Last week I went to my first Code4Lib conference. It was exciting, and also a bit overwhelming.

On Tuesday, Joshua, John Hauser from PALINET, and I presented at a Koha Camp, where I introduced a Koha 3 LiveCD (the first LiveCD I ever built!) and met one of the French Koha developers, Henri-Damien Laurent. It was Henri’s first time at a Code4Lib too, and he was inspired to blog for the first time. On Thursday, I gave a lightning talk on distributed cataloging, about which more anon.

There are already various good summaries of many of the presentations at Code4Lib, so I will focus on one session that particularly struck my interest, Karen Coyle’s keynote last Wednesday on RDA.

First, a capsule history: RDA started out as AACR3, that is, the third revision of the Anglo-American Cataloging Rules, a hefty tome that has governed cataloging practice in the UK, US, and Canada. The main goal of the revision was to extend the cataloging rules to handle virtual objects, e.g., web pages, but it was decided to take the opportunity to completely revamp cataloging. To express the discontinuity, the name was changed from AACR3 to Resource Description and Access, or RDA.

RDA had several design goals, including using FRBR as the basic conceptual model, simplifying the cataloging rules, and making the rules and resulting metadata be useful to other metadata communities such as archives and museums.

One of Karen’s points is that AACR2 is a set of rules to create highly structured and formalized text strings to express metadata about the objects of library cataloging. The AACR2 book weighs in at 600 pages, and based on the drafts of RDA chapters that have come ought, Karen estimates that RDA will have about 800 pages [and is this simpler?] and may end up being a standard … for creating highly structured and formalized text strings.

One of the issues I’ve dealt with working with library data is that these text strings, particularly as stored in MARC21 records, are not structured enough. For example, parsing a personal name from a 100$a requires looking for commas and praying. UKMARC, on the other hand, was a bit better: for western names, the surname was found in 100$a and the forename in 100$h. In addition, as with UNIMARC, the ISBD punctuation was not stored in the UKMARC record, but was meant to be generated for display purposes. I wish more aspects of UKMARC had been retained when the British Library switched to MARC21.

However, something exciting happened in 2007 — the RDA Joint Steering Committee and the Dublin Core Metadata Initiative got together and started working on combining RDA with Semantic Web technologies such as RDF and SKOS. This would allow cataloging to move beyond the straitjacket of structured text strings — RDA concepts would be expressed in terms of an RDF schema, vocabulary lists such as the language codes would be represented in SKOS and through the use of registries such as NSDL could be more easily and quickly extended, and library metadata could be more easily mashed up with other metadata. See this post by Karen Schneider for a good summary of the collaboration and some of its implications, and the DCMI/RDA Task Group’s wiki for the latest action.

Of course, a shift to RDF and a more formalized cataloging model would be a drastic change, and raises, as Karen Coyle calls them, some Frequently Unanswered Questions:

• Is FRBR the right conceptual model?
• Is RDF the right format?
• What happens to MARC?
• Who will be in charge? How will things be decided?

There are other questions — how will library catalogs, data stores, and ILSs have to change if RDA over RDF transpires? How will the existing corpus of MARC data be migrated? How will the systems vendors, open source projects, and cataloging vendors adapt, to say nothing of catalogers, cataloging teachers, and library schools?

Interesting times — and I now have a lot of reading up on and playing around with RDF to do.

Update: Karen has put up the slides [PDF] from her talk.

This entry was posted on Monday, March 3rd, 2008 at 12:21 am by Galen Charlton and is filed under Code4Lib. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.