A few notes from RubyConf 2007:
(The sessions at RubyConf 2007 were recorded by Confreaks, and will hopefully be available soon on the web. I'll update this post with a link when they are.)
Ropes: an alternative to Ruby strings by Eric Ivancich
Interesting. Strings have always been a problem area in various ways, SGI's ropes provides an interesting solution for some use cases. In particular, Strings always show up as pigs in Java memory usage, what with the many, long-ish class names and method signatures, that have lots of duplication internally; I wonder if something like ropes might help.
See notes by James Avery.
Ruby Town Hall by Matz
Didn't take any notes, but the one thing that stuck out for me was the guy from Texas Instruments who said they were thinking of putting Ruby in one of their new calculators. He asked Matz if future releases would be using the non-GPL'd regex library, as the TI lawyers were uncomfortable (or something, can't remember the exact words) with it (the licensing). See also the notes under Rubinius, below, regarding licensing.
But the big news for me, from this question, was the calculator with Ruby. Awesome, if it comes to be. I talked to the guy later and he indicated it was, as I guessed/hoped, the TI-Nspire. Sadly, he also indicated the calculator probably wouldn't be available till mid-2008.
See notes by my evil twin, Rick DeNatale.
IronRuby by John Lam
Didn't go into enough technical depth, sadly. And the technical presentation included info on the DLR and then XAML, which I don't think were really required. John had a devil's tail attached to his jacket or pants, which appeared about half way through the presentation. Really, I think everyone seemed to be quite open to IronRuby, no one seems to be suggesting it's evil or anything. Are they?
See notes by Nick Sieger.
JRuby by Charles Nutter and Thomas Enebo
Lots of good technical info. Tim Bray did a very brief announcement at the beginning about some kind of research Sun is doing with a university on multi-VM work; sounded like it didn't involve Java, and given the venue, I assume it had something to do with Ruby. Sounds like we'll hear more about it in the coming weeks.
One thing that jumped right out at me when the 'write a Ruby function in Java' capability was discussed, was how similar it seemed to me to what I've seen in terms of the capabilities of defining extensions in the PHP implementation provided in Project Zero. That deserves some closer investigation. It would be great if we could find some common ground here - perhaps a path to a nirvana of defining extension libraries for use with multiple Java-based scripting languages?
Rubinius by Evan Phoenix
Chad Fowler twittered: "Ruby community, if you're looking for a rock star, FYI it's Evan Phoenix. Please adjust accordingly. kthxbai".
I happened to hit the rubini.us site a few times this weekend, and at one point noticed the following at the bottom of the page: Distributed under the BSD license. It's been a while since I looked at the Ruby implementations, in terms of licensing, but I like the sound of this, because I know some of the other implementations' licenses were not completely permissive (see Ruby Town Hall above). Ruby has still not quite caught on in the BigCo environments yet, and I suspect business-friendly licensing may be needed to make that happen. It certainly won't hurt.
See notes by Nick Sieger.
Mac OS X Loves Ruby by Laurent Sansonetti
Oh boy, does it ever. Laurent covered some of the new stuff for Ruby in Leopard, and had people audibly oohing and ahhing. The most interesting was the Cocoa bridge, which allows you to build Cocoa apps in Ruby, using XCode, which (now?) supports Ruby (syntax highlighting, code completion?). Most of the oohing had to do with the capability of injecting the Ruby interpreter into running applications, and then controlling the application from the injector. Laurent's example was to inject Ruby into TextEdit, to create a little REPL environment, right in the editor. Lots of ooh for the scripting of Quartz Composer as well.
Apple also now has something called BridgeSupport which is a framework whereby existing C programming APIs (and Objective C?) are fully described in XML files, for use by frameworks like the Cocoa bridge, as well as code completion in XCode. That's fantastic. I've had to do this kind of thing several times over the years, and, assuming the ideas are 'open', it would be great to see more people step up to this, so we can stop hacking C header file parsers (for instance). And I think I could live with never having to write a java .class file reader again, thankyouverymuch.
I suspect all this stuff is available for Python as well.
Laurent also showed some of the DTrace support. No excuse not to look at DTrace now. Well, once I upgrade to Leopard anyway.
Someone asked "Will Ruby Cocoa run on the iPhone?" Laurent's reply: "Next question". Much laughter from the crowd. Funny, in a sad way, I guess.
See notes by Alex Payne.
Matz covered some overview material, mentioned Ruby will get enterprisey: "The suit people are surrounding us". He then dove into some of the stuff coming in 1.9. Most of it sounds great, except for the threading model moving from green threads to native threads, and a mysterious new loop-and-increment beast, which frankly looked a bit too magical to me. The green vs. native threads thing is personal preference of mine; I'd prefer that languages not be encumbered with the threading foibles provided by the platform they're running on. Green threads also give you much finer control over your threads. On the other hand, given our multi-core future, I think there's probably no way to avoid interacting with OS-level threads, at some level.
Behaviour Driven Development with RSpec by David Chelimsky and Dave Astels
I really need to catch up on this stuff, I'm way behind the times here. They showed some new work they were doing that better captured design aspects like stories, including executable stories, with a web interface that can be used to build the stories. That's going to be some fun stuff. Presentation available as a PDF.
See notes by Nick Sieger.
Controversy: Werewolf considered harmful?
Charles Nutter wonders if the ever-popular game is detracting from collaborative hack-fests. The game certainly is quite popular. I played one game, my first, and it was a bit nerve-wracking for me. But then, I was a werewolf, and the last one killed (the villagers won), the game came down to the final play, and I'm a lousy liar.
I kept notes again on a Moleskine Cahier pocket notebook, which works out great for me. Filled up about 3/4 of the 64 pages, writing primarily on the 'odd' pages, leaving the 'even' pages for other doodling, meta-notes, drawing lines, etc. I can get a couple of days in one notebook. The only downside is you need something to write on, for the Cahiers, for support, and the last half of the notebook pages are perforated. I don't really need the perforation, but it wasn't a big problem. They end up costing about $2 and change for each notebook.
I was, like usual, primarily surfing during the conference on my Nintendo DS with the Opera browser; good enough to twitter, check email, check Google Reader. It's a good conversation starter, also. At one point, a small-ish, slightly scruffy Asian gentleman leaned over my shoulder to see what in the world I was doing, so I gave him my little spiel on how it was usable for the simple stuff, yada yada. He seemed amused.