Re: LANG: Is a non-RDF triples syntax out of charter?

Dan Connolly wrote:
> On Thu, 2002-03-28 at 15:46, Jeff Heflin wrote:
> > In the telecon today, when I mentioned that I was in favor of an
> > ontology language that was not constructed using RDF triples, Dan
> > Connolly suggested that this was outside of the charter. I have since
> > gone back and reread the charter, in order to determine if my suggestion
> > is clearly out of it. I was only able to find two relevant sentences to
> > this issue, which I discuss below.
> >
> > The charter says the group must design "a Web ontology language, that
> > builds on current Web languages that allow the specification of classes
> > and subclasses, properties and subproperties (such as RDFS)." This is
> > probably what Dan is referring to. I may be getting into semantics here,
> > but I understand this as "RDFS may be one such language that we should
> > build on," not that we have to build on RDF. That is it says, "such as
> > RDF," not "including RDF." So, I don't find this a convincing reason for
> > rejecting my point of view.
> There's a whole community of folks expecting WebOnt's work
> product to work with RDF the way DAML+OIL works with RDF:
> i.e. it just adds more well-known terms for use in RDF.
> Perhaps I overstated the point when I referred to the charter...
> but if we decide that WebOnt's product doesn't
> work with RDF the way DAML+OIL does, then we're going
> to have to renegotiate with that community somehow.
> I think a change to the charter would be in order,
> but perhaps I could be persuaded that some other
> mechanism of renegotiation would suffice.

Hopefully, this group represents a large percentage of the people with
immediate interest in WebOnt. I think it would be possible for the group
to get a pretty good idea of the community's attitude towards anything
we might consider. Still, I haven't heard a whole lot of support in one
direction or the other yet. I imagine most people are waiting til they
see some concrete syntaxes and associated semantics before they form an
opinion. All I ask is that we leave the option to use a non-RDF triples
syntax open. If we have serious problems making a RDF triples syntax
work, then we can consider my option, determine how to gauge the
community's opinion, and assuming it is favorable, decide whether or not
a charter change is order.

> > The charter also say the language "will be designed for maximum
> > compatibility with XML and RDF language conventions." The phrase
> > "maximum compatibility" appears to give us some wiggle room. If we feel
> > that a certain degree of compatibility is impossible without undermining
> > the goals of our language, then maximum compatibility might be slightly
> > below that point. Also note that RDF Schema is not mentioned in that
> > sentence, and my proposal is that we still use RDF for representing
> > data, we just shouldn't use triples to represent logical definitions.
> >
> > Thus, I would say that it is not clear that we are chartered to extend
> > RDF Schema. Furthermore, there is certainly nothing in the charter that
> > says the ontology language's syntax must be formed from RDF triples.
> >
> > Jeff
> --
> Dan Connolly, W3C

Received on Friday, 29 March 2002 10:35:35 UTC