RailsConf Wrapup

Although I failed to write anything in this space during RailsConf, it’s not because I was overwhelmed or underprepared, as I had expected I might be. I was fairly well prepared for the conference (having used Rails before, which was about the only prerequisite for the talks I saw) and the talks available left me feeling just whelmed enough.

There are probably plenty of other comprehensive writeups of the conference out there (though I cannot link, as I am writing this offline in O’Hare Airport and they expect me to pay $6.95 for a few hours’ of access; yeah right), so I’m not going to do anything crazy like post a whole list of talks that I saw or (possibly crazier) post all of my notes, but there were several talks and speakers that I’d at least like to mention in this space.

First, the Keynotes. Several speakers throughout the weekend touched on the value of Rails as opinionated software, which, predictably, led to many varying interpretations of what those opinions should be. Dave Thomas’ excellent opening Keynote addressed three problems that “need” to be fixed in Rails; one of which was improved data integration, especially for legacy databases. Dave would like to see validations automatically applied to fields with schema constraints (which I don’t think anyone would really disagree with), FK support (which is a slightly more contentious issue), and support for composite PKs (which DHH openly derided in his keynote, surprise surprise). The other two concerns that Dave raised were much less contentious: one is a new, improved “Capistrano II” that would decouple the responsibilities of developing and configuring a production server, and the other is a request for less “Web 1.0” scaffolding (which, based on the content of his talk, I think DHH would readily embrace).

Friday night featured back-to-back Keynotes by Martin Fowler and Paul Graham. Martin’s talk began with the intriguing revelation that he had never actually used Rails; he said he’d asked his wife if she needed a db-backed web app, but she hasn’t needed one yet. Still, he enumerated the three things that interest him about Rails: its design, the conversations between developers and customers that it enables, and (naturally) Ruby itself. Martin struck what seemed to me to be a very different chord from Dave Thomas when he said that he hoped that Rails stayed focused on the “small” problem set that it was designed to address. It is Rails’ simplicity, he said, and its ability to create what he called “Quick and Clean” solutions that have already changed the face of software development. Even if Rails were to “die” tomorrow, he said, it’s already changed the world. Mission accomplished, DHH.

Paul Graham’s keynote took a different tack (and will likely appear online in a slightly different form soon) (later: and here it is. Naturally, his goal was to be provocative, and he stated such at the outset. “I’m going to disagree with both the Old Testament and Yoda tonight,” he said, and while the former seemed to cause a minor uproar at later talks, it’s the latter that really struck me as controversial. The crux of his talk was the value of “marginal” thought, not in the sense of less valuable, but rather less constrained by the chains of being an “insider”. To be sure, it was not a ground-breaking revelation that BigCos are significantly less likely to really change the world, but the delivery of his message was doubtless inspiring to a great many in the audience, myself included. In fact, I almost left afterward to go hack, but I would have greatly regretted missing Why The Lucky Stiff (who was mind-blowing and provocative in his own way).

The natural culmination of the weekend was Saturday night’s keynote by DHH himself, which consisted of 15 minutes or so of being arrogant and boastful (so as to not disappoint, he said), and about 75 minutes of insight into (what I assume to be) the major thrust of Rails 1.2: CRUD. Specifically, the use of those criminally under-used (and, in HTML, unimplemented) HTTP verbs PUT and DELETE. He showed off some of the sweet pretty-URL features that such a shift would buy (e.g. /people/1;edit, /people/1;destroy). His focus on the nifty new features and shift in thinking (i.e. Name Everything) coming in Rails was interesting, but not nearly as interesting to me as the anayses of why Rails is changing the world and how we can help to change it. I suppose it should have been expected from the programmer who invented the damn thing, but still, DHH’s keynote was the most disappointing of the weekend.

Beyond the keynotes, there were so many strong talks and speakers that I will inevitably shortchange them if I try to discuss them all here, so I’m going to just pick a few that really blew my mind. The first of the conference that I attended all weekend was Mike Clark’s Introduction to Capistrano, and as a Rails user who had not discovered Capistrano’s myriad sublime pleasures, it was a huge wakeup call. The one thing I absolutely cannot wait to try when I get back to a workable internet connection is cap update-code; cap deploy.

My single favorite talk of the weekend was Doug Fales’ Rails by the Waypoints, which was essentially a case study of mashups in the Rails age. His yet-to-be-released WalkingBoss is an application that I have pined for since Amanda and I returned from our road trip in 2004; since then, I have spent countless hours with the huge number of photos we took and Google Earth, reliving the strange and circuitous route we took around the country. I’ve already decided that we’ll be picking up a GPS device before our trip to Maine in a couple weeks and plotting our hikes when we get back.

Given its placement in the schedule and its attendance, Why The Lucky Stiff’s performance on Saturday night might well have been better described as a keynote, as calling it a concert almost does it a disservice. In the past, I’d seen CHUNKY BACON and never got it; now I get it. It almost seems unfair that such a brilliant programmer is also a great performer. His expose of Method#to_proc was masterful. And did you hear that David Heinemeier-Hansson died at the hands of Robert Scoble? gem install that!

Probably the single most inspiring and heartfelt talk of the weekend was Nathaniel Talbott’s on Homesteading on Saturday night. The parallels between the midwest of the 1800s and Web 2.0/The Long Tail today are strong. I expect for many people in the audience last night, it was just the push necessary to quit their jobs and go do something they love.

Finally, I’d be remiss if I didn’t mention Ken Kunz’ talk this morning about OLAP, Just the Facts (and Dimensions). I was totally ignorant of the differences between OLTP and OLAP this morning, and Ken did a tremendous job of explaining just enough to really get me excited, then introducing a plugin (acts_as_fact, available soon) to allow Rails to become a data warehouse.

Next year’s RailsConf was announced as taking place May 17-20, 2007, in Portland, OR, and O’Reilly’s going to be running it. Stuart Halloway, in an aside during his excellent talk on MetaRails this morning, urged us to enjoy this year’s RailsConf to the utmost, because next year’s is going to be very different: not necessarily worse, just different (e.g. wi-fi that actually works). I’m psyched to have been a part of the first RailsConf, and am really excited to be a part of a community that has such an amazing growth period ahead of it.