Monday, July 28, 2008


Last Friday I spent all day at RepoCamp. There were at least 40 participants from I don't know how many institutions! Major kudos to David Flanders from the JISC Common Repository Interface Group (CRIG) who did a fabulous job organizing the event, and to Ed Summers who facilitated the LoC side.

There was a lot of great discussions around SWORD and OAI-ORE. I was happy to have the opportunity to talk about BagIt with a group who hadn't encountered it yet, and we had some really interesting discussions. Talking through use cases beyond our initial simple use case -- files from Institution A are transferred to Institution B and stored for preservation with no active access -- there is an obvious need for BagIt profiles that specify what is contained in a Bag and how it's organized for other uses -- like potentially as a SIP for ingest into a repository. Folks were also really interested in the idea of "Holey Bags" where the manifest is a list of URIs for retrieving files. Ideas were batted around about crawls that start out with a minimal manifest of URIs, capture those files, generate checksums, follow links from those files to capture more files and checksums, ending up with a Bag generated on-the-fly from that crawl so you have captured the files and record of the URIs where the files were found. A really interesting suggestion was the use of an OAI-ORE Resource Map to instantiate such a capture. Or for that matter, serve as the fetch file. Bags of course can simply be files on disk, but when it's a Bag of web resources you might want more structure than just a list of locations the files came from. After listening to the discussions I'm convinced that the work we're doing (I should say Ed is doing) with a web app for a Bag deposit service that uses SWORD is going in the right direction. I think were developing some real traction with BagIt.

They video recorded all the elevator pitches and the reports to the whole group. I don't know if, where, or when they'll be available. I am not a fan of seeing myself on video.

It was nice to see some of the Fedora team -- Sandy Payette, Dan Davis, Eddie Shin -- and some UVA colleagues. Its only been 3 or so months but it feels like I left so long ago. I was pleased to meet Ben O'Steen from Oxford (we were following his Fedora IR work when I was at UVA) but I didn't get the chance to really sit down and talk with him. I had so many other interesting conversations that I need to follow up on ...

No comments: