[topicmapmail] Superclass-subclass indentation in the Omnigator

[Murray Altheim]

Lars Marius Garshol wrote:

> * Murray Altheim
> | 
> | In the same way as RDF provides a very basic graph syntax, that when
> | coupled with RDFS one can then begin to describe the semantics of
> | the syntactic constructions, [...]
> 
> I wonder why people keep saying stuff like this. Where are the
> semantics in RDFS?


Either I'm not clear or you're misinterpreting me. I've myself
had a difficulty with people comparing RDF and XTM, since one
is very simply graph syntax, the other a very specific syntax
for topic maps (about three levels up in the chain of being).
My point was that you needed RDFS as a schema language to be
able to create a markup language capable of doing what XTM does,
and that you *could* do anything you like with RDF/RDFS, such as
XUL, DAML, etc. As we both know, RDFS is just a schema language,
like DTDs or XML Schema.


> | [...] topic maps start with a basic core and allow anything to be
> | constructed, comparable more to RDFS than RDF.
> 
> XTM 1.0 specifies a syntax and an (implicit) model, in just the same
> way that the RDF M&S specification does, and at a comparable level of
> generality. RDFS corresponds much more closely to TMCL than it does to
> topic maps themselves. Ditto for DAML+OIL and OWL.


Yes, I understand this. I think most people on this list do. Maybe
I'm wrong.


> | I see no reason why topic maps couldn't be used to express anything
> | that can be expressed in another language. 
> 
> Agreed. The expression may not always be convenient, but it's
> difficult to think of something that couldn't (given the right set of
> published subjects) be expressed.


One of the things I've tried to get across to people is that
there *is* an interesting comparison between RDF-as-simple-graphs
and XTM, in that whereas RDF can't without RDFS express XTM, XTM
can by virtue of its <topic> and <association> elements provide
a fairly simple graph syntax, and that with the addition of some
PSIs could be used to create an interesting knowledge modeling
language. As I've said before, I'm eager to see Common Logic,
as I think it could easily be either expressed or transformed
in/into XTM. Then adding XTM's mapping capability and things
begin to get quite interesting. I'm not sure CL will have any
built-in mapping ability or even hooks into the outside world
(CGIF didn't, for example).

Murray

......................................................................
Murray Altheim                  <http://kmi.open.ac.uk/people/murray/>
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK

            If you're the first person in a new territory,
            you're likely to get shot at.
                                                     -- ma