RE: Implementing statement grouping, contexts, quads and scopes

>I believe the Model Theory does say that the triples are asserted in the
>document which contains them.    If we assume that what is being
>called here
>"baseUri" is the uri of the context of the document; then reading that
>document into that context, should assert those triples in that
>context.  Of
>course it's the agents prerogative whether it reads that document into that
>context.

Hmm - kind of like taking the cherries out of the basket before eating them?

>For example one could define a context called rdfs and assert that it's uri
>was <http://www.w3.org/2000/01/rdf-schema#> with the following triple:
>
>   rdfs contextUri <http://www.w3.org/2000/01/rdf-schema#>.
>
>Then it would be natural to read in the definition triples of rdfs
>from some
>trusted source into that context.
>
>Now suppose some bozo comes along with a rdf file and puts a header on it
>that looks like:
>
><rdf:RDF
>   xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>   xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
>   xml:base="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
>>
>
>... and asserts some really bizarre triples in it.  Note the xml:base!  Now
>if we read this bozo's file according to the way we are being led
>to believe
>is correct, we will totally krap on our ability to interpret rdfs files.
>But suppose we just read the bozo's file into it's own context.  Now at our
>leisure we can examine Bozo's mutations to rdfs and decide whether we wish
>to accept or reject them for our own purposes.  If we accept them, we copy
>his triples into our local context named rdfs.

Yep, the ability to make some kind of separation like this ought to be
useful.

>> but I'm not convinced of the need for formalization (although it probably
>> wouldn't hurt). I don't really see anything that might contradict my
>earlier
>> inclination to leave these matters to the agent. Allowing the agent to
>widen
>> or narrow the scope of the "current document", or in a similar way apply
>its
>> own join semantics would strike me as advantageous, if not
>essential for a
>> semantic web to get working.
>
>I totally agree :)
>
>Seth Russell
>A corollary to the axiom "Anybody can say anything about anything on the
>web." is "Anybody can read anything about anything on the web."

Nice.

Received on Thursday, 27 June 2002 16:56:01 UTC