Minutes of RDF Data Access WG 2005-09-13


1. Convene, take roll, review records and agenda

RDF Data Access WG teleconference 2005-09-13T14:30Z
Agenda
IRC log
scribe: Dave Beckett
roll call: regrets JosD
present: LeeF, AndyS, DaveB, EliasT, YoshioF, KendallC, DanC, KendallC, SteveH, RachelY, PatH

PROPOSED: to accept minutes 6 Sep as a true record
approved

next meeting: 20 Sep 2005
chair: DanC has a conflict with TAG ftf meeting, Pat Hayes to chair (ericP, KendallC offered backup)
scribe volunteer: SteveH

Actions continued without discussion:

2. Toward Protocol Last Call

ACTION KC: make changes as agreed in 6 Sep discussion
1.71 editor's draft ready for LC pub  (DONE)
ACTION EricP: publish proto-wd v 1.68 + changes from as LCWD (CONTINUES)

Discussion of Namespace for SPARQL query interfaces in
(editor's draft) which had to be updated during the publication process last-minute.

3. SPARQL Protocol and WSDL

ACTION: LeeF to draft WSDL 1.1 for SPARQL thingy with AndyS and Elias ETA 9 sep (CONTINUES)
Lee reported progress - the first draft is done but needs some polishing; there is example code in java zip file and python available, should be able to get it out soon ("this week").
ACTION: DanC to ask WSDL WG to review WSDL 1.1 and WSDL 2 SPARQL protocol stuff, once both are available  (CONTINUES)
ACTION: KC to work with WSDL WG on describing POST binding with application/x-form-encoded in WSDL 2  (CONTINUES)
cf Data Access WG questions about WSDL of 15 Aug 2005
KendallC reported that he has sent a mail with one three issues, two of which are similar. One of the issues has been turned into a WSDL last call issue with a tracking number. See the WSDL WG last call issues list.
ACTION: KC to work with WSDL WG on moving "style" from interface to binding   (CONTINUES)
FYI: simple case of IRIs for Components in WSDL 2.0 Dan Connolly (Friday, 9 September) to public-ws-desc-comments@w3.org

4. comment: SPARQL Protocol: inconsistent parameter names

ACTION: KC to respond to inconsistent parameter names commentor (DONE in KC message)

5. issues#valueTesting : "language tag issues"

ACTION: ericp to draft lang-match design, summarizing and citing RFC3066 (CONTINUES)

6. BASE IRI resolution comment

ACTION: ericP to send [OK?] message to Bjoern. re BASE IRI resolution comment (CONTINUES)

7. issues#valueTesting: handling type "error"s

ACTION: EricP to get SteveH's response to 0360 modifying truth table for OR (DONE)

Discussion of proposals from EricP (after commentor), JeenB and DaveB.

PROPOSED: to change the value testing truth table as per table 1 of 2005JulSep/0410.html
RESOLVED
ACTION DaveB: add to test suite the temperature case from comment on truth tables in commentor's message
ACTION EricP: respond to commentor on valueTesting issue

8. issues#rdfSemantics

note new issue; enumerate known options; initial straw poll
ACTION AndyS: write email with 2 use cases on rdf-semantics issues for references
Discussion of various approaches to the issue including having core/full sparql splits, a separate document which could range from a paper, WG note or larger document. Some preference to minimal solutions such as how DISTINCT was handled.

9. issues#owlDisjunction

note new issue; enumerate known options; initial straw poll
Discussed the worker example" or "little house example". These last two issues will attempt to be discussed in detail in the 20Sep, 4Oct telcons.

10. SPARQL QL grammar

rq23 grammar update Seaborne, Andy (Wednesday, 24 August)
AndyS reported the grammar state - it is now LL(1) and the major changes are all done. digits and underscores are now allowed at the start of variable names (VARNAME, rule 84).
Some grammar changes might still happen depending on how any result of language issues are resolved; could be a new builtin keyword, a qnamed function, ... test cases, e.g. SELECT ?x WHERE {.}
ACTION SteveH: review the SyntaxDev grammar tests
ACTION DaveB: review SyntaxDev grammar tests
various comments on variable names, etc.

11. issues#badIRIRef

Looking at SELECT ?x WHERE { <foo###bar> dc:title ?x } and recent grammar work.
Discussion of wording about how SPARQL references valid IRIs.
Where does it say that a query with an unbound prefix is invalid?
PROPOSED: that SELECT ?x WHERE { <foo###bar> dc:title ?x } is not a SPARQL query string; that Q_IRI_REF's must conform to generic IRI syntax
RESOLVED hayes, harris, andy, ericp abstain.