As Peter Winn has raised the issue of metadata that enhances the discovery
of humans, I'd like to explore a little further (despite the fact that it
necessitates me thinking of my colleagues as "resources" :-(
I am investigating the creation and maintenance of a database of Australian
chemists (or chemically-interested persons ;-). Is there already a
recognised schema out there for the description of personnel? Despite
Simon Cox's reply to Peter, I'm not yet clear whether DC is appropriate or
not. It would be very convenient if it was because we have good tools to
create and manage DC metadata.
At 9:05 AM +0800 on 21/7/98,Simon Cox wrote in reply to Peter Winn:
>Peter - I think that there is a major flaw in your example,
>concerning the DC.Identifier elements. Your example metadata
>purports to describe a person, whereas the identifiers that
>you give identify
>1. a web page
>2. an e-mail address.
>
>Neither the web-page or e-mail address are identifiers for
>instantiations of Stu. Arguably, they are resources in their
>own right. The 1:1 principle, which was extensively discussed
>at DC-5, asserts the notion that DC metadata should refer to
>a single specific instantiation of a resource.
>"Identifier" should really be some kind of unique code for
>the resource itself. Better examples in the case of a person
>might be SSN, TFN, etc.
>
>However, pursuing this line a little further, while the web-page
>is relatively easy to describe in DC, the e-mail address is
>trickier (I guess, if pushed, I'd suggest it could be
>DC.Type="interactive" since it only has meaning when interacted
>with, but I'm not real confident here). Furthermore, both
>of these resources are clearly _related_ to Stu-the-person,
>and thus it should be possible to provide pointers between the
>metadata for each and Stu-the-person using DC.Relation. I wonder
>what Relation-type should be used? For the web-page either
>"IsBasedOn/IsBasisFor" or "References/IsReferencedBy". However,
>for the e-mail address it is not at all clear.
>The choices are (see Relation working-group report [1] )
>
>IsPartOf/HasPart
>IsVersionOf/HasVersion
>IsFormatOf/HasFormat
>References/IsReferencedBy
>IsBasedOn/IsBasisFor
>Requires/IsRequiredBy
>
>Given the two problems described here for e-mail address
>(which DC.Type? and which DC.Relation-type wrt the person)
>maybe I've argued all the way back around to e-mail address
>actually being a valid "identifier" for a person.
>Doesn't feel quite right.
>
>Comments, anyone. (esp. David, Misha?)
>
>(yr values for DC.Subject are a little wierd too - they are not
>really keywords applicable to Stu-the-person - perhaps the
>contents of his mind, but ...)
>
>[1] http://purl.oclc.org/metadata/dublin_core/wrelationdraft.html
>
>
>[log in to unmask] wrote:
>>
>> DC.Title = "Weibel, Stu"
>> DC.Type = "Physical Object"
>> DC.Identifier = "http://www.oclc.org:5047/~weibel/" SCHEME="URL"
>> DC.Identifier = "[log in to unmask]" SCHEME = "Email Address"
>> DC.Subject = "Dublin Core, Metadata, OCLC, IETF, HTML, URI, IW3C2, ALCTS,
>> ..."
>> DC.Date = "nnnn-nn-nn /" (Stu's Birthday)
>> ...
>--
>__________________________________________________
>Dr Simon Cox - Australian Geodynamics Cooperative Research Centre
>CSIRO Exploration & Mining, PO Box 437, Nedlands, WA 6009 Australia
>T: +61 8 9389 8421 F: +61 8 9389 1906 [log in to unmask]
>http://www.ned.dem.csiro.au/SimonCox/
___________________________________________________________________
Alan Arnold, School of Chemistry and
Director of Flexible Education, University College (UNSW)
Australian Defence Force Academy, CANBERRA ACT 2600 Australia
voice:+61 2 6268 8080 fax:+61 2 6268 8002 web: http://www.ch.adfa.oz.au/apa/
|