SCRUM: SCRUM trainings were completely sold out. SCRUM methodology itself deserved a whole – crowded – track.
  • Agility: while only a few speeches explicitly talked about agile, still it is an interesting issue. Developers listened to other's developers trying to measure how agile they were an how they could improve. It's definitely a path and not a label, so talking about it still matters.
  • Domain Driven Design and Intentional Software: from different angles they both try to tackle the cultural mismatch between Domain Expert and Domain Modelers. Eric Evans by presenting techniques and a clean conceptualization of the steps needed to have the modelers successfully interact with the experts. Charles Simonyi's Intentional Software takes the other way round, by presenting tools that allows domain expert to express the software in their own Domain Specific Language, sort of MDA without all of those meta- words. Both approaches have strengths and weaknesses: I am a DDD fan, but I'll keep focused on both.
  • SOA: this acronym looked dead and buried. There was a SOA track in 2006, and this year only a few mentioned it. I have the feelings that SOA was – and still is – a great idea, but completely messed up by big vendors (or big buyers). Sort of what happened to UML some time ago…
  • Software Architecture: I followed part of the track. I ended up having a broader vision of the overall possibilities (I've really been stuck with JEE for too long).
  • Ethics: a lot of speeches focused on ethical aspects of the software development process. "The Pragmatic Programmer" was the reference book of the whole conference, giving a somewhat introspective feeling. Maybe we didn't learn that much about the new cool stuff, as much as what we've learnt about who we are and what our job is.
  • Semantic Web: unfortunately it wasn't even mentioned.

After all, I was a little less excited than the year before, maybe because I missed the surprise effect, and also because some speeches were resounding the ones of the year before. There were no astonishing news on the stage, but we were not looking for the new messiah, so I guess this is pretty good.

Post a Comment