- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Fri, 07 Mar 2003 04:40:38 +0000
- To: Brian McBride <bwm@hplb.hpl.hp.com>
- CC: Jeremy Carroll <jjc@hplb.hpl.hp.com>, RDF Core <w3c-rdfcore-wg@w3.org>
Brian McBride wrote: > At 18:16 06/03/2003 +0000, Jeremy Carroll wrote: > > >> Brian, please prioritize discussion of this issue since >> it is critical path for WebOnt. >> >> Proposal to close reagle-01, reagle-02. >> >> Summary: >> Use exc-c14n without comments throughout. >> Suggest that parser should (but not SHOULD) canonicalize. > > > I thought we were defining the syntax and semantics of a langauge, and > assuming a processing model. > > If this will solve the webont problem, then can we say nothing at all? > They just have to make sure they use a canonicalising parser? > > Brian > The actual changes I propose that suggest that parser should canonicalize are: 1: adding test(s) that are based on that asusmption (We have not included tests in the test suite that require C14N). 2: adding following note to concepts: [[ Note: For systems which reason about RDF graphs it is suggested that the canonicalization be performed on XML input. The internal representation and non-XML external representations should be in canonical form. ]] There is some webont resistance to the cost of c14n, and the added complexities they perceive. It has not helped that Pat felt a need to spell out the L2V mapping of rdf:XMLLiteral in the semantics doc. By spelling out the obvious implementation we may make it clearer that webont implementors just do a string compare. This text does not preclude other implementations even for reasoning systems. Jeremy
Received on Friday, 7 March 2003 02:22:13 UTC