See also: IRC log, previous weeks's minutes
next meeting: 2 Jan
next scribe: SimonR
<scribe> ACTION: KendallC to close formsOfDistinct issue [CONTINUED] [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action01]
<scribe> ACTION: KendallC to remember that the wee, lost filter tests should be put [CONTINUED] [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action02]
<scribe> ACTION: PatH to change the entailment section around to talk about SPARQL first, then more general conditions in a normative appendix [CONTINUED] [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action03]
<scribe> ACTION: ericP to seek clarification on http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Nov/0004 [CONTINUED] [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action04]
<LeeF> Jeen on the test suite process: http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0205.html
<scribe> ACTION: EricP to sort out some string literal thing for the operator table [ed: this action is about text for simple literals and xsd:string and text for RDFterm-equal] [CONTINUED] [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action05]
<scribe> ACTION: Jeen propose test suite process (not do it all). [DONE] [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action06]
http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0205 -> Jeen's proposed test suite process
<scribe> ACTION: LeeF to review rq24-algebra [DONE] [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action07]
http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0206 -> LeeF's review of rq24 algebra
LeeF: we have lots of unapproved
tests that are not distinguished from those next to outdated
ones
... we discussed CR exit criteria where we had at least two
implementations that implemented each SPARQL "feature"
... assuming still need this, we need to record up front what
the set of "features" is
... then assign tests some set of features
SimonR: what do when we have more than one right answer
ericP: One way to handle it is to avoid tests with more than one answer
SimonR: what about a graph with inference capabilities
LeeF: those tests fall under either SPARQL extensions, or the closed graph is known
<ericP> +1
<sdas2> +1
AndyS: want test core SPARQL
first things in SPARQL. not focus on extension points
... would like to put your [ericP's] classified extended SPARQL
queries in the back of the queue
... don't want extra-core features on the critical path
LeeF: if i were doing the tests, i would prefer explicit classification
<SimonR> I'm not complaining about focusing on non-extended SPARQL or only tests which permit only a single correct answer. I'm just checking that we're doing this deliberately and consciously.
<LeeF> Very deliberately :-)
ericP: i want to make sure that
test be able to have more than one feature
... propose that we annotate tests with features
SimonR: we already have URIs for features from the document section headings
ericP: reasonable thesis, but needs testing
LeeF: reluctant to make such a decision without Jeen
LeeF: re protocol test suite, Elias says "no worries"
mince!
<LeeF> sorry :)
i heard lots of hiss, then got booted. assumed hiss must of been my leave music
rq24 algebra
Lee's rq24 algebra review
LeeF: main points:
... 2 snaps up
... need connections to grammar
LeeF: soln modifiers could be
more explicitly tied to concrete syntax
... as a reviewer, endorse adoption into rq24
AndyS: Jeen said he may find time for comments, but no commitment
Souri: hope to review by Jan 9
ACTION: Souri to review http://www.w3.org/2001/sw/DataAccess/rq23/rq24-algebra by 9 Jan 2007 [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action08]
LeeF: any words from Pat re rq24-algebra?
AndyS: said it looked OK a couple weeks ago.
LeeF: i know of no major issue
stumbling block decisions
... need:
... Souri's review
... PatH's entailment text
... decision on algebra around mid Jan
... merge into rq24
... need end to end reviews from WG
... would like LC WD decision around 1st week of Feb
... want a short-as-possible LC, then CR
... then back to CR
... then test suite work and exit criteria
ACTION: LeeF to pester PatH about entailment text [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action09]
LeeF: will review outstand issues
and hope that the new rq24 will address those issues
... for example, FredZ's issues from the end of the
summer
... will ask Elias to do the same with the protocol doc
... 2 months ago there were 3 small issues which raised no
pulses during a WG meeting
... would like all (3) to go to LC/CR at the same time
... any thoughts on issues that i haven't address?
<AndyS> HP supports that outline plan.
LeeF: speak now or forever hold your peace^H^H^H^H^H^H^H still welcome to raise in the future
<Zakim> AndyS, you wanted to ask if that makes LC content assumed to be CR content?
AndyS: you expect the LC and the CR are as close as possible?
LeeF: yes, as close to byte-wise as the comments will allow
AndyS: summary above indicates the test suite work is serial. can do that stuff in parallel
EARL
pretty EARL pics from an Algae run
ericP: I'd like to suggest we use EARL as a reporting vocabulary for the results of implementation test runs
SimonR: What about the XUnit [ed: is that right?] XML format? There are a lot of existing tools that we could use for that.
SimonR: (I think EARL would be cool) speaking as devil's advocate, what about XUnit?
AndyS: are they solving the same problem?
SimonR: report results of runs of tests
NUnit for .net, JUnit for java
<LeeF> DUnit for DAWG!
<LeeF> SPUnit for SPARQL!
<LeeF> ok, sorry, i'm done.
ACTION: SimonR to look at .Unit [recorded in http://www.w3.org/2006/12/19-dawg-minutes.html#action10]
<ericP> .+Unit?
ADJOURN
<AndyS> Merry Christmas and a Happy New Year to everyone from the UK