Re: range, domain: Conjunctive AND disjunctive semantics both supportable

At 05:37 PM 10/2/00 +0100, Ian Horrocks wrote:
>You are missing something ...
>
>1) Jan's approach doesn't express the fact that the range restriction
>is EXACTLY (A or B) but that it is some class that subsumes (A or
>B).

I take your point.

I'm currently digesting your OIL-in-RDFS paper ("Adding formal semantics to 
the Web" -- sorry, have lost URL), which looks very interesting.

Unless I'm wrong again (quite likely), I note that conjunctive semantics 
for domain would mess up your proposed definition for <oil:hasOperand>, but 
I guess that's easy to fix.

#g

PS:  I think the paper has a typo in Figure 4:  I think the second 
"class-def plant" should be "class-def branch".


------------
Graham Klyne
(GK@ACM.ORG)