[sc34wg3] SAM-issue term-scope-def

[Nikita Ogievetsky]

Hello Bernard,

> > I am with Lars and Kal on this.
>
> Well ... I thought Kal was with Graham and me. So we are all together?

Of course we are all together, it is just a fluctuation :-)

> > In around 70's in USSR people got the notion
> > of a new plastic and aluminum millennium and threw away antic and
> > highly valued bountiful things of bronze, cupper, silver and even gold.
> > Nobody could ever understand russian soul but lets not run after
plastic.
> > (I hope you know what I mean)
>
> Barely. Has it something to do with "Russian Semiotics" ? ;-)

Not at all :-)
I meant that we often tend to underestimate our present values
when we run after new shining ideas.

> > Bernard, I do not see your problem:
>
> I'm sad you don't. It's not a problem in fact, just an example.
>
> > "Henry is King of Navarre and King of France from 1589 to 1610"
> >
> > <association>
> >     <instanceOf><topicRef xlink:href="#king-of-a-country"/></instanceOf>
> >     <member>
> >         <roleSpec><topicRef xlink:href="#role-king"/></roleSpec>
> >         <topicRef xlink:href="#henry"/>
> >     </member>
> >     <member>
> >         <roleSpec><topicRef xlink:href="#role-country"/></roleSpec>
> >         <topicRef xlink:href="#navarre"/>
> >     </member>
> > </association>
> > <association>
> >     <instanceOf><topicRef xlink:href="#king-of-a-country"/></instanceOf>
> >     <scope><topicRef xlink:href="#from-1589-to-1610"/></scope>
> >     <member>
> >         <roleSpec><topicRef xlink:href="#role-king"/></roleSpec>
> >         <topicRef xlink:href="#henry"/>
> >     </member>
> >     <member>
> >         <roleSpec><topicRef xlink:href="#role-country"/></roleSpec>
> >         <topicRef xlink:href="#france"/>
> >     </member>
> > </association>
>
> Of course you can express it that way, and many others. I would have done
something like
> that too ...
> I expressed it with scopes just for the illustration. Maybe I could have
found something
> less far-fetched.

Of course, let me help you:

"Nikita said that Bernard agreed with Graham that scopes are obsolete."

> > I was proposing structured scopes two years ago (as I think many others
did)
> > But that proposal was put on a waiting list.
> > Ands I must admit that I had waited quite successfully.
>
> What do you mean exactly by that?

I mean that as implementer I had to use what I had and found that it was
doable.

> I don't think structure of scope is the issue. The issue
> is : what is the nature of the relation between the scoping topic and the
scoped
> association? And IMO this question boils down to specify a role, no more,
no less ...

Agree. This is actually exactly what I am doing in RTM and new QTM proposals
See, for example, slides 29 and 31 in
http://cogx.com/xtm2rdf/extreme2001

Where each association
in RDF representation has a "validIn" property (scope)
whose object is a s-node (using TMPM4 terms).
S-node in XTM is just a set. While in RTM it can be a richer construct.

Some QTM examples are on slides 11 and 24 in
http://cogx.com/kt2002

> > Lets use scopes for what they are good for (like in Lars's examples for
> > example)
> > for everything else there are association members.
>
> > For distinguishing roles of scope themes there is the topic  typing
> > mechanism.
> > (type of topic that plays the role of scope theme).
> > Primitive but works.
>
> I don't like it that much. Confusing a type (class) with a role is
suboptimal. Again, what
> has to be specified is the *role* of each scoping topic in its relation
with the
> association. And we have a mechanism to do that very accurately and quite
simply through
> association reification. Why not use it?

Why not indeed? You mean create an association describing a context and
then use its subject to define a scoping theme?

I misunderstood that you are suggesting to make the very notion of scope
obsolete.
I think that separation of contextual axes from characteristics axes is
very-very important.


--Nikita.