Re: WOWG: first language proposal

The problem with using only triples means that you have to resort to
using reification or "dark triples" to correctly convey the meaning of
the syntax. Even Berners-Lee seems to agree that an RDF triples syntax
for expressing logical statements is problematic. See
http://www.w3.org/DesignIssues/Toolbox.html, where he says it could be
done in RDF using reification and a special "quote" construction, but
then in all of his other examples, uses an plain XML (non-RDF) shorthand
syntax (you can tell because it doesn't obey the RDF striping model).
Note, even if we went the route suggested by him, you still need to add
the "quote" element.

Of course, the logical constructs mentioned by Berners-Lee are more like
rules than WebOnt, so you may wonder if the issues still apply. So
instead, look at WebOnt's closest relative, DAML+OIL. Do you think the
syntax is that ugly because the Joint Committee wanted it to be? No, it
was the restrictions of RDF Schema and using triples to define the
language that forced us into that syntax. If you don't think DAML+OIL
syntax is ugly, then try explaining to someone why we need the odd
parseType="daml:Collection" in order to say that a class is the union of
two other classes.

Jeff

Jos De_Roo wrote:
> 
> Jeff Heflin wrote:
> >>>Finally, an important issue will be finding a way to map your abstract
> >>>syntax into XML/RDF and still preserve its simplicity. I believe that in
> >>>order to get a good, intuitive syntax, we'll have to seriously consider
> >>>dropping the idea of using triples to represent the language, i.e., do
> >>>not layer on top of RDF Schema (but this is a point I've already raised
> >>>in another thread).
> 
> Jos De_Roo wrote:
> >> I haven't seen anything in the past 3 years that would motivate
> >> such an idea, really, http://www.w3.org/DesignIssues/Overview.html
> 
> Frank van Harmelen wrote:
> >Jos,
> >Can you clarify what you mean by "such an idea"?
> >
> >Do you not see to the need "to map the abstract syntax into XML/RDF and
> still
> >preserve its simplicity" or to " not layer on top of RDF Schema" ?
> 
> oops, sorry for the confusion Frank, but I meant the latter one
> i.e. the idea to drop the use of triples to represent the language
> (also in the perspective of the logic/proof layers)
> all cases I've seen so far are examples of circular models
> such as _:x :p _:y . _:y :q _:z . _:z :r _:x but I don't
> see any problem with that
> 
> --
> Jos De Roo

Received on Thursday, 28 March 2002 10:47:55 UTC