Re: universal languages

>
>Regarding the role of RDF in all this, there's not much more to RDF than
>the modest claim that Web data standards should use Web identifiers
>(URIs) to name pretty much everything. I'm certainly not married to the
>idea that everything ever needs to be squeezed through the RDF 3-tuples
>hourglass. RDF is plenty useful but there's more to all this than nodes
>and arcs. I do think URIs are pretty essential though...

Clearly it is essential that URI's *can* be used by the web data 
standard. However, one can use URI's as identifiers in KIF (well, not 
ANSI-KIF which is case-insensitive, but we are fixing that). What I 
don't see the reasoning behind is the insistence that URI's should be 
used to name *everything*. In fact that seems brain-damaged (its a 
bit like insisting that HTML consist of nothing but URL's, with no 
actual text allowed), since one is almost certainly going to want to 
refer to things that don't have a URI (such as my Uncle Jim's left 
thumb, or the number 6, or a step in a proof which someone claims to 
exist but can't actually refer to) and one is definitely going to 
want to refer to classes not every member of which has a URI (such as 
the class of people who don't have a URI, to take an example at 
random). Since URI's can be readily identified from their syntax, 
allowing non-URI identifiers wouldnt get in the way of anything, so 
why is this restriction 'pretty essential'? What depends on having 
it, or would break if it wasnt there? URI's don't (can't possibly) 
provide a unique-names assumption (since the same thing can be 
mentioned at many places on the web), so their global uniqueness 
doesnt provide any semantic leverage.

Pat Hayes

---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes

Received on Monday, 5 February 2001 10:54:55 UTC