It’s done! The third day was terrific. We knew we were putting the last two speakers (Stan Ruecker and Michael Geist) in a tough slot, having to try to engage a tired, overstimulated audience, but both came through for us. How often do you see a dozen questions for the final speaker in a three-day conference? And no one walked out.

Our session went well, though I felt bad that my laptop sabotaged some of Art’s special effects. I meant to reboot it before we started, but forgot; it bogged down (thanks, Windows) and made my Firefox extension look slower than it really is. Oh well.

More random thoughts:

  • Lorcan Dempsey’s keynote, following on the first two days of the conference, consolidated for me some of the directions we need to move. We’ve been saying for a while that we need to put our services where the users are, but it’s becoming clearer what that means and how we should do it. Turning the library inside out, uncoupling all our services, putting them out there in the form of lightweight, RESTful web services: we can do that. And we can start now.
  • I think I really get COinS-PMH now. Really. And as part of the idea of a registry of library online services (as Jeremy is working on in the Ockham Digital Library Services Register), it makes terrific sense. Now I want good, seamlessly self-installing tools to take advantage of the different services I’ll be discovering as I browse. I want to be able to generate a profile of all the tools I use and store it on a server, so that I can connect to it from a public-access workstation and have my heavily customized environment available to me. Firefox extensions and GreaseMonkey scripts will look primitive in comparison to the browser-based tools we should have in few years.
  • Stan Ruecker’s visualization experiments were stunning; I love the one where a vertical lines swipes across a screen-full of tiny text from left to right, greying out the terms you don’t want and enlarging the ones you do, and then swipes back and pushes the hits into the left margin. I can’t wait to get some of these in a form where I can put them in front of our data. Modularity!
  • One current task is to add an OAI-PMH interface to an existing web search interface. I shouldn’t let myself get into this position again: I should build the service interface first, with unit testing (including radioactive records), and only do the user when that’s all done.
  • I’m tired.
  • Subversion is wonderful. I had started to add a feature to the Firefox extension I built as a demo, but I ran out of time to finish it. I had changed half-a-dozen files and thoroughly broken the thing. But rolling back to the working version was the work of a moment. Which is about all I had.
  • I must get serious about METS. And RDF. And Python. And SRU. And OpenSearch.
  • It should be lovely in Ottawa this time next year.