For the fourth year in a row, I spent an entire weekend in a hotel with a bunch of guys with laptops. It was RubyConf, of course, and as always I found it to be educational and thought-provoking.
When I went to my first RubyConf, in Austin in 2003, the scene was quite different: Rails didn’t exist, and most of the 50-or-so Rubyists in attendance had to program in another language at their day job. For RubyConf 2006, the 250 tickets sold out in a matter of hours. It’s a much bigger language now, and as it grows it tackles different problems on different scales.
One example: This year’s recurring theme was that of alternate implementations, as represented by non-Matz projects such as JRuby, RubyCLR, Rubinius, and Cardinal. Some of the attendees convened an implementers summit, to reduce duplication of effort and slowly converge on a unified test suite.
That these implementations are of Ruby 1.8 nicely complements the work on Ruby 2.0. In his keynote, Matz reiterated a point from the year before: That Ruby 1.8 is good enough a language to work in today, so delays in the arrival of Ruby 2.0 shouldn’t be a matter of concern. In the past I’d worried that this was just hand-waving—forget about new syntax, what about better garbage collection, concurrency, etc.?—but this year put many of my anxieties to rest. What Matz is leaving unsaid is that he prefers to be a language designer, not an implementer, and that Ruby is now big enough that Matz can fulfill a narrower role in its development. Harder implementation questions belong to others now, whether that’s Sasada Koichi, who’s in charge of the Ruby 2.0 virtual machine, or those working on the aforementioned alternative implementations of 1.8.
As the community grows, its makeup shifts slowly every year. One indication of these shifts was Joe O’Brien’s announcement of eRubyCon, the first ever enterprise-centric Ruby conference. Some of my friends in these Web 2.0 times might consider this a distasteful development, but I think real work gets done in what’s called “the enterprise”, and real work counts for a lot even when it’s less fun than spending angel money on an Ajax-driven social-networking music-recommendation podcasting-over-SMS startup. Besides, this isn’t the first change the Ruby community has gone through. During Masayoshi Takahashi’s “History of Ruby” talk, I was reminded of how strange it must have been to have been an original Rubyist in Japan when English-speaking Rubyists started to flood into the community. Once upon a time, Ruby’s core mailing lists were all in Japanese. Those days are behind us now.
The community is still unhealthy in one particular way: There are still far too few women. This year’s RubyConf had about five women out of 240 attendees—and while I suspect that our ratios may be only as bad as those in, say, the Perl and Python world, we shouldn’t content ourselves with that. Of course it’s hard to say how much men can do about this. I’ve started to prod the handful of women who come to Ruby-NYC to start their own local, women-only programming group, but what’s prodding good for, really?
As for the talks: Others have blogged extensively about them so I won’t go into much detail. The lightning talks were far better than they had any right to be. Zed Shaw’s talk on fuzzing was great, and his coining of the “that guy” meme may be the most enduring aspect of the entire conference. One day Ruby may be a passe language, but programmers will still be using the phrase “that guy” to heckle each other down at EnterpriseErlang 2023.
Rich Kilmer’s talk on Indi revealed a series of astounding hacks aimed at building a cross-platform, zero-installation Ruby on Rails and Flash app that runs off of a USB drive. (My company, Diversion Media, was kind enough to get a camera man to videotape both Rich’s and Matz’ talk.)
Tim Bray’s Unicode talk was entertaining and insightful: I expected from no less given the phenomenally sharp writing on his blog. He argued that it’s practically impossible to do automatic capitalization and decapitalization in an automated manner: This provoked a surprising amount of pushback in the Q&A that followed. A debate that isn’t finished, to be sure.
And the conference itself? Still fun, still inspiring, still thought-provoking, still a little exhausting. I would’ve paid extra for reliable wifi and electricity, and for it not to be out in the ‘burbs. As for size, 240 people seemed to be the right balance between keeping it familiar and letting people in. The tickets may have sold out fast, but among the folks I knew everybody who absolutely had to get a ticket (and could afford it) got one. I can’t imagine this can last forever, but in the meantime, I’ll see some of you in 2007, somewhere on the East Coast.