RE: URI Declarations [Grounding]

> From: Pat Hayes []
> [ . . . ] one cannot get grounding by making
> purely formal assertions,

I agree.

> even if they are called
> 'declarations' and you have to believe them, on pain of being
> very naughty. David wants to get grounding, but he's going
> about it the wrong way.

No, that's *not* what I'm trying to do, though I evidently need to improve my description to be clearer.  (Hopefully these discussions will help me to do so.)

URI declarations are not about achieving grounding any more precisely than any other way of achieving grounding.  They are about clarifying the *mechanism* by which an application should *select* the appropriate grounding for a given URI, to whatever extent such a grounding is available and is expressible as a set of assertions (which may well include reference to natural language descriptions, published scientific articles, etc.).  Given a statement:

     _:foo _:bar <> .

the architectural guidelines for URI declarations help address the question: "What algorithm should an application use to determine the "meaning" of <>?"  (To the extent that an application can *ever* determine the "meaning" of it, that is.)

Another way to think of a URI declaration is that it helps answer this question:

QUESTION: When we say that a URI *denotes* a resource, what should that mean operationally to a Semantic Web application?

ANSWER: It means that if the application encounters that URI as denoting a resource, it should assert the "core assertions" from the URI's declaration: use of the URI implies agreement to those "core assertions".

Note: Case 4b -- domain hijacking -- described in
is an exception in which the community accepted meaning needs to supercede the architecturally defined meaning.  That case can be efficiently handled by a Web proxy that redirects, to an alternate declaration, the application's follow-your-nose attempt to dereference the URI in search of its URI declaration.

David Booth, Ph.D.
HP Software
+1 617 629 8881 office  |

Opinions expressed herein are those of the author and do not represent the official views of HP unless explicitly stated otherwise.

Received on Wednesday, 5 March 2008 07:57:31 UTC