Re: comment (non-member) on N-ary relations

Natasha,

Thanks for your helpful responses !

One follow-up on the Pattern 1 discussion point (about not inadvertantly
suggesting probability concepts are tractable in standard OWL/RDF).

You make a great point:

> I am
> afraid that the idea of using speed won't much less controversial
> than probability in some quarters :) One may interpret this as some
> suggestion on representing temporal information, and of course we
> don't want to do this in this note. I hope that the disclaimer that
> we are *not* trying to make any suggestions on how to represent
> probabilistic information, would be an acceptable compromise.

I suppose this highlights a more general point: Pattern 1 lends itself to
asserting all kinds of "How" semantics. But many kinds of "How" semantics
are formally representable only using expressive modal logics, and OWL is
not equivalent to an expressive modal logic.

So maybe a footnote should just make the broad point that while Pattern 1
invites representing all kinds of "How" assertions, users should bear in
mind that for many of these--like those involving time, probablility,
belief, obligation, etc.-to actually reason validly by machine is a
fundamentally complex task, that in many cases is inherently intractable
using current formalisms, including OWL DL.

John

Received on Saturday, 16 July 2005 19:34:50 UTC