Chatlog 2011-03-15

From SPARQL Working Group
Jump to: navigation, search

See original RRSAgent log and preview nicely formatted version.

Please justify/explain all edits to this page, in your "edit summary" text.

13:56:14 <AndyS> Present: SteveH, kasei, cbuilara, AndyS, LeeF, bglimm, MattPerry, OlivierCorby, Sandro
13:56:17 <AndyS> LeeF, you could review all but sec 18 and factor in comments by Birte and Matt.  Majority are sec 18 comments - we can process those first then you review.
13:56:26 <LeeF> zakim, this will be SPARQL
13:56:26 <Zakim> ok, LeeF; I see SW_(SPARQL)10:00AM scheduled to start in 4 minutes
13:56:30 <LeeF> AndyS, thanks
13:56:34 <LeeF> Chair: LeeF
13:56:36 <LeeF> Scribe: Andy Seaborne
13:56:39 <LeeF> Scribenick: AndyS
13:56:46 <LeeF> Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2011-03-15
13:57:03 <LeeF> Regrets: Axel, Alex, Chime, Paul
14:00:59 <AndyS> Topic: Admin
14:01:59 <LeeF> PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2011-03-08
14:02:26 <AndyS> seconded
14:02:36 <LeeF> RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2011-03-08
14:02:53 <AndyS> LeeF: No new comments to assign
14:03:06 <LeeF> Next regular meeting: 2011-03-22 @ 14:00 UK / 10:00 EST (scribe: Lee F) NOTE DIFFERENT TIME FOR NON-US LOCATIONS
14:03:07 <AndyS> ... Axel will do a round up next week.
14:03:26 <AndyS> scribe next time: LeeF
14:03:38 <LeeF> steveh at risk
14:03:48 <LeeF> topic: to last call
14:04:14 <AndyS> Query reviews from Birte and Matt
14:04:29 <AndyS> ... in email
14:04:58 <AndyS> http://lists.w3.org/Archives/Public/public-rdf-dawg/2011JanMar/0433.html
14:05:06 <AndyS> http://lists.w3.org/Archives/Public/public-rdf-dawg/2011JanMar/0436.html
14:05:28 <LeeF> close ACTION-391
14:05:29 <trackbot> ACTION-391 Review Query, particularly Section 18 closed
14:05:35 <AndyS> just waiting for LeeF now :-)
14:05:50 <AndyS> ... and Axel.  Spot the connection.
14:06:46 <AndyS> AndyS update reviews http://lists.w3.org/Archives/Public/public-rdf-dawg/2011JanMar/0421.html and http://lists.w3.org/Archives/Public/public-rdf-dawg/2011JanMar/0427.html
14:08:09 <LeeF> bglimm: 2 significant issues - no connection between EXISTS/NOT EXISTS and algebra & property paths break entailment extension mechanism
14:08:17 <LeeF> AndyS: re first, that's an omission
14:10:08 <AndyS> ... entailment doc was going to discuss entailment+prop paths -- still some discussion about what's in query doc.
14:10:31 <AndyS> Kendal : his org is not able to review the entailment doc.
14:10:45 <AndyS> LeeF: One view from Oracle (link?)
14:10:52 <bglimm> I can ask Bijan, but I asked in our group here and no volunteers :-(
14:11:18 <AndyS> ... will ask semweb co-ordination group - maybe someone from OWL2.
14:11:47 <bglimm> I can also ask Ian whether he can think of somebody
14:12:01 <AndyS> ... do want 2 (+) reviews before LC 
14:12:06 <MattPerry> Oracle entailment review from Zhe Whu (http://lists.w3.org/Archives/Public/public-rdf-dawg/2011JanMar/0217.html)
14:12:11 <LeeF> ACTION: Lee to ask for pre-last call reviews of entailment doc from SWCG and OWL WG
14:12:12 <trackbot> Created ACTION-411 - Ask for pre-last call reviews of entailment doc from SWCG and OWL WG [on Lee Feigenbaum - due 2011-03-22].
14:12:55 <AndyS> Topic: Actions
14:13:02 <AndyS> None to close
14:13:06 <LeeF> topic: Dataset protocol & service description
14:14:08 <AndyS> Little email discussion after last telecon
14:14:32 <LeeF> PROPOSED: The SPARQL 1.1 Service Description describes endpoints that implement the SPARQL 1.1 Protocol; it does not describe URIs that respond to the SPARQL 1.1 RDF Dataset HTTP Protocol.
14:14:32 <Zakim> +Sandro
14:14:36 <sandro> (sorry I'm late)
14:14:39 <AndyS> ... proposal is therefore the minimal case that service description covers protocol, not HTTP dataset protocol
14:15:07 <AndyS> ... anyone have concerns about this approach?
14:15:11 <LeeF> <silence>
14:15:19 <kasei> +1
14:15:38 <LeeF> RESOLVED: The SPARQL 1.1 Service Description describes endpoints that implement the SPARQL 1.1 Protocol; it does not describe URIs that respond to the SPARQL 1.1 RDF Dataset HTTP Protocol.
14:16:04 <AndyS> LeeF: Overview doc needs to point this out. 
14:16:25 <AndyS> Kasei: SD doc does not mention dataset protocol.
14:17:06 <AndyS> LeeF: make clear statement about what is described and leave to overview doc.
14:17:25 <LeeF> ACTION: Lee to make sure overview document is clear about (non-)relationship between service description and rdf dataset http protocol
14:17:25 <trackbot> Created ACTION-412 - Make sure overview document is clear about (non-)relationship between service description and rdf dataset http protocol [on Lee Feigenbaum - due 2011-03-22].
