- From: Seaborne, Andy <andy.seaborne@hp.com>
- Date: Sat, 23 May 2009 15:54:49 +0000
- To: Axel Polleres <axel.polleres@deri.org>, Pat Hayes <phayes@ihmc.us>
- CC: Eric Prud'hommeaux <eric@w3.org>, Sandro Hawke <sandro@w3.org>, "public-rdf-text@w3.org" <public-rdf-text@w3.org>
> -----Original Message----- > From: Axel Polleres [mailto:axel.polleres@deri.org] > Sent: 23 May 2009 01:17 > To: Pat Hayes > Cc: Eric Prud'hommeaux; Seaborne, Andy; Sandro Hawke; public-rdf-text@w3.org > Subject: Re: enforcing the prohibition > > Pat Hayes wrote: > > > > On May 22, 2009, at 1:31 PM, Axel Polleres wrote: > > > >> 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? Yes - that's the clause about the scoping graph can only contain current-RDF-style literals with no use of rdf:PlainLiteral as a datatype. > >> We can just say that graphs with > >> explicit rdf:PlainLiteral typed literals aren't well-formed. No, if you mean tehthing being queried with the BGP - the restriction should apply to the boundary of SPARQL to the entailment regime. The scoping graph not the graph being queried. It's quite reasonable that the queried graph contain rdf:PlainLiteral. It may not be an RDF graph; consider SPARQL/OWL or a graph which does grok rdf:PlainLiteral. What matters is the form of the results, not what's in the OWL/D-entailment engine. Andy > > > > We can say that, yes, but there is nothing in the current RDF or SPARQL > > specs that say this. > > just wanted to make sure with you that this is a feasible approach... > > > So what should a conforming SPARQL/RDF engine do, > > if it comes across one? Apart from reporting it to the OWL/RIF militia, > > that is. That is why I think having a named 'convention' that engines > > can say they support, or not, is useful. An engine which does can flag > > this as an error with a clear conscience, and its owners can cite the > > relevant W3C document when challenged, and nobody has to refer to OWL or > > RIF (inviting the response: so what, I'm not using those, just RDF...) > > Note, just saying that you support {rdf:text}-entailment isn't going to > > be enough. > > ... just as SPARQLOWL entailment migt refuse graphs containing graphs > that comprise incomplete OWL axioms (e.g. having malformed RDF lists, etc.) > > > > > Pat > > > >> > >> Axel > >>
Received on Saturday, 23 May 2009 15:57:55 UTC