Meeting Record

There are some format problems with the chatlog. Please correct them and reload this page. They are labeled on this page in a red box, like this message.

It may be helpful to

Zakim IRC Bot: the conference code is 26632 (tel:+1.617.761.6200 tel:+33.4.89.06.34.99 tel:+44.117.370.6152), AndyS_ ←

Lee Feigenbaum: AndyS_, did you take a look at the updated crq? did I miss anything important? :) ←

Andy Seaborne: I only had time to scan it - just back from ESWC - but it looked fine to my scanning. ←

Lee Feigenbaum: last time i talked to him he was floating in the middle of a lake ←

Lee Feigenbaum: at-risk feature that allows implementations to return between 1 and n copies of a given solution (where n is the "proper" number as given by the algebra) ←

Lee Feigenbaum: Motivated by implementation experience from Simon and SteveH, primarily ←

Andy Seaborne: Allows impls to do some dup removal but not perfect DISTINCT. Certain implementers argued for it. ←

Ivan Mikhailov: hm. So if I use reduce, I may get different answers depending on which engine I use? ←

Lee Feigenbaum: you may receive different answers against the same engine twice ←

Andy Seaborne: Yes :-( but common practice might cluster around adjacent duplicate removal (ever hopeful here) ←

Zakim IRC Bot: the conference code is 26632 (tel:+1.617.761.6200 tel:+33.4.89.06.34.99 tel:+44.117.370.6152), ivan ←

Zakim IRC Bot: the conference code is 26632 (tel:+1.617.761.6200 tel:+33.4.89.06.34.99 tel:+44.117.370.6152), timbl_ ←

Lee Feigenbaum: RDF is a directed, labeled graph data format for representing information in the Web. RDF is often used to represent, among other things, personal information, social networks, metadata about digital artifacts, as well as to provide a means of integration over disparate sources of information. This specification defines the syntax and semantics of the SPARQL query language for RDF. ←

Unknown timbl_: Can we rephrase the SPARQL question as "return bindings such that this pattern, substituted for each set of bindings, is upported by the kB"? ←

Unknown timbl_: Then you can simply advertize the capabilities of the KB as metadata ←

Andy Seaborne: Sounds like it is about the endpoint not the QL per se (service description is postponed). ←

Unknown TimBL: during CR, collect as many examples of implementation and note all their features. That set of features can be a starting point for service description requirements [ Scribe Assist by Unknown RalphS ] ←

Unknown TimBL: but in general SPARQL should not require any particular entailment [ Scribe Assist by Unknown RalphS ] ←

Unknown RalphS: ... point out that a taxonomy of services has not yet been created ←

Unknown timbl_: informatively rather than normatively, point out a list of things that implementations already do, and characterize the things that have not been created yet [ Scribe Assist by Lee Feigenbaum ] ←

Unknown RalphS: I should think that the implementation report is a fine place to start to document this but I don't think it's necessary to have a complete enumeration of service features to exit CR (so long as the features in the CR spec are all shown to be implemented) ←

Ivan Mikhailov: I wasn't recommending eliminating UNBOUND from the language; I was ←

Ivan Mikhailov: relegating it to secondary status within the language, i.e., making it ←

Ivan Mikhailov: a computed predicate and not according it a reserved word. Its ←

Unknown MacGregor: seems to need computed predicates [ Scribe Assist by Unknown RalphS ] ←

Unknown Timbl: no obvious how you use a computed predicate [ Scribe Assist by Unknown timbl_ ] ←

Andy Seaborne: ack : UNSAID is sim to log:notIncludes - the graph scope is explicit ←

Unknown RalphS: a well-crafted operator, UNBOUND is a hack that the language would be ←

Unknown RalphS: below). Thus, my suggestion would be to drop it from the language, ←

Unknown RalphS: but, as Andy notes, UNSAID is quite much harder than UNBOUND. so the 'hack' can be demonstrated to be useful by implementations. ←

Unknown timbl_: is there an algorithmic mapping from unsaid to unbound? [ Scribe Assist by Lee Feigenbaum ] ←

Unknown timbl_: 2 things - unsaid as future work for SPARQL, during CR look at relationship between unsaid and !bound [ Scribe Assist by Lee Feigenbaum ] ←

SPARQL Working Group

Minutes of 07 June 2007