See also: IRC log
<scribe> Scribe: EliasT
<LeeF> So picky.
<AndyS> EricP - extracted examples - any updates? I hope to make the corrections I noted in email
<AndyS> many of the examples aren't checked though
<AndyS> if there is an update soon I coudl do all fixes at once
<Sven_Groppe> JacekK is me
<LeeF> agenda is ok
<AndyS> Jeen -- sound check?
<DanC> minutes 11 apr
<LeeF> oo
<LeeF> has it always been that polite?
<DanC> RESOLVED to approve minutes 11 Apr
<DanC> I'm in the mood to meet next 30 May
jeen: Do we need a full meeting to work on the test suite?
DanC: Well, to approve tests you need a WG decision.
ericP: Looking for items that need WG decisions in agenda
DanC: valueTesting seems to be needing more WG work.
<AndyS> AFS notes that fn => http://www.w3.org/2004/07/xpath-functions
DanC: if we don't decide when to meet next, we need a 7-day notice before we can meet.
<DanC> PORPOSED: to meet next 2 May, EricP or KC to chair.
<DanC> JB: regrets 2 May
<libby> I can't make it either
<libby> 9 shoudl be ok
<DanC> PORPOSED: to meet next 9 May, EricP or KC to chair.
<LeeF> probably OK
<AndyS> Even http://www.w3.org/2005/xpath-functions is wrong because it seems to be /# for .../xpath-functions/#not
<LeeF> i can if i'm there, but i'm a bit at risk
<LeeF> say, 70% chance of it being OK
<LeeF> :-p
<DanC> PORPOSED: to meet next 9 May, EricP or KC to chair, Lee or Jeen to scribe
<DanC> RESOLVED: to meet next 9 May, EricP or KC to chair, Lee or Jeen to scribe.
PROPOSED: to meet next 9 May, EricP or KC to chair, Lee or Jeen to scribe
<DanC> ACTION: EricP to publish http://www.w3.org/2001/sw/DataAccess/json-sparql/ [CONTINUES] [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action01]
<DanC> comments to sparql-dev. Note
<scribe> ACTION: EricP to change status to reflect that comments should go to sparql-dev [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action02]
<DanC> ACTION: DanC find decision record for bnodes in predicate [CONTINUES] [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action03]
<DanC> Geoff's comment
http://lists.w3.org/Archives/Public/public-rdf-dawg/2006AprJun/0047.html
<DanC> Equality in FILTER [#ValueTesting] Seaborne, Andy (Tuesday, 18 April)
<DanC> (boolean isn't a subtype of integer, is it?)
<DanC> AFS: XQ F&O says False < True
Souri: Does XPath say that?
<AndyS> http://www.w3.org/TR/xpath-functions/#func-boolean-less-than
<AndyS> Hmm - no boolean-compare which is odd
<DanC> AFS: having < apply to booleans helps with sorting
<AndyS> I'd be happiest with a general what does "=" mean principle
DanC: would prefer to make WG
decision based on tests
... are replies to Geoff should be backed by tests.
<AndyS> Geoff's comment is http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Apr/0003.html (not 0002)
<ericP> ACTION: EricP to propose text and tests to add {boolean < > = != boolean} to SPARQL [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action04]
<DanC> ah... indeed, 0002 is the wrong one
<AndyS> As we discussed last week, I have acked him offlist that I fwd'ed it to the WG list
<Sven_Groppe> must leave because of Kindergarten opening times...
<DanC> enjoy, Sven
DanC: can you look for roman numeral test?
<AndyS> DanC: it should be that no processor says "2 != II^^:roman"
<SteveH> +! to what danc is saying
<SteveH> +1 even
<SteveH> (= sometimes being equivalence and sometimes value compare?)
<ericP> FILTER ( "XI"^^xsd:romanNumeral = "two"^^xsd:numericText ) in a naive processor falls through to the {RDFTerm = RDFTerm} test and = FALSE
<AndyS> What about ?p = <foo> ??
<ericP> when the processor understands both xsd:romanNumeral and xsd:numericText, we get TRUE
DanC: it would be interesting to see what users think getting type errors when we have unknown terms.
<DanC> (er... integer and decimal do intersect)
<ericP> FILTER ( ?p sameAs <foo> )
ericP: if we have error propagation then tests would only change from error to TRUE when implementation learns about new types. (since all of our xsd types don't intersect)
<ericP> FILTER ( ?p sameTermAs <foo> )
<ericP> FILTER ( !(?p sameTermAs <foo>) )
<DanC> (cwm calls it log:equalTo, I think, which is also goofy.)
<ericP> maybe FILTER ( !(?p ¥ <foo>) ) ?
ericP: the common usage is to be it numeric comparison and same term comparison, however the syntax gives us a non-monotonic behavior.
AndyS: it could be monotonic if
we throw an error.
... if it doesn't have a clue, throws an error.
<ericP> FILTER ( "XI"^^xsd:romanNumeral = "two"^^xsd:numericText )
ericP: how do we differentiate between numeric and term compare?
<ericP> (?x = "XI"^^xsd:romanNumeral) (?y = "two"^^xsd:numericText)
DanC: you put it in the BGP
<ericP> FILTER ( ?x = ?y )
ericP: sorry for being dense
<AndyS> s/XI/II/ ??
<DanC> (?x = "II"^^xsd:romanNumeral) (?y = "two"^^xsd:numericText)
DanC: it should be an error.
<ericP> FILTER ( ?x = <foo> )
<DanC> (should as in: that's what AFS is proposing)
AndyS: it should return an error.
<AndyS> lol
<AndyS> Can do same-rdf-term by (1) BGP or (2) FILTER with str/datatype/lang (a bit verbose but uncommon?)
ericP: I believe AndyS proposal to remove RDFTerm equal when the two are typed literals.
<DanC> AFS isn't able to confirm based on the way that got recorded, but such is life
<AndyS> EricP's or mine?
<LeeF> All this has no bearing on error propagation, right?
AndyS: yours.
<AndyS> The discussion email is: http://lists.w3.org/Archives/Public/public-rdf-dawg/2006AprJun/0047.html
<AndyS> Summary "=" on literals means sameValueAs
<DanC> ACTION: EricP to draft text for open-world = [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action05]
<DanC> ACTION: Andy to draft of open-world = tests [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action06]
<DanC> ACTION: DanC to follow up re optional test based on op:dateTime triple [CONTINUES] [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action07]
<DanC> ACTION: DanC to follow up re optional test based on op:dateTime triple [WITHDRAWN] [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action08]
<DanC> ACTION: AndyS to contribute a couple tests for computed-properties/virtual-prop/builtin. one usingn lists, and one not using lists [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action09]
<DanC> ACTION: EricP and Jeen to fix OPTIONAL coding [DONE] [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action10]
<DanC> jeen's work is noted in http://lists.w3.org/Archives/Public/public-rdf-dawg/2006AprJun/0020.html
<jeen> http://www.w3.org/2001/sw/DataAccess/tests/data/extracted-examples/extract
<ericP> http://www.w3.org/2001/sw/DataAccess/tests/data/extracted-examples/extract
<DanC> ACTION: EricP to respond to Dirk on MIME type stuff [CONTINUES] [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action11]
<DanC> ACTION: EricP to follow up on mime type requests (2Mar) [CONTINUES] [recorded in http://www.w3.org/2006/04/18-dawg-minutes.html#action12]
<DanC> ADJOURN.
<SteveH> bye