Comments on: Why RDA is doomed to failure

By: Bruce

Bruce — Fri, 21 May 2010 23:54:29 +0000

If you wonder whether this disconnect between computer science and library science (specifically cataloging) is real, stroll down your QA76 range of shelves sometime and marvel at the distribution of shelving locations for something like Oracle how-to books.

By: Irvin Flack

Irvin Flack — Wed, 05 May 2010 05:56:33 +0000

Following up on Karen’s and Ross’s comments, I’m reminded about that joke about the guy looking for his lost keys under the street light — not because that’s where he dropped them but because that’s where he could see. Or, to throw in another metaphor: you visit a surgeon, you get an operation. Cataloguers are experts on the rules — so that’s what RDA at heart still is, a set of rules.

By: Karen Coyle

Karen Coyle — Tue, 27 Apr 2010 15:47:15 +0000

If you look at the diagram called “Singapore Framework” on the Dublin Core site [1], it illustrates all of the necessary elements of a functioning, modern metadata scheme. The framework is based on RDF, but it could really be based on any other foundation technology. What we don’t seem to have learned in the library world is that the cataloging rules do not a metadata schema make. The rules are about how you make decisions, but you need to have defined data elements, vocabularies, and, above all, you need to have some sense of what functionality you wish your metadata to support. I feel like we go about it entirely backwards, first creating rules, then trying to fit it all into a data format.

[1]http://dublincore.org/documents/singapore-framework/

By: George Duimovich

George Duimovich — Tue, 27 Apr 2010 14:02:45 +0000

In “Directions in Metadata” Karen Coyle notes that the current vendors have been reporting near ZERO feedback / customer demand for anything related to RDA. True, it’s still early – the spec hasn’t been formally released – but in a slow moving community, any change seems to need a lot of “ramp up” time, for both the library community and its vendors.

Very too bad, since there’s a sense of urgency that’s missing in all of this discussion. I think the OSS community is going to shape up to be best positioned to respond to changes, but moving forward with some reasonable consensus from libraries is going to be the challenge. There still remains a gulf between the well-informed IT & catalogers vs. the laggards from the catalog card generation who don’t understand how our MARC/AACR2 standards present huge data issues that prevent us from moving forward.

By: Wally Grotophorst

Wally Grotophorst — Fri, 23 Apr 2010 16:24:37 +0000

By: Laura

Laura — Fri, 23 Apr 2010 16:23:10 +0000

One minor quibble. RDA is intended to be a replacement for AACR2 — a descriptive standard, rather than MARC — a transmission standard. Granted MARC has evolved over the years to do both description and transmission in practice since there have been rules akin to application profiles in terms of how to enter data into a MARC record.

By: Ross Singer

Ross Singer — Fri, 23 Apr 2010 15:13:26 +0000

Hmm, there’s a lot here and while I think some of this would be easier to talk about synchronously, you have to go with the forum you have, not the forum you want.

First off, let me put it on the record that I don’t disagree with your thesis. I can’t say whether or not RDA will fail (or what that “failure” or “success” means, really) but its timidity in actually modeling the data leaves a lot to be desired.

Now, on to your arguments… Equality (with regards to information) is always going to be subjective. Witness the agita that owl:sameAs is currently wreaking on the Linked Data universe (esp. the hardcore semantic web set) to see. Machine based linking is always going to have error. Homonyms, mistaken assumptions, and human error are just going to have to be accounted for. Without a doubt RDA need to drop the string matching qualities of the status quo in MARC/AACR2 in favor of real identifiers. Still, this isn’t going solve the equality issue 100% because, honestly, a cataloger may not be 100% sure of what s/he is describing.

Also, abbreviated titles are actual things. Like “JAMA”. I’m not sure the actual provenance of these titles, but they are distinct from the actual title (and generally considered important and used).

My last point would be how you compare “RDA” and “MARC” in your last part. Really, you’re comparing RDA with AACR2 (esp. since the powers that be are trying to figure how RDA will be transmitted via MARC). The major issue is that RDA doesn’t distance itself nearly enough from AACR2 to be entirely worthwhile. Everything is still a literal and there is still a very “record-centric” mindset (even in the RDF schemas). This is most obvious when you see things like “titleOfTheWork” and “projectionOfCartographicContentExpression” instead of, I don’t know, just modeling the damned FRBR entities like they should.

So, instead, we have a somewhat-major change in cataloging rules that will require a lot of time and energy and still provide no “real” relationships between resources and entities.