Comments on: Data structures and Serializations

By: Jakob

Jakob — Fri, 06 Aug 2010 07:03:50 +0000

Yes, FRBR is one example of a Domain Model – librarians can do this. But with FRBR they failed to define a serialization. Domain Models help you to talk about things with human beings. But to exchange data you need a serialization of the model. I agree that the model must come first, but when you stop there, you end up doing no data exchange but philosophy (which is nice too).

By: Painting a picture of metadata « (d)atalog(ue)

Painting a picture of metadata « (d)atalog(ue) — Fri, 14 May 2010 14:33:41 +0000

[...] librarian would put their metadata in a data format (or “content format” or “data structure“). Some examples are binary or XML. It is the carrier for the content, just like how a CD [...]

By: Joe Montibello

Joe Montibello — Wed, 21 Apr 2010 20:32:12 +0000

Hi,

I’m not at all familiar with Domain Modelling – so far what I know about it comes from this blog post, plus a less useful Wikipedia article, plus a random white paper I googled up. (http://www.aptprocess.com/whitepapers/DomainModelling.pdf)

My question is this: would modelling the domain for a library system consist of coming up with something like the set of behaviors that FRBR describes, and then building a data structure based on that?

Thanks for an interesting post, anyway. Joe M.

By: MJ Suhonos

MJ Suhonos — Wed, 21 Apr 2010 02:08:56 +0000

Bill, just to clarify my perspective on the issue, I fully agree with everything you’re saying above. In fact, your explanation is probably the clearest I’ve seen to date. And the thread is definitely ridiculous. :-)

By: Jonathan Rochkind

Jonathan Rochkind — Tue, 20 Apr 2010 21:12:49 +0000

A brief exchange me and Bill had in IRC, which I think is further illuminating:

(5:10:13 PM) jrochkind: BillDueber: I’d say the problem is that MARC is BOTH a “data model” AND a “data structure.” Even though was never designed as a data model, it has become one.

(5:11:02 PM) BillDueber: jrochkind: Right. We long ago passed the point where the model drives the data structure. It’s [now] the other way around. [which is a bad thing]