RE: How do you deprecate URIs? Re: OWL-DL and linked data

> From: Dan Brickley
> [ . . . ]
> Call me old fashioned, but owl:sameAs means what it means as
> defined in the OWL specs. If people are going to miss-use it,
> I'd rather they did
> so by admitting they're publishing possibly-false-claims,
> than by trying to stretch and bend it to have some new "as
> used in real life" vaguer meaning.

I don't see "truth" or "falsity" as being the right measure here, since there is no objective notion of truth on the semantic web.  I think "usefulness" is a more appropriate measure.

The problem is that models of the real world are almost always false in some sense -- they are necessarily approximations.  Indeed, we don't even know all the ways in which our models are wrong, since new scientific discoveries happen all the time that change our notions of what we consider true or false.

I do not think the semantics of owl:sameAs should be changed, nor do I advocate sloppiness.  But I *do* think it is important to acknowledge that: (a) an RDF model that is useful for one application may be inadequate (or "false") for another application; (b) that's inevitable and it's okay -- the semantic web does not collapse as a result; and (c) we need to be prepared to deal with that reality when we reuse data.

It is pointless to call someone else's data "false" if in fact it is useful for a number of applications *and* it could be useful for your application *if* you accept the fact that their model is different from yours and you will need to perform some surgery to make their data compatible with yours.

Sure, it's good to warn people if one makes a simplifying assumption that is likely to be incompatible with other people's data.  But people publishing RDF data cannot predict all of the ways their data may be combined with other data, so it isn't possible to warn about everything.



David Booth, Ph.D.
HP Software
+1 617 629 8881 office  |  dbooth@hp.com
http://www.hp.com/go/software

Statements made herein represent the views of the author and do not necessarily represent the official views of HP unless explicitly so stated.

Received on Thursday, 10 July 2008 02:31:35 UTC