Re: Meaning of URIRefs

At 8:57 AM -0700 2002-10-25, Bob MacGregor wrote:
>If I understand Sandro's position correctly, then if I use a URI
>foo:bar (when oh when are we going to finally add qnames to RDF?),
>I'm committed to "believing" all of the information
>stored in the corresponding document.

Let me see if I understand Sandro's argument correctly. In the Dog -> 
Mammal example, we have two documents, <Doc-1> and <Doc-2>.

<Doc-1> states:
     <Doc-1#Dog> rdfs:subClassOf <Doc-1#Mammal>.

<Doc-2> states:
     <Doc-2#Spot> rdf:type <Doc-1#Dog>.

There are four statements in discussion:

S1: <Doc-1#Dog> rdfs:subClassOf <Doc-1#Mammal>.
S2: <Doc-2#Spot> rdf:type <Doc-1#Dog>.
S3: <Doc-2#Spot> rdf:type <Doc-1#Mammal>.
S4: <Doc-1#Dog> rdfs:isDefinedBy <Doc-1>.

In addition, there is also the fact that S1 is asserted by <Doc-1>, 
which I'll call F1. This is always true.

F1: <Doc-1> asserts S1

Lastly, according to RDFS, S1+S2 entails S3.

The question is whether <Doc-2> is asserting S3 when it asserts S2.

Sandro's argument, as I understand it, is similar to this:
1. Because <Doc-1#Dog> is a fragment of <Doc-1>, it is defined by <Doc-1>.
    That is, S4 is always true.
2. If a term is defined by a document, then everything that document says
    about the term is true. That is, F1+S4 entails S1.
3. Points 1 and 2 together mean that S1 is always true.
4. Therefore, S2 entails S3. So when <Doc-2> asserts S2 is is also
    asserting S3.

Is this a fair interpretation?

If so, I'll say that I disagree with point 1 and am uncertain about point 2.


Alternately, Sandro might be saying that the use of any term from 
<Doc-1> implies belief in every statement in <Doc-1>.

Let's say that <Doc-1> defines dolphins as a type of fish.

S5: <Doc-1#Dolphin> rdfs:subClassOf <Doc-1#Fish>.
F2: <Doc-1> asserts S5

The stronger form would be: F2+S4+(any statement involving 
<Doc-1#Dog>) entails S5. In which case, if you wanted to talk about 
dogs without asserting that dolphins are fish, you would need to use 
a alternate term for dog, dophin, or fish.
-- 
Dave Menendez - zednenem@psualum.com - http://www.eyrie.org/~zednenem/

Received on Saturday, 26 October 2002 16:36:16 UTC