Re: A plea for peace. was: RE: DAML+OIL (March 2001) released: a correction

Aaron Swartz wrote:
> 
> Dan Connolly <connolly@w3.org> wrote:
> 
> > But in practice, I don't expect to use reification
> > to actually convert between logical formulas
> > and RDF 1.0 syntax.
> 
> So, to be clear, you are suggesting what?

I didn't mean to suggest anything in particular.
I'm suggesting lots of experimentation, I guess.

> I would guess you want an RDF 1.5 or RDF 2.0 with built-in primitives for
> logic?

Well, yes, I expect a standardized representation of logical
formulas to be widely deployed.


> And this would be done after the work of the current RDF Core group,
> no?

I dunno when it will/would be finished, but it's already
started; a lot of relevant work is going on right now.

Take a look at Drew McDermott's idea, for example.

  Version 0.1 
  http://www.cs.yale.edu/homes/dvm/daml/proposal.html
  Last Modified:  Wednesday, 13-Dec-00 23:40:15 GMT

  <-
  http://lists.w3.org/Archives/Public/www-rdf-logic/2000Dec/0072.html


> And what are your thoughts on the current work doing logic in N3 -- it seems
> that this uses reification (the {} braces) to do logic and works just fine.

The N3 tools can sort of theoretically reduce {} formulas to
RDF 1.0 syntax, but practically, they don't exploit
the connection. The N3 tools don't support recognition
of logical formulas in reified form.
In the general case, that's a theorem-proving excercise,
since we could have subproperties of rdf:subject and so on.

N3 {} formulas are their own little world right now. I'll
be more happy to say that they work "just fine" after
we've shown interoperability with some other tools:
RuleML tools, or forethought's RIL tools, or some
KIF tools, or XSB or something. And after I convince
myself that the use/mention tricks you can do with N3 {}'s
make sense.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Thursday, 5 April 2001 13:23:33 UTC