[topicmapmail] Opera topic map question

[Lars Marius Garshol]

* Alexander Mikhailian
>
> Once I started browsing the actual topic maps (in XTM 1.0, of  
> course), I
> naturally assumed that the instanceOf element was there to convey the
> meaning of a hierarchical relationship and that "instanceOf" is a
> misnomer for "type". This is actually very easy to believe, due to the
> ambiguous use of "instanceOf" in XTM 1.0 that is resolved in the XTM  
> 2.0
> by the introduction of the type element.

Actually, "instanceOf" and "type" are synonyms, as is quite easy to  
see. The two statements

  - lmg's type is person, and
  - lmg is an instance of the person topic type

mean the same thing.

> XTM 2.0  tries to make this confusion less likely by disambiguating
> between different uses of the instanceOf element in XTM 1.0. But it
> leaves completely behind the scope the description of the particular
> status of supertype-subtype and type-instance associations, to the  
> point
> that a reader willing to implement the XTM 2.0 format will have to  
> read
> through TMDM as well.

Any implementor of XTM 2.0 will have to read TMDM anyway. It's  
impossible to tell what the interpretation of the syntax is without  
knowing the data model this syntax represents.

> Which some will definitely not do, rightfully expecting the XTM 2.0  
> format description to be self-contained.

I think it's quite clear from the XTM 2.0 standard that it's not self- 
contained, given that most of the document is given over to describing  
the mapping to TMDM.

> P.S. Is it right to assume that the supertype-subtype association in  
> XTM
> 2.0 is just the rename of the superclass-subclass association in XTM
> 1.0?

Yes, it is. (Strictly speaking, the semantics of the XTM 1.0 PSIs were  
never specified, so one could try to argue that they are not  
necessarily the same, but I wouldn't bother.)

--Lars M.