[tm-pubsubj-comment] ISSUE 9 - PS Doc scope (domain) of use

[Steve Pepper <pepper@ontopia.net>]

Bernard wrote (in 
http://www.oasis-open.org/committees/tm-pubsubj/docs/recommendations/issues.htm):

>-- Is "domain" metadata the best way to define scope of
>recommended use, and intended audience?
>
>-- Is it the same for all the PS Doc Set, or can it be refined
>at each PS Indicator level?
>
>Proposal:
>
>"dc:domain" seems to be relevant for this declaration of
>recommended use.

What is dc:domain? I can't find it at www.dublincore.org...

What is the requirement?

Do we want to

(1) give humans an indication of the kind of PSIs to expect in
the PSI set?

(2) enable computers to find potentially relevant PSI sets based
on something equivalent to keyword searching?

Or both?

If (1), then surely dc:description is sufficient:

   Definition: "An account of the content of the resource."

   Comment: "Description may include but is not limited to: an
   abstract, table of contents, reference to a graphical
   representation of content or a free-text account of the
   content."

If (2), then surely dc:subject is most appropriate:

   Definition:  The topic of the content of the resource.

   Comment: Typically, a Subject will be expressed as keywords,
   key phrases or classification codes that describe a topic of
   the resource. Recommended best practice is to select a value
   from a controlled vocabulary or formal classification scheme.

Of course, the term "subject" (and the use of "topic" in the
text) invites confusion, but we did decide to use topic map
terminology as our starting point and relate this to
corresponding DC terminology as appropriate. Since topic maps
don't have terminology for this particular thingy, we could
use the term "keywords".

Keywords should be given as published subject identifiers
wherever possible with strings as a deprecated fallback.

Should identifiers for PSIs in the PSI set be included among the
keywords? I think not. What would they be used for, as input to
queries in order to find the PS Doc? But if you already know the
identifier it should resolve straight to the PS Doc...

We should recommend *against* using PSIs that are in the PSI set
as keywords for that PSI set. The keywords should be more
general than the subjects in the set. For example, a PSI set for
countries might have the keyword "geography" (or rather a PSI
for that subject); its keywords would not include "country" or
"France".

>It seems consistent to attach it to the entire PS Doc, and not
>to each individual PSI.

I agree.

>Useful PS Doc should have some "ontological consistency" and
>should not be bags of motley subjects. The domain should be
>defined itself by a PSI.

I'd rather say that the domain covered by the PSI set should be
described in human interpretable form using "description" and in
machine interpretable form using "keywords" consisting preferably
of published subject identifiers.

Perhaps we should describe an algorithm for transforming Library
of Congress Subject Headings into published subject identifiers in
order to provide a starter set of "keywords"?

Steve

--
Steve Pepper, Chief Executive Officer <pepper@ontopia.net>
Convenor, ISO/IEC JTC1/SC34/WG3  Editor, XTM (XML Topic Maps)
Ontopia AS, Waldemar Thranes gt. 98, N-0175 Oslo, Norway.
http://www.ontopia.net/ phone: +47-23233080 GSM: +47-90827246