Re: problems with concise bounded descriptions

From: <Patrick.Stickler@nokia.com>
Subject: RE: problems with concise bounded descriptions
Date: Fri, 1 Oct 2004 14:42:16 +0300

[Edited to make the example clearer and fix possible problems in it.]

> > -----Original Message-----
> > From: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
> > Subject: Re: problems with concise bounded descriptions
> > Date: Fri, 01 Oct 2004 05:24:57 -0400 (EDT)

[...]

[Problem 7:  This definition does not provide enough 
 information to
 distinguish the node from other distinguishable nodes 
in the graph.
 Consider, for example, the RDF graph: 
	_:z ex:r _:a .
	_:w ex:r _:b .
 	ex:r rdf:type owl:InverseFunctionalProperty .
 	_:a ex:r _:b .
 	_:b ex:r _:a .
 	_:a ex:s "NODE A" .
 	_:b ex:s "NODE B" .
 Then the CBD of both _:z and _:w in this graph is
	_:x3 ex:r _:x1 .
 	_:x1 ex:r _:x2 .
 	_:x2 ex:r _:x1 .
  so two distinguishable nodes have the same CBD.

> I'm sorry, but I don't see why any such requirement 
> would exist.
> 
> What does it matter if the CBDs of two resources are the same
> graph?

Because the only point I can see for CBDs is to enable access to all the
information a server knows about a node in an RDF graph.  This example
shows that your definition of CBD fails to enable this.

> Aside from the fact that one agent cannot request from
> another resource a CBD of a resource denoted solely by
> an anonymous node.

Your definition allows for CBD of blank nodes to be determined, and there
are many cases where this even more useful than determining the CBD of a
URIref node.  (For example, a foaf server returning information about all
instances, including anonymous ones, of foaf:Person.)

> If we modified your example to one using URIrefs:
> 
>    ex:a ex:r ex:b .
>    ex:b ex:r ex:a .
>    ex:a ex:s "NODE A" .
>    ex:b ex:s "NODE B" .
> 
> then the CBD of ex:a is
> 
>    ex:a ex:r ex:b .
>    ex:a ex:s "NODE A" .
> 
> and the CBD of ex:b is
> 
>    ex:b ex:r ex:a .
>    ex:b ex:s "NODE B" .
> 
> which are distinct; and also some agent X needing a 
> CBD of ex:a or ex:b can ask some other agent Y which
> may know something about those resources.

So what?   All this shows is that in some cases your CBD definition may
return useful information.

Perhaps you are claiming that the problems only exist for CBDs of blank
nodes.  Well, perhaps the particular technical problem only occurs for CBDs
of blank nodes.

However, change my example above, replacing _:z with ex:z and _:w with
ex:w.  Then the CBDs for both ex:z and ex:w do not allow subsequent access
of all relevent information, showing that your definition of CBD is
inadequate even for URIref nodes.

> Patrick


Peter F. Patel-Schneider
Bell Labs Research

Received on Friday, 1 October 2004 13:01:41 UTC