U of T, IFLA, xISBN
- University of Toronto Libraries, which I think is the third biggest university library system in North America, launched their new Endeca-based catalogue. Congratulations to them for doing a very nice job. Alastair Boyd, head of cataloguing, said they did FRBRization at the Expression level, not the Work level. His example was a search for ariadne auf naxos, where you can see the libretto and score with multiple Manifestations underneath each.
- IFLA Cataloguing Section’s Annual Report 2008 (74 KB PDF). They’re the people from whom FRBR, FRAD, and FRSAR emerged. The minutes of the meetings in Québec City in August 2008 (244 KB PDF) are up too. (I thought they were new but I noticed I linked to them last November.)
- Xiaoming Liu posted about Guessing Publisher from ISBN Prefix, a new addition to xISBN.
Do you know why the U of T decided to go with expression-level collocation? I guess my gut instinct would’ve been to go with work-level collocation with the ability to narrow to expressions, but it’s hard to know without seeing what it would look like.
I did try a couple searches in their catalog where I wonder if the expression-level collocation may be negatively impacting results (tho it may just be the searches I picked).
I tried a search on “Shakespeare midsummer night’s dream.” On the first page, I don’t see a hit for the actual text of the play except under “Shakespeare, William, 1564-1616. Plays. Selections. (3 records)” I don’t know how their relevancy ranking is working, but I wonder if incorporating some work-level collocation would boost up the actual work in an example like this. Or perhaps they think people are more likely to be looking for criticism?
I then did a search for Zen and limited by author to Dogen. The first hit is
“Dogen, 1200-1253. Shobo genzo. Selections English. (2 records)”
which includes
“Shobogenzo, Zen essays / by Dogen ; translated by Thomas Cleary” and
“Moon in a dewdrop / writings of Zen master Dogen ; edited by Kazuaki Tanahashi ; translated by Robert Aitken … [et al.].”
Oddly enough, the second hit on the top list is
“Shobogenzo, Zen essays / Paperback ed. by Dogen ; translated by Thomas Cleary,”
which is presumably the same expression as the first one in the above subset whereas I would not have thought of the Tanahashi version/selection as an identical expression to the Cleary version/selection.
It seems to me that this shows the danger of relying on 240 uniform titles for collocation when they are so inconsistently applied and also that, since selections can mean different selections, putting those together might lead to counterintuitive collocation. The way they’ve done it also hides the title the user is likely to recognize–I can see value in somehow displaying at least some of the related 245 fields at the top level.
That said, I applaud the fact that they’re experimenting with using FRBR to try to usefully condense the number of records initially displayed and attempt to collocate things that are basically the same to the user.
Comment by Kelley McGrath — 22 January 2009 @ 3:30 pm