RE: QNames, semantics, serialization

>Drew,
>
>Drew McDermott wrote:
> > I believe I am echoing Pat Hayes when I say that I don't understand
> > why it is considered desirable, let alone essential, let alone
> > *possible*, that there be only one name for each object.  What on
> > earth could guarantee such a thing?  Lasers in space that destroy any
> > computer found to have a nonofficial URI for an object?
> >
>Although I *really really* wouldn't want to contribute to this waaay too
>long discussion, I have to say that the issue is that one would like
>there only to be one *object* for any *name*, not the other way around
>(which you seem to be talking about - and I agree with your comments
>there).

In a given interpretation, yes.  But now we have to face the awkward 
but inescapable fact that there is no way to guarantee that the 
interpretation you have in mind when you write your RDF is the same 
one I have in mind when I read it. The best we can ever do is to 
assume that I manage to get enough information from what you say (and 
what I already know) that I can draw the conclusions that you want me 
to be able to draw about the things that you hope I am thinking 
about. Really pinning down actual *reference* to *real things* is a 
very, very  tricky business, one that is way beyond the ability of 
current semantic theories to analyse, and therefore probably best 
left aside for now in these discussions.

>That said, I think the possibility outlined earlier that two QNames
>accidentally map into one URI is so remote that I would suggest we let
>go of this discussion and move onto more productive things :-)

Ive given up on *that* discussion long ago :-)

Pat

---------------------------------------------------------------------
(650)859 6569 w
(650)494 3973 h (until September)
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes

Received on Thursday, 30 August 2001 19:13:12 UTC