Pointer Records for Metadata

[JQ Johnson]

Brian Nelson argues that PDF should be extended to allow inclusion of
Dublin Core metadata elements.  I concur, but...

There are many applications where decoupling the metadata from the data
(and storing a pointer to the metadata in the data) is a desirable
approach.  Examples include cases where the authority for the metadata is
independent of the authorship of the article (e.g. PICS ratings,
publishing information associated with DOIs, some copyright management
information, etc.), cases where one wants to preserve metadata independent
of the format of the particular work (I want my metadata to survive if you
convert my work from HTML to a GIF image), cases where the metadata
reflects collection-level information but you want to reference it at the
item level (let's not repeat a bunch of collection-level data in 10,000
web pages, where we'll have to change every page if the metadata values
change), where for efficiency one wants to be able to search the
authoritative metadata without accessing the individual data, etc.

I believe these issues were hashed out quite thoroughly in the Dublin Core
meetings, though I didn't particpate there so someone who did might like
to comment.

Note that libraries typically store Dublin-Core-like metadata (catalog
entries) separately from the data (books).  Doing otherwise would make the
card (oops...showing my age) catalog substantially less useful.

JQ Johnson                      Office: 115F Knight Library
Academic Education Coordinator  mailto:jqj at darkwing.uoregon.edu
1299 University of Oregon       phone: 1-541-346-1746; -3485 fax
Eugene, OR  97403-1299          http://darkwing.uoregon.edu/~jqj/