W3C

RDF DAWG Weekly

27 Feb 2007

Agenda

See also: IRC log

Attendees

Present
EliasT, LeeF, AndyS, Orri_Erling, PatH, SimonR
Regrets
ericP, Souri
Chair
LeeF
Scribe
EliasT

Contents


 

 

<LeeF> Scribe: EliasT

<SimonR> Andy, found a solid answer: dc:author never existed. Totally in my imagination. :) http://askdcmi.askvrd.org/default.aspx?id=17213&cat=1720

<LeeF> SimonR, are you calling in?

<LeeF> and iv_an_ru, will you be calling in?

<iv_an_ru> Start without me

<SimonR> Trying, Lee.

<SimonR> Feel free to kick off. I'll wrangle it eventually.

<LeeF> minutes from 13 February: http://www.w3.org/2007/02/13-dawg-minutes

AndyS: seconds minutes from Feb 13

<SimonR> I'll scribe next week.

<LeeF> meet next Mar 6, SimonR to scribe

review action items

<LeeF> ACTION: Elias to add wording for PROPOSED: ed(The SPARLQ Protocol does not derefrence query URIs so 5.1.3 does not apply. Per 5.1.4, services must define their own base URI, which may be the service invocation URI.) [DONE] [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action01]

<LeeF> ACTION: Lee to talk to protocol editors re: POSTing application/sparql-query [DONE] [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action02]

<LeeF> ACTION: AndyS to add text clarifying the prohibition on blank node labels in multiple BGPs to rq25 [CONTINUES] [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action03]

<LeeF> ACTION: EricP to run the yacker tool over and annotate the existing tests [CONTINUES] [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action04]

<LeeF> ACTION: LeeF to remember that the wee, lost filter tests should be put [CONTINUES] [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action05]

ORDER BY, language tags, and unknown types

<LeeF> Eric's suggestion: http://lists.w3.org/Archives/Public/public-rdf-dawg/2007JanMar/0111.html

<LeeF> Andy's response: http://lists.w3.org/Archives/Public/public-rdf-dawg/2007JanMar/0118.html

<LeeF> original comment: http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2007Feb/0005.html

<AndyS> Original : http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2007Feb/0005.html

<LeeF> Andy's summary: http://lists.w3.org/Archives/Public/public-rdf-dawg/2007JanMar/0098.html

<SimonR> The only comment I have is that I vaguely thought that language codes were not case sensitive. The list traffic I read gave me the impression we weren't stripping case before doing comparisons.

<LeeF> Andy: The text currently says you can't use extended definition of < to do ORDERing, which I take as a bug

<LeeF> Andy: We can't define a total ordering

<LeeF> SimonR, language codes are case insensitive, according to RDF Concepts

<SimonR> LeeF, I'm surprised and newly enlightened!

AndyS: SPARQL doesn't support language tags
... I think we should support language tags in the spec

SimonR: I think we should. If we don't, we'll get hell from i18n.

<SimonR> (That was Pat, not me.)

<patH> that was patH

<patH> we brits al sound alike.

<AndyS> @en-uk

LeeF: Unless there's a proposal that we revisit language tags support, we should discuss extensions and less-than only.

AndyS: We do date, but would like datetime.

<AndyS> Mute is 61# / unmute is 61#

AndyS: We do datetime, but would like date support.

LeeF: does anybody have any thoughts on making it explicit on the text that implementations of extensions that extend less than can affect the order of solutions?

PatH: No strong feeling. Pitty we can't follow some precedence.

Orri: We inherit the SQL ordering.

SimonR: We support LIMIT/OFFSET so we need to provide a stable ordering support.

<AndyS> "001"^^xsd:integer and "01"^^xsd:int

LeeF: Given AndyS and ericP's comments + the community feedback, convinces me that we need to allow for implementation to extend ordering in the less-than operator. Although this could result in less interoperability in our spec in regards to ordering.

ericP: The main risk we run is that a new implementation will affect somebody's expected slicing.

SimonR: In order to solve this problem for real, we need to have a separate protocol that holds on to the answers in order for you to obtain correct slices from it.

Orri_Erling: We can inherit a scrollable cursor from SQL.

<LeeF> PROPOSED: ORDER BY comparisons respect extended implementations of < that operate on types beyond what's given in the operator table

SimonR: Mulgara supports a holding-results set.

AndyS: Seconds it.

SimonR: Is < the only operator that gets extended in this way?

AndyS: We only use one operator to build the table.

patH: what is R.E.S.P.E.C.T?

<LeeF> PROPOSED: ORDER BY comparisons may use extended implementations of < that operate on types beyond what's given in the operator table

<LeeF> so RESOLVED.

<SimonR> Would extended versions of equality (particularly for datatype processing, affect DISTINCT?

<AndyS> DISTINCT is term distinct, not value distinct.

<LeeF> ACTION: ericP to add text to spec noting that ORDER BY comparisons may use extended implementations of < that operate on types beyond what's given in the operator table [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action06]

unexpected DISTINCT

<AndyS> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2007Feb/0014.html

<LeeF> test case: http://www.w3.org/2001/sw/DataAccess/tests/#modifer-limit

<LeeF> ericP: Fred is in favor of explicitly NOT having auto distinct

ericP: If people really wanted that, we could add a keyword, but only until someone proposes.
... I'm with Fred.

<LeeF> zkim, mute ericP

<LeeF> SimonR: Mulgara always does auto distinct

<LeeF> ACTION: LeeF to talk to SteveH and JeenB about auto distinct behavior in implementations [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action07]

rq25 status and reviews

<AndyS> I'm currently working through the reviews sent by KendallC, Bijan and SimonR.

<LeeF> Thanks all reviewers for their reviews sent so far and welcome any more reviews from the WG.

protocol status

LeeF: Does anybody have anything invested in posting SPARQL queries?

AndyS: Joseki can POST a plain query. It doesn't have to be in the specification.

Orri_Erling: For our purposes it's ok as it is.

ericP: I don't care.

<LeeF> ACTION: LeeF or EliasT to reply to Bjoern regarding (not) POSTing application/sparql-query documents [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action08]

ADJOURNED

Summary of Action Items

[NEW] ACTION: ericP to add text to spec noting that ORDER BY comparisons may use extended implementations of < that operate on types beyond what's given in the operator table [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action06]
[NEW] ACTION: LeeF or EliasT to reply to Bjoern regarding (not) POSTing application/sparql-query documents [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action08]
[NEW] ACTION: LeeF to talk to SteveH and JeenB about auto distinct behavior in implementations [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action07]
 
[PENDING] ACTION: AndyS to add text clarifying the prohibition on blank node labels in multiple BGPs to rq25 [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action03]
[PENDING] ACTION: EricP to run the yacker tool over and annotate the existing tests [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action04]
[PENDING] ACTION: LeeF to remember that the wee, lost filter tests should be put [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action05]
 
[DONE] ACTION: Elias to add wording for PROPOSED: ed(The SPARLQ Protocol does not derefrence query URIs so 5.1.3 does not apply. Per 5.1.4, services must define their own base URI, which may be the service invocation URI.) [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action01]
[DONE] ACTION: Lee to talk to protocol editors re: POSTing application/sparql-query [recorded in http://www.w3.org/2007/02/27-dawg-minutes.html#action02]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2007/02/27 15:38:39 $