- From: Dave Beckett <dave.beckett@bristol.ac.uk>
- Date: Fri, 04 Apr 2003 10:24:42 +0100
- To: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- cc: w3c-rdfcore-wg <w3c-rdfcore-wg@w3.org>, duerst <duerst@w3.org>
>>>Jeremy Carroll said: > > > See (member only) > http://lists.w3.org/Archives/Member/w3c-archive/2003Apr/0047.html > for the full message. > > > From: Martin Durst > -----(Edited) Original Message----- > ... > > Jeremy: > > > >I wanted to express some urgency on the request concerning IRIs and NFC. > > ... > > > >It would be very helpful to RDFCore if we can have a simple yes/no to the > >question "Should RDF restrict its use of IRIs to NFC?", ideally before our > >telecon tomorrow (approx 21 hours from now). > > > ... > > I think making no mention for NFC in IRIs is better, because otherwise > you may get into conflicts with the IRI spec if it is tweaked. (I don't > expect any big changes regarding NFC in IRIs, but tweaks might happen.) > > On the question of NFC for RDF overall, as you personally know, we have > backed up a bit from the position that everybody MUST check and reject. > Check and reject for NFC is now just a SHOULD. Great, that sounds good i.e. we will be compatible (by weakening our NFC MUST to a SHOULD) but not directly based on the still-changing IRI work. Apart from wordings, lThat means any of our test cases that previously tested the MUST need to be removed and made optional. That will need a new type of test. > What is definitely still true, in all cases, is that 'applications' > (such as RDF) MUST NOT normalize stuff themselves. OK. I think that is already record. Dave
Received on Friday, 4 April 2003 04:27:24 UTC