RE: SUO: RE: Topic Maps
Dear Pat,
> >Dear John,
> >
> >You completely miss the point. This is not an issue of KIF
> or Topic Maps.
> >Topic Maps only provides a structuring and linking
> capability - precisely
> >what KIF doesn't have.
>
> What "structuring capability" is provided in ISO/IEC 13250 ? See for
> example http://www.ornl.gov/sgml/sc34/document/8/draft27.pdf
MW: Topic Maps are really one step above an html hyperlink, in that they
allow you to say something about the nature of the link, whereas a hyperlink
only says "follow me if you dare". More particularly, Topic Maps allow
associations (relation instances) to be created between "resources", which
can be pretty much what you want, but I would have in mind that the
resources were specific theories in KIF and that Topic Maps could be used to
support the
Theory X
Uses Y, Z
...
End Theory
stuff that you and the other KIF folk seem unwilling to put into KIF itself.
>
> >There are two points:
> >
> >1. Topic Maps can help us to organise our KIF theories -
> manage the network
> >for us. All we get with KIF is something monolithic.
>
> I fail to see quite how topic maps are going to do this managing, or
> even help in doing it.
MW: See above.
> They seem to be a purely hyper-textual device,
> for organizing textual resources. Being based on a primitive notion
> of 'association link' (it is a pity that the ISO committee didn't
> have a psychologist on board, who could have explained the 30-odd
> years of frustration into which cognitive psychology was led by
> relying on association networks) they aren't going to be able to
> express much useful logical structure. So even with the ISO
> impimateur, I doubt if Topic Maps are likely to be very much real use.
MW: I can agree with you that Topic Maps are not all that could be wished
for. It is only version 1 though.
>
> I think a much more productive direction to go in would be to
> reconcile KIF with the emerging web-based semantic markup languages,
> notably DAML+OIL, which provide both a way to use URIs as an
> organizing tool to connect ontologies on the web and could be fully
> integrated into the logic itself. There already is an axiomatic
> semantics for DAML+OIL written in KIF, for example.
MW: This might be the case if we wanted to express the ontology directly in
an XML format of some sort, but that is not what I am suggesting. Rather I
am suggesting providing an organising structure with Topic Maps.
>
> >2. If we are going to get the SUO used, we need to make it easily
> >accessible. We could aid this with the use of Topic Maps.
>
> What is the user community for Topic Maps?
MW: The target user community for Topic Maps is anyone who uses a search
engine (at least indirectly).
> KIF is already easily
> accessible: there are many widely available software tools for
> reasoning with it, translating into and out of it, and parsing it.
> Several of these have been in wide use for about 10 years now, and it
> can be written in XML if anyone feels they want to do that.
MW: Again you are missing the point. This is not about representing the
ontology in Topic Maps, but organising the parts of the ontology, and making
it accessible through Topic Maps.
There is also an issue of making it intelligible to the masses, but I do not
see that Topic Maps can help with that.
>
> >What is being suggested has nothing to do with how the
> ontology is defined.
> >There is an issue that KIF is unintelligible to all but a
> few cognoscenti,
>
> Well, I do feel obligated to point out that KIF has been treated as a
> kind of standard by almost everyone in the entire ontology research
> community for about a decade, and is in any case only conventional
> first-order logic written in (an admittedly somewhat messy) LISP-like
> syntax, so this 'few cognoscenti' is rather a large percentage of the
> world's active workers in this area.
MW: Precisely. And if that's not clear enough: how many hundreds of millions
is that?
Regards
Matthew
============================================
Matthew West
Operations & Asset Management
Shell Services International
H3229, Shell Centre, London, SE1 7NA, UK.
Tel: +44 207 934 4490 Fax: 7929
Mobile: +44 7796 336538
E-mail: Matthew.R.West@is.shell.com
http://www.shellservices.com/
============================================