George, thanks, that's very informative.
Steve's wikipedia reference also very interesting.
I note there a statement I find difficult to understand fully:
"Metadata is more properly called ontology or schema
when it is of broad or narrow utility."
and following on to the definition of ontology
http://en.wikipedia.org/wiki/Ontology_%28computer_science%29
the debates about upper ontologies are fascinating.
At 12:42 2005-05-24, George Macgregor wrote:
>Yes! If users would find such access points useful then they are perfectly
>acceptable. My only proviso would be that the item needs to be adequately
>recorded in the catalogue and has the necessary descriptive metadata (i.e.
>not just size, power, features, but a statement of responsibility, general
>material designation, etc.). Or, are you suggesting that there should only
>be access points and no descriptive metadata??? Hmmm, now that would be
>controversial!
I can see how one might conceptualise that kind of information as
metadata of a search term - while at the same time being
not metadata but attributes of the washing machine.
"Search metadata", OK, "metadata" by itself, not so.
Not every attribute of the washing machine is relevant to
(an actual person's) search, so not all attributes would be metadata.
>FRBR goes some way to clarifying your metadata categories, Simon.
>FRBR is expected to make big waves in the library community, and I expect
>the shocks will be felt by everyone in metadata-land, not just in libraries!
>Incidentally, FRBR relies on controlled vocabularies and the use of
>authority files, but let's not get into that again! ;-)
Indeed, very tasty, I like FRBR, as it looks like some clear thinkers
have been in there, and managed to overcome the obfuscators!
Off the top of my head, the lessons I might draw from FRBR are
(apologies in advance for possible terminology gaffes)
- have a good ontological domain model
- understand what are the potentially relevant attributes of objects
- clarify the appropriate relationships and *their* attributes
Looks to me as if FRBR not only removes the need
to use the term "metadata" altogether, but might
also be a conceptual framework which would shed light
on the washing machine issue.
Honest, I have absolutely nothing to do with library science...
(obviously a missed vocation :-) )
Regards
Simon
|