[OAI-implementers] pseudo-qualifiers in simple DC

[Riley, Jenn]

Hello all-

I'm looking for opinions from OAI service providers on the use of
"pseudo-qualifiers" in simple DC values exposed via OAI. By
"pseudo-qualifiers" I mean the inclusion of strings indicating element
refinements or encoding schems (the same sorts of things the real
qualifiers in qualified DC do) within the actual simple DC element
content string.  The purpose of doing this could be two-fold: 1) making
the simple DC metadata more intelligible for end-users seeing metadata
records, and 2) allowing service providers to use this information
intelligently to increase the quality of service they can provide.

We have relationships with some service providers, so in those cases
we'd know if they'd be able to use metadata formats more robust than DC
if we exposed them, and how they'd want simple DC elements structured to
use these "pseudo-qualifiers." But I'm concerned that tailoring our
simple DC for a few service providers would reduce the utility of the
records for the larger OAI community. Basically I have three questions:

1) What's the general opinion of service providers out there on
including pseudo-qualifiers in DC elements?
2) If these pseudo-qualifiers are used, is it best to put them at the
beginning or end of the element content?
3) If these pseudo-qualifiers are used, what characters are best to use
to delimit the pseudo-qualifier from the rest of the element content?

I imagine the answers to these questions would differ among DC elements.
It seems pretty straightforward that:

<dc:relation>IsPartOf
http://www.dlib.indiana.edu/collections/cushman</dc:relation>

would be OK. Including the IsPartOf pseudo-qualifier wouldn't affect the
end-user experience, and the URL is recognizable easily for service
provider processing. However, qualifying a subject or contributor might
be more problematic:

<dc:subject>[TGM I] Bodies of water</dc:subject>
<dc:contributor>Berlin, Irving [composer]</dc:contributor>

If a service provider were to not do any post-processing of these string
to remove and/or intelligently use these pseudo-qualifiers, they could
interfere with effective combination of metadata from multiple sources.
If the qualifier is at the beginning, a subject or name alphabetical
browse would be ineffective. If the qualifier is at the either end,
collocation of all resources with the same subject or contributor would
be affected. But the utility of the record, especially for the
pseudo-qualified contributor, is greatly increased for the end-user if
this pseudo-qualifier is included.

So what are most service providers out there doing? Do you work with
data providers in your area of interest to define mutually agreed-upon
usage of pseudo-qualifiers? Do you look for them in harvested data? Do
you find they foul up your carefully-laid plans?

Jenn

========================
Jenn Riley
Metadata Librarian
Digital Library Program
Indiana University - Bloomington
Main Library E170
(812) 856-5759
www.dlib.indiana.edu