Dear all,
The most recent telecon we held about DCAM, last August [1], concluded with
some tentative agreement about the need for something like a Serialization
Profile -- something which would define the relationship between a Description
Set Profile, in the abstract sense, and various serializations, or even string
formats.
Looking back over our discussions [2], I think that some of our difficulty in
sustaining momentum on this topic has been due to the lack of a key use case to
drive discussion. ISBD provided one use case, and there are issues around the
use of Syntax Encoding Schemes that need further attention. However, the ideal
use case would involve a compelling need to enable communities of practice to
define new profiles, plus running code to transform those profiles into
displays, data validators, input forms, and the like.
The Bibframe project has proposed the notion of a "Community Profile" ([3],
slide 54). A Community Profile, as I understand it, would specify a template
for describing things like Book, Paperback, and Hardcover ([3], slides 38 and
40). While still in early stages, the Bibframe project is approaching the
problem by testing simple examples of transformations from MARC into RDF [4].
Whatever problems the core Bibframe model may still have at this stage, I very
much like the incremental, test-driven approach. If Bibframe is to replace
MARC, then as I see it, the future of, say, RDA will depend on the ability to
express RDA in Bibframe, which means the ability to express RDA in Community
Profiles. Inasmuch a Community Profile is conceptually close to the DCMI
notion of an Application Profile, it means finally getting around to making
application profiles for RDA -- a task first set as a goal way back in May 2007
[5].
I propose that we re-boot the DCAM discussion, but that we set DCAM as an
abstract framework (and an acronym) aside for now and frame the discussion as
one about how to template Community Profiles. I actually like the term
Community Profiles alot better than Application Profiles because it captures
the notion, central to the concept since its introduction into the DC context
in 2000, of the specification of a community consensus.
I expect that this discussion might circle back to the notion of a
Serialization Profile as the implementation-oriented complement of a Community
Profile, but I suggest we not try to derive generalizations (i.e., "abstract
models") before looking at specific, concrete requirements, such as those
emerging in the context of Bibframe.
Do you agree, and if so, what could be our next steps?
Tom
[1] http://wiki.dublincore.org/index.php/DCAM_Revision/TeleconReport-20120814
[2] http://wiki.dublincore.org/index.php/DCAM_Revision
[3] http://www.slideshare.net/slideshow/embed_code/16143235
[4] https://github.com/lcnetdev/marc2bibframe
[5] http://www.bl.uk/bibliographic/meeting.html
--
Tom Baker <[log in to unmask]>
|