[topicmapmail] Are Facets Really Simple After All?

[Murray Altheim]

Kal Ahmed wrote:
> On Sun, 2003-11-30 at 18:32, Murray Altheim wrote:
> <snip>
> 
>>   2. Using a Topic occurrence. I'm leaning towards advocating that
>>      RDF-style property assignment be handled via Topic occurrences,
>>      whenever the application doesn't demand or suggest they be
>>      fully reified as Topics. There's some screwy semantics to doing
>>      this that I'm not entirely comfortable with, as Topic occurrences
>>      are occurrences of their Topics, not properties of their Topics.
> 
> Occurrences are "information that is specified as being relevant to a
> given subject" (ISO 13250:2002, 3.28 - the same phrase is repeated in
> XTM 1.0, section 2.2.3). I think it is a mistake to read too much into
> the name "occurrences".
Kal,

If I took that vague statement as the sum total of what we knew
about Topic occurrences, I'd have given up long ago. If "information
that is specified as being relevant to a given subject" were to only
mean "relevant" as in "somehow connected" and not more explicitly
"representative of" or "indicative of", then occurrences would be
semantically meaningless or at best extremely vague, and essentially
every Topic Map ontology would have "is relevant to" as its basic
association type. In an ontology of medicine, military strategy,
eBusiness, etc., basically *everything* "is relevant to" the overall
subject.

My take on occurrences is that they are simply what they are called,
occurrences of a Topic. There's enough vagueness in that statement
alone to fly an airplane through -- expanding that to be anything
relevant to a specific Topic means that every Association between
two or more Topics is basically establishing relevancy, and the
entire Topic Map would collapse. E.g., every one of the 700 Topics
in my authoring ontology is somehow relevant to authoring, so I'd
end up with one Topic with 700 occurrences. "is relevant to" is
pretty useless. It's essentially a synonym for the most basic
relationship: "is related to".

If we go back to the original idea of Topic Maps, as back-of-book
indices, if one were to query the index for "description logics" in
an index, nobody would expect to have returned say, the created date
of the index entry itself. This is what you're suggesting is a valid
use. I can't buy that could possibly be the intention of the standard,
only that the text is vague. And we always interpret standards. The
fact that both of us have been part of its development should inform
us about a perhaps more correct interpretation than conflating the
metadata about a Topic with its occurrences.

 > I have always considered it perfectly valid
 > (indeed the only way) to encode properties of topics where the property
 > value is not something that you want to talk about any further within
 > the topic map.

That there is an inability to encode properties of Topics doesn't
suggest that a form of markup abuse is permitted. The reason I
started the thread on facets was to look into some sort of best
practices for doing this kind of thing. Unless all of the vendors
and developers were to agree upon some sort of PSI identifier (i.e.,
something that really should be in the standard itself), we'd find
applications displaying or using Topic metadata as if that were
Topic occurrences.

Murray

......................................................................
Murray Altheim                    http://kmi.open.ac.uk/people/murray/
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK               .

   "Mr Bush, you have to admit it's a pretty remarkable thing for a
    man just to go to a hotel room door and open it and have a woman
    standing there and have sex with her," said Marshall Davis Brown,
    lawyer for Sharon Bush. "It was very unusual," Bush said.
    http://www.nzherald.co.nz/storydisplay.cfm?storyID=3536318