[Thursday] Sitting in the airport getting ready to head out; it’s been a while since I’ve spoken to over a thousand people (well, I do every day here, but that’s different) so I’m nervous. Happy, too; I’ve managed to schedule my Sun-related work so I can get to a few sessions that sound super-interesting; I expect to get my brain upgraded. [Friday] I’m filling in conference notes and pictures as the days go by.
Solr · I had breakfast with Erik Hatcher, who is one of the heavy-lifters in the Lucene world. He’s working on Solr, which is a REST interface to Lucene. Back in 2003, in my On Search, I called for exactly this, so obviously I think it’s a good idea.
Opening · Chad Fowler delivered an absolutely smashing welcome. He detoured from the obvious tech stuff, acknowledged that the Rails community has a reputation for arrogance (there was a scattering of applause; hmm) and challenged the roomful of geeks to get involved, give some money to a good cause, and otherwise make a difference. Go Chad!
Below is a picture of Chad and Rich Kilmer setting up for DHH’s keynote.
DHH · No obscenities, just a straightforward, entertaining state-of-Rails survey. The state is, more than anything else, RESTful. The new stuff in the new Rails is all about mapping code to resources and representations.
Also, there’s some lovely-looking debugging code that ought to be integrated with the stuff the IDE people are building.
Also, there’s a bunch of stuff to facilitate caching and HTTP performance. Which is an area that needs lots of attention; for example, Patrick Mueller illustrates apparent mis-use of HTTP by Twitter, which may well be contributing to its performance problems.
I talked to David after the keynote about what they’re doing with ETags, and the default Rails implementation is going to help save bandwidth, but do nothing to offload the CPU. I’m not sure that this is something that can be fixed in the framework; at the end of the day, only an application knows when it’s mangled some resource or another. But anything Rails can do to help programmers Do The Right Thing would be a major plus.
The Sessions · I enjoyed parts of Clean Code by Bob Martin and Doing REST Right by Scott Raymond. Both sessions were jam-packed; for example, check out this picture:
The crowd are, on the face of it nearly all Mac-toting males:
I haven’t had enough time to do the schmoozing yet to find out what kind of people they are, where they come from. But I’m really distressed to be attending yet another nearly-all-male event.
Trade Show · There was an actual trade show; I guess Rails is going mainstream:
Below is our own Arun Gupta demonstrating JMaki with Rails; apparently people liked it.
Avi and Ze · Avi Bryant, once again, challenged the Ruby crowd, pointing out that Smalltalk, a nearly-identical language, has immensely higher performance and a culture of extending Smalltalk in Smalltalk; so why can’t Ruby have the same? He’s right, I think.
Ze Frank closed the evening; I wasn’t expecting much, since his little video shows mostly didn’t reach me. But he was excellent; funny and also on-target, talking about the process of building his Web persona and presentation. There was some fairly standard stand-up comic fare in between, but it was crisp and his timing great; definitely a good way to close the evening.
Comment feed for ongoing:
From: Erik Hatcher (May 19 2007, at 06:36)
Tim - Thanks for the Solr mention! It's a project that deserves a lot of serious attention.
To add more about our breakfast conversation... Tim is a long-time library geek as well as, of course, a search-engine geek (I assert they are are a necessary combo!). We had a nice chat about MARC data and a few cringing laughs about TEI, the markup "standard" that everyone uses differently.
I pointed Tim to both NINES Collex - http://www.nines.org/collex - and Blacklight - http://aleph.lib.virginia.edu/blacklight
In peeking over Tim's shoulder has he explored Blacklight, I think he found what he was looking for right off the bat. As it should be.
Again, thanks for the mention Tim!
p.s. The link to me you provided was to the main code4lib.org site, but I'm still just a newcomer to this wonderful world of library geeks and humbly bow to the many great folks that make me proud to be considered a "library geek". I blog there under http://code4lib.org/erikhatcher
From: Patrick Mueller (May 19 2007, at 07:40)
I'd love for my son, Peter Mueller, to have made comments about ETags. Alas, no, it was me, hist father, Patrick Mueller :-)
BTW, I don't believe I claimed that not using ETags was causing performance problems; I've been around too long to make unfounded claims like that, without some evidence. It seems like an obvious thing to investigate though. Never know, maybe the extra cpu, or even extra data access to handle ETags will end up making performance problems even worse.
From: Tim (May 19 2007, at 12:25)
Oops, sorry, Patrick. Cleaned up both your name and what you are reported to have said.