- From: Seaborne, Andy <andy.seaborne@hp.com>
- Date: Tue, 12 Oct 2004 16:29:32 +0100
- To: "Phil Dawes" <pdawes@users.sourceforge.net>, <public-rdf-dawg-comments@w3.org>
Hi Phil, Thank you very much for the observation. Keep them coming. At the moment, DAWG is only just on the point of having a protocol specification so the relationship between the various elements of the working group's outputs have yet to be worked out. It is certainly true that target identification is a necessary protocol feature and that the API does also provide a way to identify the target. There seems no necessity for FROM but it can be convenient. There are other local use cases such as scripts - SPARQL is not just a query language embedded in programs - but the API case can be expected to be quite common. Andy on behalf of DAWG -------- Original Message -------- > From: Phil Dawes <> > Date: 26 September 2004 15:43 > > Hi Andy, Hi Dawg comments > > I'm struggling to see the value of having the 'FROM' functionality in > the query language, even in the local case. > > Andy Seaborne writes: (in the Named Containers post) > > [...] > > > > ==== FROM > > > > This is as much about "protocol" as query but its needed for the > local > query case where there isn't a protocol layer." > > > > In the local case I would consider the protocol layer to be the > programming API, and the FROM functionality fits equally nicely in > this layer as it does for the remote protocol. > > Moreover, 'FROM' buys little in terms of interoperablitiy in the local > case, since the client still has to use a seperate api to stage the > RDF store(s) and actually issue the query. > > (all IMHO!) > > Cheers, > > Phil
Received on Tuesday, 12 October 2004 15:30:37 UTC