RubyConf 2005, Day 1
So, I don’t mean to be a fanboy, but holy shit, dude, fucking Martin Fowler is here. And Sam Ruby, so I hear, though I haven’t spotted him. Martin Fowler and Sam Ruby come to your conference, man, that’s the big time.
Ruby’s going through some changes, and only time will tell how they play out. Largely due to the buzz behind Rails, this year’s RubyConf has 200-some people at it; three times as many as the year before. There are also, I’m told, five women in attendance, which is sadly enough five more than came last year or any year before that.
I gave the first talk this morning, on Top-to-bottom Testing in Ruby. You can listen in here, and read a few commentaries here and here. I can’t say how successful the talk was, but during the Q&A there was a spirited discussion, involving DHH and others, so I suppose that’s a good sign.
As a side note, I wish I’d taken a closer look at HighLine before giving my talk; I probably would’ve used it in my command-line code example instead of EasyPrompt. After my talk, Gregory Brown was telling me that HighLine was quite conscious about stealing the best bits from EasyPrompt and from Cmd—we took a quick look together, and it looks like EasyPrompt may be more or less obsolete due to Gregory and James Edward Gray II’s work. That’s fine with me; I’ve got too many tiny libraries on my plate as it is.
Next, Akira Tanaka talked about open-uri and the Virtual File System. It was a sort of meandering-yet-informative talk, spending a lot of time on general principles of making an API easy to use and remember. His points about HTTP as a kind of network file system reminded me that I have yet to write that lib to override require to read files as URIs—though that’s probably just a solution looking for a problem …
After lunch, a power-troika of dense, language-level talks. Charles Nutter talked about JRuby, an implementation of Ruby in Java that’s complete enough to be used by certain Java projects already. Koichi Sasada followed up on last year’s talk on YARV with a progress report. And Eric Hodel talked about Seattle.rb’s progress with MetaRuby, an effort to implement Ruby in Ruby.
Charles and Eric’s talks touched on many similar issues. Eric said that writing tests has been difficult since you can’t run Test::Unit without most of Ruby working in the first place. And both projects highlight what would appear to be a need for a more formal Ruby specification. Of course, somebody has to write that spec, and writing specs aren’t fun. But maybe that’s part and parcel of the language’s growth; there’s more housekeeping.
I presented the question to Matz at his evening roundtable; he said that if somebody wants to contact him about writing a spec for 1.8, he’d be interested. 1.9 and 2.0, of course, are far too fluid to be worth spec’ing out at this point.
As the community grows, it’s interesting to see which roles get filled and which are left vacant. More than a few projects which are hard and non-glamorous get left by the wayside, but meanwhile everybody’s put out a web framework … I’m certainly not saying I’m any better than anyone else when it comes to this, or else I’d be snatching leadership of Ruby-DBI out of Daniel Berger’s hands. (From what he’s blogged and our conversations today, it’s fair to say he’d be happy to hand it off.) But you never know. Maybe more tests can get added to the Rubicon. Maybe people will get together this weekend and form a specifications team. And Patrick’s making good progress on his Ruby-Web library, so there’s one bright spot.
At any rate, now it’s 11 p.m. and there are about 60 people tap-tapping away on their laptops in the hotel lobby, and out on the patio, by the pool. Last year, they kicked us out of the lobby at about 1 a.m. We’ll see how long we last tonight …