[sc34wg3] SAM-issue term-scope-def

Bernard Vatant sc34wg3@isotopicmaps.org
Thu, 4 Jul 2002 11:48:46 +0200


>
> * Bernard Vatant
> |
> | So - sorry to repeat myself ...You can't set generic rules for
> | scopes, even for scopes on names, because scope assignment can carry
> | opposite semantics:
> |
> | either "This is one condition in which the assertion is valid" (and
> | it might be valid in some others)

*Lars Marius Garshol
> This is what the standard has to say, as I've pointed out repeatedly.
> This case is just an example of missing information. That is, the fact
> that the topic map fails to say that "economie" is a valid name for
> the subject in Samoyed does not mean that it isn't valid.

Exactly. So please reread carefully what you said in the message I replied to just before.

> | or "This is the only condition in which the assertion has to be
> | considered valid"
>
> It is acceptable for applications to assume this, but not for the
> standard to do so.

I disagree. If the standard says that scope is sufficient but not necessary, an
application that assume that it is both sufficient and necessary is not conformant. I
think we have to be crystal clear on that.

> | The more I think about it, the less I think we can make general
> | sense of scope in a way that could be supported by any kind of
> | formal model, because we can't restrict it to be one of the two
> | terms of the above alternative, and support both under a single
> | concept is contradictory.
>
> I don't see that as so very problematic. If you do, it would be good
> if you could explain why.
>
> | Maybe we need to introduce the notions of "necessary scope" and
> | "sufficient scope" after all?
>
> What notions are they? I would be happy if you could introduce them to
> me, at least. :)

<sigh> You are unusually exasperating today ;-) <sigh/>
I had introduced them already I thought, but maybe not explicitly enough ...

-- "necessary scope" means:
"If A is valid, then scope applies".

It's better understandable and usable in its contrapositive equivalent expression:
"If scope does not apply, then A is not valid"

Necessary scope will support filtering mechanism, e.g. if scope does not belong to users
profile, then they do not see the characteristic (the characteristic is not seen as valid
by/for those users)

-- "sufficient scope" means:
"If scope applies, then A is valid".

If I catch you well, SAM would recommend:

1. Scope shall be interpreted as sufficient in any case.
2. It should be useful in certain cases to interpret also scope as necessary scope.

But of course, it would be far better if the TM itself could declare if the used scope has
to be interpreted as necessary, sufficient, or both, so that every application uses it the
same way.

Hope that helps ...

Bernard