- From: Danny Ayers <danny666@virgilio.it>
- Date: Thu, 27 Jun 2002 22:48:37 +0200
- To: "Seth Russell" <seth@robustai.net>, "RDF Logic" <www-rdf-logic@w3.org>
>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