Meeting Record

14:27:38 <AxelPolleres> I have two problems

Axel Polleres: I have two problems ←

14:28:14 <AxelPolleres> 1) the definition in Section 18.1.6 Property Paths Patterns sounds misleading to me

Axel Polleres: 1) the definition in Section 18.1.6 Property Paths Patterns sounds misleading to me ←

14:28:50 <AxelPolleres> 2) in the formal definition in http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#defn_evalZeroPath the only thing which is unclear to me is  how nodes(G) is defined

Axel Polleres: 2) in the formal definition in http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#defn_evalZeroPath the only thing which is unclear to me is how nodes(G) is defined ←

14:28:50 <AndyS> Dropping the second "all" is OK - removing the rest is a big change and goes against the original decision.  :X :p{0} ?A matches ?A to :X regardless of whether :X is in the graph or not.

Andy Seaborne: Dropping the second "all" is OK - removing the rest is a big change and goes against the original decision. :X :p{0} ?A matches ?A to :X regardless of whether :X is in the graph or not. ←

14:30:11 <AxelPolleres> ok, got you for that bit...

Axel Polleres: ok, got you for that bit... ←

14:30:21 <AxelPolleres> trying to summarise:

Axel Polleres: trying to summarise: ←

14:31:15 <AxelPolleres> A zero length path matches all subjects and objects in the graph, as well as IRIs explictly given as endpoints of the path pattern to themselves.

Axel Polleres: A zero length path matches all subjects and objects in the graph, as well as IRIs explictly given as endpoints of the path pattern to themselves. ←

14:31:17 <AxelPolleres> ok?

Axel Polleres: ok? ←

14:31:45 <AxelPolleres> (the "to themselves" is the bit that was missing for me)

Axel Polleres: (the "to themselves" is the bit that was missing for me) ←

14:32:26 <AxelPolleres> because a path pattern returns pairs of bindings, doesn't it?

Axel Polleres: because a path pattern returns pairs of bindings, doesn't it? ←

14:34:02 <AxelPolleres> Apart from that, I se that it's a change in semantics, but why do you want the endpoints of patterns to match, even if they don't appear in the graph?

Axel Polleres: Apart from that, I se that it's a change in semantics, but why do you want the endpoints of patterns to match, even if they don't appear in the graph? ←

14:34:15 <AxelPolleres> Intuitivelyt, this doesn't make too much sense to me.

Axel Polleres: Intuitivelyt, this doesn't make too much sense to me. ←

14:34:22 <AxelPolleres> ... to be honest.

Axel Polleres: ... to be honest. ←

14:34:39 <AxelPolleres> (and I do apologise for having overlooked that so far)

Axel Polleres: (and I do apologise for having overlooked that so far) ←

14:36:15 <AxelPolleres> if you are ok to drop that, I am happy to draft the resp. adaption for Definition http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#defn_evalZeroPath

Axel Polleres: if you are ok to drop that, I am happy to draft the resp. adaption for Definition http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#defn_evalZeroPath ←

14:36:33 <AxelPolleres> at least, as a proposal to the list...

Axel Polleres: at least, as a proposal to the list... ←

14:40:50 <AxelPolleres> To concretise this:

Axel Polleres: To concretise this: ←

14:41:38 <AndyS> 1/ The path TC decided things that way 2/ it's needed to make BIND(<uri> AS ?X) ?X :p{0} ?Y matching (<uri> is not in the graph)

Andy Seaborne: 1/ The path TC decided things that way 2/ it's needed to make BIND(<uri> AS ?X) ?X :p{0} ?Y matching (<uri> is not in the graph) ←

14:42:06 <AndyS> How you consider an RD graph in terms of

Andy Seaborne: How you consider an RD graph in terms of ←

14:43:39 <AndyS> graph theory depends on how you consider the node set of the graph.  As RDF is open world, makes sense (to me at least) that all URIs as graph nodes "exist".

Andy Seaborne: graph theory depends on how you consider the node set of the graph. As RDF is open world, makes sense (to me at least) that all URIs as graph nodes "exist". ←

14:45:01 <AxelPolleres> ok, wait

Axel Polleres: ok, wait ←

14:45:09 <AndyS> :X :p{0} ?A is axiomatically true.   It does not depend on the data, the existence of :p in the graph (consider :p is rdfs:subClassOf*)

Andy Seaborne: :X :p{0} ?A is axiomatically true. It does not depend on the data, the existence of :p in the graph (consider :p is rdfs:subClassOf*) ←

14:45:28 <AxelPolleres> that doesn't make sense to me

Axel Polleres: that doesn't make sense to me ←

14:45:44 <AxelPolleres> in BIND(<uri> AS ?X) ?X :p{0} ?Y

Axel Polleres: in BIND(<uri> AS ?X) ?X :p{0} ?Y ←

14:46:09 <AxelPolleres> <uri> is certainly not the endpoint of the path pattern?!?

Axel Polleres: <uri> is certainly not the endpoint of the path pattern?!? ←

14:47:06 <AxelPolleres> why would you need that to match, BTW, if <uri> isn't in the graph?

Axel Polleres: why would you need that to match, BTW, if <uri> isn't in the graph? ←

14:47:39 <AndyS> OK - bad example - but you are hassling me on multiple channels at the moment.

Andy Seaborne: OK - bad example - but you are hassling me on multiple channels at the moment. ←

14:49:41 <AxelPolleres> Andy, hassling you was certainly *not* the intention. I just suggested to switch to IRC, asked you whether that's ok for you, and you seemed to agree, sorry. If you prefer to take this back to email, let me know.

Axel Polleres: Andy, hassling you was certainly *not* the intention. I just suggested to switch to IRC, asked you whether that's ok for you, and you seemed to agree, sorry. If you prefer to take this back to email, let me know. ←

14:51:14 <AndyS> And you directly IM'ed me.

Andy Seaborne: And you directly IM'ed me. ←

14:53:05 <AndyS> So the timeline is email - 14:17:03, IM 14:17:56, IRC 14:19:09

Andy Seaborne: So the timeline is email - 14:17:03, IM 14:17:56, IRC 14:19:09 ←

14:58:59 <AxelPolleres> Andy, all I wanted to get your attention on personal IRC message to discuss a technical  matter on the IRC channel (recorded) which I hoped to speed up finding agreement. Let us take this back to email, if you don't have time, but please don't blame me, I had proposed IRC only if ok you. Thanks! TTYL

(No events recorded for 5 minutes)

Axel Polleres: Andy, all I wanted to get your attention on personal IRC message to discuss a technical matter on the IRC channel (recorded) which I hoped to speed up finding agreement. Let us take this back to email, if you don't have time, but please don't blame me, I had proposed IRC only if ok you. Thanks! TTYL ←

Formatted by CommonScribe

SPARQL Working Group

Minutes of 17 December 2010