Re: Graham.Klyne@Baltimore.com, patrick.stickler@nokia.com, jjc@hplb.hpl.hp.com, jos.deroo.jd@belgium.agfa.com

>[what a subject to discuss...]
>
>Pat,
>
>[about http://www.coginst.uwf.edu/users/phayes/w3-rdf-mt-draft-J.html]
>
>If XL is really a *global* mapping

What Im trying to convey is that XL is considered to be fixed outside 
the particular interpretation. It may be clearer to just omit that 
word 'global'.

Fixed.

>  I assume that
>each literal node is mapped into *one-and-ony-one* literal value

Right, which is why we have to treat different occurrences as 
distinct nodes (or else rule out the P-ish class of schemes by fiat).

>or have I misunderstood global?
>then I don't see why
>.. A literal node is a particular occurrence of a literal
>.. literal nodes are unique to each graph.
>.. several occurrences of a literal produce several rectangles with the same
>label
>.. literal nodes are not 'merged' in this way
>
>I think that the most convenient global mapping is
>   XL(literal-as-simple-name) = literal-as-simple-name
>i.e. the identity mapping

Well, yes, it is surely the most convenient from an MT point of view, 
but that would say that every literal denotes a string. I originally 
assumed that XL was defined on literals themselves (not occurrences 
of them), but I got mauled for doing that by Peter Patel-Schneider, 
if you recall, precisely because he wanted to have a 
datatype-sensitive mapping from literal occurrences to meanings.  So 
I'm leaving the issue as open as I possibly can. XL might give every 
literal the same value, or it might give every single literal token 
in the universe a different value, choosing it by divine inspiration, 
or anything in between. This version of the MT just washes its hands 
of the problem and leaves it to some other box to decide. Obviously 
this is not a final account (as the document says in about ten 
places), but it is about the weakest assumption that I can possibly 
make for now.  Most of the entailment stuff works even with this 
minimal an assumption, which is encouraging (since that means it will 
go on working no matter what we decide about datatyping.)

Pat
-- 
---------------------------------------------------------------------
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 Wednesday, 30 January 2002 12:48:44 UTC