14:18:29 <AndyS> Topic: relationship between URI of a graph store at which they can issue dataset protocol requests? 
14:19:04 <AndyS> LeeF: a bit unclear currently : not an issue in SPARQL 1.0 in practice,
14:20:09 <AndyS> SteveH: /sparql/dataset and /data/dataset 
14:21:04 <AndyS> AndyS: Fuseki uses /dataset/sparql,  /dataset/update and /dataset/data (and /dataset)
14:22:37 <AndyS> LeeF: SD assumes you have a URI to GET to start process
14:23:56 <AndyS> LeeF: Leave as is for now there being no proposals
14:24:09 <AndyS> Topic: More operations besides POSTing to a Graph Store 
14:24:00 <LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011Feb/0013.html
14:24:22 <LeeF> """
14:24:23 <LeeF> However my comment relates to management of Datasets, not graphs. I
14:24:23 <LeeF> think it would be useful to be able to:
14:24:23 <LeeF> * Add/Remove Datasets
14:24:23 <LeeF> * Add/Remove Graphs from Datasets
14:24:23 <LeeF> The specification already discusses the addition and removal of Graphs.
14:24:25 <LeeF> """
14:25:26 <AndyS> q+ to ask why point 2 isn't always covered
14:25:29 <kasei> I don't understand what the first one means. How would you then use a new dataset if it's not a 'default' dataset?
14:26:03 <kasei> q+ to clarify datasets vs. SD
14:26:06 <AndyS> LeeF: We do not provide a lifecycle for datasets
14:26:09 <LeeF> ack AndyS
14:26:09 <Zakim> AndyS, you wanted to ask why point 2 isn't always covered
14:27:02 <AndyS> Leigh may have a model of a set of graphs that can be assembled in datasets
14:27:23 <LeeF> ack kasei
14:27:23 <Zakim> kasei, you wanted to clarify datasets vs. SD
14:27:54 <AndyS> Who's handling the comment?
14:29:20 <AndyS> Thread for LD-3
14:30:52 <AndyS> leeF: advice to Chime that the request isn't clear
14:31:19 <LeeF> AndyS: work in RDF WG on quads may enable Leigh's request in the future
14:31:21 <AndyS> AndyS: Some work in RDF-WG may impact this.
14:31:39 <AndyS> LeeF: will follow up with Chime.
14:32:00 <AndyS> Topic: D-entailment canonicalization issue
14:32:41 <AndyS> Birte: problem is that there are multiple lexical forms for the same value.
14:33:10 <AndyS> ... no specification as to whether canonicalization should be done.
14:33:36 <AndyS> ... if no canonicalization done, may get multiple answers for same value.
14:34:05 <AndyS> ... and also issue as to which datatypes are supported.
14:34:26 <AndyS> LeeF: What is state of the art?
14:34:30 <SteveH> silence?
14:35:07 <AndyS> Birte: XSD numerical datatypes common 
14:36:11 <LeeF> AndyS: what about overlapping value spaces of decimal and double?
14:36:21 <LeeF> bglimm: XSD says that the value spaces are disjoint
14:36:44 <AndyS> LeeF: What are the options?
14:37:00 <LeeF> http://www.w3.org/2009/sparql/docs/entailment/xmlspec.xml#id0x0fec6340
14:37:40 <LeeF> bglimm: 1) remove D-entailment 2) require canonicalization and identify a set of datatypes
14:37:44 <AndyS> Birte: (1) remove d-entailment (2) require canonicalization + some datatypes (3) no changes to doc
14:37:45 <LeeF> AndyS: a fixed set of datatypes?
14:37:49 <LeeF> bglimm: no could be extended
14:39:16 <MattPerry> I would vote for option 2
14:40:55 <MattPerry> Oracle does it for most xsd types
14:41:23 <MattPerry> numeric, date, boolean, string
14:42:23 <bglimm> OWL 2 doesn't support xsd:date, just xsd:dateTime & xsd:dateTimeStamp
14:42:54 <AndyS> What's dateTimeStamp?
14:43:20 <AndyS> Canonicalization is done by a few systems - they do not claim D-ent
14:43:22 <bglimm> I would have to look it up
14:43:53 <MattPerry> Same at Oracle, we don't say anything about D-entailment
14:44:16 <MattPerry> just canonicalization
14:44:18 <AndyS> AndyS: e.g. <s> <p> 1 . <s> <p> 01 .
14:44:56 <AndyS> Birte: one triple counts for query answers
14:45:27 <bglimm> D-Entailment includes also RDFS Entailment
14:45:55 <AndyS> LeeF; D-ent only of interest to systems that do not canonicalize.
14:46:04 <AndyS> .. any such systems?
14:46:38 <AndyS> ... now understand better.  Take one week to survey and if too few, remove section.
14:46:45 <MattPerry> That's fine
14:47:06 <AndyS> AndyS: Acceptable.
14:47:22 <AndyS> LeeF: any other business?
14:47:41 <AndyS> AndyS: tests?
14:48:12 <AndyS> LeeF: some pending - maybe a "tests telecon" - some open actions on checking and extending tests.
14:49:01 <AndyS> LeeF: Repeat for encouragement - checking does not take too long.
14:50:30 <AndyS> LeeF: Semtech June seems a bit late for a F2F
14:50:56 <AndyS> ... could get together for impl report etc as we might be in CR
14:51:17 <MattPerry> bye
14:51:18 <AndyS> ADJOURNED
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000200