- From: Alexandre Bertails <bertails@w3.org>
- Date: Sun, 04 Nov 2012 15:35:33 +0100
- To: "Wilde, Erik" <Erik.Wilde@emc.com>
- CC: "nathan@webr3.org" <nathan@webr3.org>, "public-ldp-wg@w3.org" <public-ldp-wg@w3.org>, Andrei Sambra <andrei@fcns.eu>
On 11/03/2012 03:10 PM, Wilde, Erik wrote: > hello nathan. > > On Nov 3, 2012, at 4:51, "Nathan" <nathan@webr3.org> wrote: >> I'd suggest that the vocabulary/predicates used are the thing that need >> specified, since "MUST be RDF" doesn't really help anybody understand >> the messages. > > that's what i was suggesting with mapping the error document model to an RDF schema of some sort. MUST would be too strong, though, but we might want to say that "error documents SHOULD support RDF representation(s)". not sure we should go as far as suggesting a specific syntax. If we don't make it a MUST, then a client will be required to support many alternatives as it has no certitude that RDF (being TURTLE or not) can be returned in all cases. I believe that this is an unwanted burden on the client that we must avoid. Alexandre. > > cheers, > > dret. >
Received on Sunday, 4 November 2012 14:35:42 UTC