- From: Ron Daniel <rdaniel@interwoven.com>
- Date: Mon, 2 Apr 2001 15:53:08 -0700
- To: "'Aaron Swartz'" <aswartz@swartzfam.com>, "'Dave Beckett'" <dave.beckett@bristol.ac.uk>, "'RDF Interest'" <www-rdf-interest@w3.org>
- Cc: "'spec-comments'" <spec-comments@prismstandard.org>
Aaron said: > I understand this -- the issue is not that the RDF is bad, > but that generic > RDF processors cannot be used as PRISM processors. In other > words, I can't > take SiRPAC or Redland off-the-shelf and use it to manipulate > PRISM. This > may not be a major issue, but it does bother me that you're > requiring the > use of special tools to use PRISM and not being truly open RDF. If this is the consensus of the RDF interest list I can certainly take that new material back out. But please consider: 1) PRISM is an application of RDF. RDF is a generic framework. PRISM has a particular purpose, and thus there is a surrounding context of assumptions and decisions. Just like an XML parser has to be supplemented in order to do the application-specific logic, RDF tools will have to be supplemented to do the PRISM-specific stuff. In my case, the RDF tool I have been using is RDFFilter, so dealing with order does not set off alarm bells for me. Its just extra PRISM-specific logic to track the order they came out of RDFFilter. That will not be perfectly accurate, but it sure looks like a 90/10 point to me. 2) PRISM applications are REQUIRED to produce legal RDF that can be handled by any fully-compliant RDF processor. Only PRISM processors would be required to preserve the order of statements. So, generic RDF software will read and write the descriptions just fine. They won't be aware of some higher level application-specific info. They may swap the order of some things. That is to be expected. They were not developed to be PRISM compliant. Ron
Received on Monday, 2 April 2001 18:55:10 UTC