Metalogger

September 3, 2007

repository display, oai harvesters — omitting the document link from the oai-dc?

Filed under: Uncategorized — Neil Godfrey @ 6:48 am

When checking out the potential clashes in cases where a repository portal display configuration is tied to the oai-dc datastream, one question that came to mind was the possibility of omitting the http link to an offsite document entirely from the oai-dc. That way, the oai-dc data provider could offer the repository’s metadata page as the dc.identifier without any messy entanglements with assigning a different dc value for a document according to its location (archived in the repository or offsite).

But that option also has significant drawbacks. We have to assume that Service Providers will be able to index and search records from the data provided in the oai-dc. This must inform any Data Provider’s (repository’s) decision about what to include in the oai-dc. And some SPs expect to be able to search full text — which means they will need to know the link to the document itself, whether that is in the dc.identifier or some other value such as the dc.relation.

Advertisements

2 Comments

  1. Trying to meet both human and machine user needs for aggregated data entirely in simple DC over OAI is largely a losing proposition, as many aggregators have found. Using other metadata formats supplementing oai_dc is likely to be necessary. For one particularly robust option, see http://www.dlib.org/dlib/october06/cole/10cole.html.

    Comment by Jenn Riley — September 10, 2007 @ 1:02 am

  2. Just catching up again here and appreciate the link, Jenn. Thanks. (For anyone having trouble attempting to open the link, delete the full stop that got included at the end of the html bit.)

    Assume you don’t mind my using your input as one more bit of ammo when we approach certain repository developers to get smarter.

    Comment by neilgodfrey — September 11, 2007 @ 8:17 am


RSS feed for comments on this post.

%d bloggers like this: