Re: request for feedback on domain and range semantics.

On Fri, 2002-10-25 at 12:08, pat hayes wrote:
> 
> (BCCd to three WGs to avoid cross-postings in replies)
> 
> I am hereby seeking feedback in order to help me make a decision. As 
> editor of the RDF MT document, I have discretion to decide whether 
> rdfs:range and rdfs:domain should have 'IF' or "IFF" semantics.  What 
> turns on this, in case anyone hasn't been following, is whether 
> ('iff') or not ('if') the following inference should be considered 
> valid:
> 
> P rdfs:range A .
> A rdfs:subClassOf B .
> -->
> P rdfs:range B .
> 
> I have argued for the utility of disallowing this entailment, and 
> several people have agreed; but there also seems to be a widespread 
> feeling that the entailment is intuitively 'reasonable'. Moreover, 
> several people have noted a preference for having a uniform rule one 
> way or the other, and I think it is essential that we give subClassOf 
> and subPropertyOf an 'iff' semantics. On the other hand, the 'if' 
> alternative makes life a little easier for inference engines.
> 
> So far, all the arguments I have heard, including my own, are 
> basically aesthetic.

I wonder if the bit about inference engines is really aesthetic...
I can't quite put my finger on it, but it seems to me that
the IF semantics are expressible as horn clauses while
the IFF semantics are not.

As long as the IFF semantics can be captured as so called
"closure rules" (which are horn clauses, right?), then
yes, it's a coin-toss... I kinda lean toward IF on
the grounds that it's less constraining (and in some
sense more expressive...) but I asked TimBL about
this and he expects IFF semantics.

> My request is, does anyone have a "can't live 
> with" technical objection to either alternative? If so please send me 
> an email in the next few days.  Thanks.
> 
> Pat Hayes
> -- 
> ---------------------------------------------------------------------
> IHMC					(850)434 8903   home
> 40 South Alcaniz St.			(850)202 4416   office
> Pensacola               			(850)202 4440   fax
> FL 32501            				(850)291 0667    cell
> phayes@ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Friday, 25 October 2002 13:15:19 UTC