[sc34wg3] SAM-issue psi-generics (was: SAM-issue term-scope-def)

[Lars Marius Garshol]

* Nikita Ogievetsky
| 
| What I am saying is that the fact that an author decided to make a
| topic proxy for a subject can be lost if "marc : topic" is replaced
| with simply "marc".

I don't understand what you mean at all. XTM 1.0 says that if you
don't have an explicit class, you are by default an instance of the
PSI #topic.

So the two following are therefore equivalent:

  [marc]
  [marc : topic]

That's what issue "psi-generics" is about.

| So topic here is similar to DAML "thing".

Does DAML have that term?
 
| So, I am a subject? Hmmm...
| I think that a subject is a mental proxy for an individual.
| And a topic is a computer proxy for a subject.
| So I am neither a subject nor a topic.
| However my subject is an instance of a subject
| and my topic is an instance of a topic.

I think this is a philosophical point of view that might have some
validity, but it doesn't fit very well with how topic maps have been
defined up to this point. So far topic maps have said topics represent
subjects, and subjects are real-world things. I don't think there's
much to be gained from quarreling with that.

We have a standard that we are currently rewriting. What we need to do
is clear up the problems with the standard we have, and that requires
a certain focus. It would be good if this thread could focus a bit
more soon.
 
| To illustrate that class topic make sense,
| imagine a QA API for XTM. When you QA engine encounters
| a topic element it understands that you are saying:
| 
| - I want to talk about a topic.
| [...]

Dear Nikita, this was no help at all. We have to be very formal and
precise here in order to ensure that we are talking about the same
things. That's why I've used LTM syntax everywhere, so that it's
absolutely clear what I'm saying.

Here is my problem, formulated in terms from the current
specifications. I think we have to focus on that in order to be able
to make any progress on this at all.

XTM 1.0 seems to say that if we read in

  <topic id="marc">
    <baseName>
      <baseNameString>Marc de Graauw</baseNameString>
    </baseName>
  </topic>

we have to produce a SAM instance where this topic has a
class-instance association with the #topic PSI. Given that all topic
maps contain several topics with no class which represent subjects
that are *not* topics, this can't be right. That's my concern, and
that's what this issue is about.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >