[sc34wg3] SAM-issue term-scope-def

Jan Algermissen sc34wg3@isotopicmaps.org
Sun, 30 Jun 2002 21:40:08 +0200

Marc de Graauw wrote:
> [Jan Algermissen]
> > No, as I said, I do NOT agree with that. If the extend of validity of a
> > basename characteristic assignment is the scope {Dutch} then (at least I)
> > draw the conclusion that it is NOT valid outside that scope. As you say,
> > 'We are open only on Monday' implies that 'we are closed on Tuesday'.
> >
> > My point was, that 'economie' is NOT a valid basename for the topic
> > if NOT *{Dutch}* applies, but that this says nothing about the validity
> > of the potential validity of 'economie' as a mere name.
> >
> Jan,
> I'll try to get it right this time.

No problem, it seems that Lars was right when he suggested that I should be
more careful in using these terms.

> You say, if I understand you correctly, that in the case described
> 'economie' is not a valid basename when {Dutch} does not apply, but that it
> can be valid as a mere name. That raises the question what a 'mere name' is.

Let's start with what a basename is: According to the present topic map
standards ISO 13250 and XTM 1.0 no topic can have the same basename in
the same scope and thus scopes constitute namespaces in which no two subjects
can have the same name. This provides the ability to identify topics
unambiguously by their basenames. (see Note 31 of ISO 13250)

This is what I referred to when I said that basenames have the semantics of
unambiguity. The association between a topic and its basename provides more
than just a name (or 'mere name' or 'label') for the topic, it provides
an unambiguous name (together with the namespace (scope) of course).

So, my understanding of a scoped basename is, that it serves as an unambiguous
name when the scope applies and that it does not serve as such when the scope
does not apply. In other words, the scope expresses the extend of validity of the
unambiguity of the name, not the extend of validity of the name alone.

Now, what is a 'mere name' ? 
The fact that existing topic map markup languages provide a convenient way
to express topic-basename-characteristics (such as the <baseName> element
in XTM) does not prevent you from defining other association types in order to
express topic-name-characteristics that do NOT have the semantics of unambiguity
(and could therefore be used to avoid topic-naming-contraint-based merges if you
do want to use ambigous names, such as acronyms for example)

You could call your association type 'topic-label' and create associations of
that type between your topics and the strings that are the labels. 

> Not a variant, since 'economie' has in the example not been declared as a
> variant. 

Of course not, <variant> should be used to express different appearances of one
and the same name (such as a GIF image as opposed to a string).

> Do you mean an conformant application might for instance show
> 'economie' as a name to a user when {Dutch} does not apply, as long as that
> application does not treat 'economie' as a basename? So that in a sense
> <baseName> elements have a double use: as real basenames, which are unique
> identifiers, and as name-strings, which can be used in whichever way an
> application chooses? 

No, but it seems like a reasonable idea for starting to resolve the issue if
the topic naming constraint should be applied on basenames or not in the future.

[interestingly you use the term 'real basenames' ;-) ]

> That makes sense, but it would imply that one can
> restrict the extent of validity of the name 'economie' as a basename but not
> as a name-string (mere name). I.e. the scoping of a <baseName> element would
> affect only its use as a basename, not it's use as a name-string. I find
> this couterintuitive and I do not think the standards (XTM, 13250) describe
> it this way.

You are right, they do not define it that way.

> If this is what you mean, what do you think of this problem? If this is not
> what you mean, then what is a 'mere name'?

See above.

> BTW, it is appealing to say what you say (and what SAM suggests now) - when
> a basename is scoped, it is *not* valid outside this scope. It adds a lot of
> simplicity to the model. We might get around the counterintuitive
> consequences by saying: when scoped, a basename is not a valid name *for
> this topic* outside this scope. It might be a valid name *for the subject*,
> that we do not know, and the Topic Map does not say anything about this. To
> get back to our example:
> [economy = "economie" / dutch]
> This says that when {Cherokee}applies but {Dutch} does not, 'economie' is
> not a valid basename for topic economy. It is however possible that
> 'economie' is a valid name for the subject that topic economy represents
> when {Cherokee}applies but {Dutch} does not.

Ah, you DID understand me ;-)

> This seems like an intuitive way to define things and keep simplicity. It
> would say, however, that 'economie' is not valid as a 'mere name' either for
> *topic* economy when {Dutch} does not apply, just that it might be valid as
> a 'mere name' for the *subject*.

When I was thinking about an example for the topic-label association above,
I realized that after all, your initial problem isn't solved:
Suppose I say:

'"tennis" is a label for topic T1 in the scope {English}'

does this imply that 'tennis' is not a valid label for T1 in German ???
The 'extend of validity' of labaling T1 with 'tennis' is expressed to
be the scope {English}. Since we can consider {German} to be beyond that
extend it should be logical to say that 'tennis' is not a valid label
for T1 in German....

That's rediculous, isn't it ?

What does this mean for

- the use of languages as scopes ?
- the original issue of the interpretation of scope ?

> This brings to light some interesting differences between our standards:
> ISO13250: topic name = A string of characters specified as a name of a
> *topic*
> SAM: A base name is a name or label for a *subject*
> XTM: A *topic* may have zero or more names

Can you explain what you mean, I do not understand.


> [Lars Marius Garshol]
> > How about phrasing the definition of scope as shown below?
> >
> >   All topic characteristic assignments have a <term>scope</term>,
> >   which defines the extent to which the statement represented by the
> >   assignment is valid. Outside the context represented by the scope
> >   the statement is not known to be valid. Formally, a scope is
> >   composed of a set of subjects that together define the context. That
> >   is, the topic characteristic is known to be valid only in contexts
> >                                ^^^^^^^^^^^^^^
> >   where <emph>all</emph> the subjects in the scope apply.
> This means the existing definition would remain as it is (say *not valid*
> instead of *not known to be valid*), though we could add a sentence
> explaining these issues, i.e.: "This restriction does not say anything about
> the relation between the subject and the characteristic in contexts where
> where not all the subjects in the scope apply. It only restricts the
> relation between the topic and the characteristic."
> And we could say:
> A base name is a name or label for a topic, and, indirectly, for the subject
> the topic represents.
> What do you think?
> Marc
> _______________________________________________
> sc34wg3 mailing list
> sc34wg3@isotopicmaps.org
> http://www.isotopicmaps.org/mailman/listinfo/sc34wg3

Jan Algermissen
Consultant & Programmer

Tel:   ++49 (0)40 89 700 511
       ++49 (0)177 283 1440
Fax:   ++49 (0)40 89 700 841 
Email: algermissen@acm.org
Web:   http://www.topicmapping.com