- From: Frank Manola <fmanola@mitre.org>
- Date: Fri, 28 Feb 2003 10:04:45 -0500
- To: Bob MacGregor <macgregor@isi.edu>
- CC: www-rdf-comments@w3.org, macgreg@isi.edu, bwm@hplb.hpl.hp.com
[Brian--see comment about the need for WG agreement at the end; as a result, perhaps this requires an issue] Bob-- Thanks for the reply. Once again this has been very helpful. OK, here's my idea for what needs to be done to the Primer reification section, following your suggestions: 1. The reificaton section would have an outline something like this: * RDF provides a vocabulary to describe RDF statements (for making "statements about statements") * this is intended to support statements describing the *provenance* of other statements (e.g., who wrote them, when they were written, and so on). * here is how you would use this vocabulary to describe provenance [example]. * notice that in order to provide this description, you need to have a URIref that identifies the actual statement (or graph) you are talking about. RDF does not provide a built-in way of generating such URIrefs, or of indicating how URIrefs are to be associated with individual statements or graphs, any more than it provides a built-in way of generating URIrefs for the tents we described in Section 3, or of associating URIrefs with specific tents. As in those examples, the mechanisms for associating specific URIrefs with specific resources (statements in this case) are outside of RDF. 2. At the end of the reification section, I could add a *brief* paragraph on propositional attitudes going something like this: * some of the statements about statements people might want to make are things like "Lois believes that Superman is strong", or "Phil thinks that George is a clown". These statements describe what are called "propositional attitudes"; e.g. in the first example, the attitude Lois takes (that she believes it) to the proposition that "Superman is strong". * it is straightforward to *syntatically* represent such statements using the reification vocabulary (assuming a URIref has been assigned to the appropriate statement) using the technique already described [example] * However, consistently processing statements involving propositional attributes is known to be extremely complicated [and involves the potential for generating apparent contradictions; cite references]. So anyone using RDF for such purposes needs to be extremely careful, and be aware of the issues described in the references. Does this sound reasonable? If so, I think making these changes involves more than "editorial discretion", so I'd like Working Group agreement that these changes are OK. --Frank -- Frank Manola The MITRE Corporation 202 Burlington Road, MS A345 Bedford, MA 01730-1420 mailto:fmanola@mitre.org voice: 781-271-8147 FAX: 781-271-875
Received on Friday, 28 February 2003 09:44:55 UTC