Future of MARC and the Semantic Web

I've been blogging for about 6 months now and have 100+ library blogs in my reader. But one topic I'm really interested in doesn't get much press: the future of Machine-Readable Cataloging (MARC).

When it is mentioned, MARC usually gets a bad rap. It's often viewed as worn out legacy metadata better suited for card catalogs with an antiquated late 1960's data structure that mystifies computer programmers when they first encounter it.

However, lately I've read a few things that say otherwise. Some say MARC is very good metadata--it just needs to be morphed into something more Web-friendly. This more positive view describes MARC as a complex metadata format, with a good level of granularity--well suited for what it was intended for at the time it was developed.

Here's one of those voices. In a post primarily about libraries and the Semantic Web, Richard Wallis comments on MARC:

So does the library world pack up its knowledge and collaborative spirit and jump on the SW bandwagon, leaving the good, but old creaking and inflexible library standards such as Z39.50 and MARC, behind? - In the very long term probably yes, but over the next while what libraries are doing today, in the way they are doing it (with the exception of the format of the odd MARC tag), is still very valuable. [emphasis mine]

I would contend that, from a cataloguing point of view, libraries are providing more value than current software packages and their user interfaces can make full use of. Some of this value is being shown by the work around faceted browsing, FRBR and the like, but this is just scratching the surface. Transferring/transforming the data in to RDF, opens up opportunities to walk or browse through the semantic relationships expressed between records and external resources to deliver a more holistic view of a resource and its place in the world.

The discussions I have watched go past can be characterized; as a binary choice between RDF and MARC. That is wrong. As we, and others, have shown MARC is a rich source of data that can drive the expansion of the semantic web. [emphasis mine]

Comments

[Alex Grigg]

While MARC records certainly have their issues, the primary issue being that they are overly complicated for the uses to which they are put, the main problem I see with the format is that our OPACs don't do anything with the wealth of information those records contain. Most of the elements of FRBR have been included for decades, but it is only just recently that we are finding a useful way to display that information to library users.

[Chris Schwartz]

Thanks for this comment, Alex. You bring up an important point that most of the FRBR elements are already included in MARC records. I've been thinking about this lately (although not clear enough to blog about it) in relation to expanding FRBR's application in the new content standard, Resource Description and Access (RDA). I'm sort of feeling that some of this is an added burden to catalogers/metadata librarians that could really be handled by machine processing.

[Francois Scharffe]

Note that a transformation of MARC records to Dublin Core is available [1] via Marc XML. It is just about fields of MARC relevant to DC, but can be extended with mappings to specific ontologies. We are actually working in EASAIER [2] on an extension in the music domain.

[1] http://www.loc.gov/standards/marcxml/
[2] http://www.easaier.org

[Richard Wallis]

Thanks for quoting me.

On the subject of browsing semantic relationships extracted from already cataloged, in MARC, bibliographic data - in my presentation [http://blogs.talis.com/panlibus/archives/2007/10/access_2007.php] at Access 2007 [http://access2007.uvic.ca/] I showed a tool our developers use to do this in a graphical way.

I captured some screen shots in to the slides you will find from the above link. To get a more dynamic view, look out for the video when they get published on the Access site.

[Richard Wallis]

Seems that the URLs in my previous posting somehow got corrupted.

So I'll try again.

Access 2007: http://access2007.uvic.ca/
My Presentation: http://blogs.talis.com/panlibus/archives/2007/10/access_2007.php