- From: Axel Polleres <axel.polleres@deri.org>
- Date: Fri, 22 May 2009 19:31:37 +0100
- To: Eric Prud'hommeaux <eric@w3.org>
- CC: "Seaborne, Andy" <andy.seaborne@hp.com>, Sandro Hawke <sandro@w3.org>, Pat Hayes <phayes@ihmc.us>, "public-rdf-text@w3.org" <public-rdf-text@w3.org>
Maybe I missed that in the thread, but as for defining D-entailment for SPARQL, we should be fine, because we can restrict BGP matching extension accordingly, right? We can just say that graphs with explicit rdf:PlainLiteral typed literals aren't well-formed. Axel Eric Prud'hommeaux wrote: > On Fri, May 22, 2009 at 09:55:03AM +0000, Seaborne, Andy wrote: >> >>> -----Original Message----- >>> From: public-rdf-text-request@w3.org [mailto:public-rdf-text- >>> request@w3.org] On Behalf Of Sandro Hawke >>> Sent: 22 May 2009 01:27 >>> To: Pat Hayes >>> Cc: Axel Polleres; public-rdf-text@w3.org >>> Subject: enforcing the prohibition >>> >>> >>>>> One thing I am not sure still: It was pointed out that we cannot >>>>> prevent people from writing graphs using rdf:text as a datatype >>>>> explicitly. >>>>> Is that a problem? >>>> Well, I think we can very actively discourage them from doing so, and >>>> warning them to expect trouble, and exactly what to expect, if they >>>> do. In fact, nothing will actually break if they do, unless they >>>> expect these things to mean the same as plain literals without using >>>> datatype entailment. Its more likely that they, the publishers. won't >>>> have any problems, but some poor schmuk the other side of the world >>>> won't get their queries answered properly. But if the spec has plainly >>>> said this using rdf:text (or whatever) as a dataype will cause these >>>> problems, and it does, then its going to be easy for people to find >>>> the culprit, which I think is all that we really need to do. Social >>>> pressure will do the rest: blogs will immediately point out that XXX's >>>> RDF is corrupted with the forbidden datatype, etc.. >>> I'm neutral on this option, but one more stick we *could* use is to >>> require RIF systems to reject RDF graphs that use rdf:text as a >>> datatype. >> This seems harsh. "Be liberal with what you accept." > > i have a similar conclusion, but my arguments are: > 1 don't add a new graph validation layer, burden for implementors. > > 2 someone may have clever ideas for it in the future. > >> Andy >> >> >>> RIF already does this with the rif:iri, to try to make sure >>> it doesn't leak out. >>> >>> ...documents importing RDF graphs containing typed literals of the >>> form "http://iri"^^rif:iri must be rejected. >>> >>> -- http://www.w3.org/2005/rules/wiki/SWC >>> >>> We haven't yet added any ImportsRejectionTests to check on this, but we >>> plan to. I don't think OWL 2 such a notion, and I wouldn't want to add >>> it just for this. >>> >>> -- Sandro > -- Dr. Axel Polleres Digital Enterprise Research Institute, National University of Ireland, Galway email: axel.polleres@deri.org url: http://www.polleres.net/
Received on Friday, 22 May 2009 18:32:18 UTC