[topicmapmail] Opera topic map question

[Lars Heuer]

Hi Alexander,

[...]
> May I slightly change your example so that my question becomes more
> clear? I will add three new topics: "be" "abstract" and "concrete" and
> use them in the association as type and roles.

Thanks. Please let me transfer your example to CTM [1], I think it is
more readable than XTM. The following CTM fragment produces exactly the
same topic map:

       be(abstract: person, concrete: alexander)

> Now, many questions may be asked:

> 1. Can this be done at all?

Yes.

> 2. What can be said about the link from <association/> to #be?

The TMDM says, it is the type of the association which describes the
nature of the relationship.
The TMRM goes a bit further and mandates that the above statement
implies the following statements:

        tmdm:association ako tmdm:statement .

        be ako tmdm:association .

In conclusion, the TMRM says, that every association is a subtype of
tmdm:statement and that every association-type is a subtype of
tmdm:association.

I think, that's the answer of your questions (and answers question 3
and 4, too)?

> Now, suppose I modify the above fragment to add the topic "relation" and
> to make a supertype-subtype association between #be and #relation.

CTM:
    be ako relation


> 5. What can be said about the transitivity of the link 
> #relation 
>   -> #be 
>     -> the association of type #be?

You can infer that be is an instance of "relation" and if you take the
TMRM into account, "be" is an instance of "relation" and
tmdm:association.

Best regards,
Lars
-- 
http://www.semagia.com
http://www.topicmaps.de/mailinglist/ German Topic Maps mailinglist