[topicmapmail] Regarding Facets

[Murray Altheim]

Kal Ahmed wrote:
> I've already replied privately to Nikita on this one, but seeing as
> there was a request to keep this thread public...
> 
> On Thu, 2003-03-13 at 05:40, Nikita Ogievetsky wrote:
> 
>>Kal Ahmed wrote:
>>
>>>I think this is the stumbling block. In XTM, an occurrence is defined
>>>as:
>>>"any information that is specified as being relevant to a given
>>>subject." [http://www.topicmaps.org/xtm/1.0/#desc-occurrence]
>>
>>May be we should take this off-line...
>>Yes, according to the above definition, most people (me included) are using
>>occurrences to describe properties.
> 
> Me too! And I think that we are right in doing so.

I agree, with the proviso that the facet (as a named value) is reified
as a topic and associated - as a facet - to the original topic, not
using occurrences of topics as their own facets.

>>The problem with this is the difference between the name of <occurrence>
>>element and its semantics.
>>I.e. unless we rename <occurrence> into <relatedPieceOfInformation> ...
>>the name of the syntactical element does not correspond to its meaning...
>>(unless I need to brush-up my English of course :-))
>>And that's what I was really trying to say here.
>
> I understand your position, the word 'occurrence' is potentially
> confusing. Of course as one of the founder members of TopicMaps.Org, you
> will know that this name has "historical" significance in that it ties
> directly to the name of the matching concept in ISO 13250.
> 
> I would suggest that the problem lies not in the syntax, but in the
> specification and in the (lack of) tutorials that go with the syntax
> specification. I would also suggest that conversations like this are an
> important part of the education process that will hopefully result in
> fewer people being confused when they see <occurrence> in XTM syntax. So
> I would argue for more energy going toward explaining the syntax we have
> rather than changing it.

I liken this somewhat to the difficulty someone might have in
using RDF or GXL "raw"; there's little or no semantics in any
of these languages apart from their originally-intended ones,
and people have all sorts of domain-based semantics that they
want to add. These are properly the realm of PSI sets, and I
venture to say, use of these PSI sets appropriately as the
semantic basis of topic maps-as-ontologies.

For example, the idea of facets is something that some people
think of as simple name-value pairs, others will want to be
able to type and/or scope. Myself, I definitely want scope on
a lot of my properties. In ontological engineering, there's
typically the graph nodes making up the ontology, then each
node, and sometimes edges too, have their own potential for
properties.

In my authoring ontology I have concepts for character and
place, and then character and place each have properties. These
properties vary sometimes in time or space, so I want to scope
those properties. Because I use PSIs to identify the taxonomical,
mereological, and other type of relations in the base ontology,
I can effectively hide or ignore the property relations (facets)
when I'm displaying or navigating the graph, and search
only within facet relations, say for all male characters over
the age of 50. As you can easily imagine, what is considered
a property and what is considered part of the "base" ontology
has mostly to do with context of use and not structure of the
ontology, hence the happy coincidence that this difference is
not syntactic but only based on the semantics of the association
relations.

Murray

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

     "In Las Vegas Mr Gates also demonstrated a prototype
      fridge magnet which can be programmed to receive traffic
      reports, sports results and advertisements from local
      restaurants using the same FM signal as the wristwatch."
                                  -- The Guardian, 10 Jan 2003.