See also: IRC log
<DanC> Scribe: DaveB
<_> 26 July agenda
<DanC> 19 July minutes
Next meeting: next week, chair: DanC, scribe: AndyS
actions in section one of the agenda:
<_> ACTION: EricP clarify which regex lang [DONE] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<_> ACTION: PatH to review new optionals definitions, if any. [DONE] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<_> ACTION: DanC to check after August on SteveH regarding test preparation for publication as WG Note [CONTINUED] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<_> ACTION: DanC to follow up re optional test based on op:dateTime triple [CONTINUED] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<_> ACTION: DaveB to to propose source test to approve [CONTINUED] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<_> ACTION: EricP to finish extendedType-eq-pass-result.n3 [CONTINUED] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<_> ACTION: AndyS to add the above graph test cases (analagous to valueTesting test cases) (don't expect quick delivery) ref http://www.w3.org/2001/sw/DataAccess/ftf5-bos.html#item05 [CONTINUED] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
comment from july 2004
<kendall> is ".srf" better?
<DanC> AndyS: seems OK, though I found one conflict with ".srq"
<kendall> where "better" == "no conflicts"
<kendall> I'm a little bit leery of this conflict w/ some sql systems, but it's not a big deal.
<DanC> (I was thinking .sqr )
<db-scr> Souri arrivesi
<db-scr> DanC - add it to the result draft and get reviewers
<DanC> Scribe: EricP
<scribe> ACTION: DaveB to add mime type to results format [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<AndyS> A Google search for filetype:sqr gave no returns
<scribe> ACTION: EliasT to review XML Results mime-type registration [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<scribe> ACTION: DanC to review XML Results mime-type registration [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<DaveB> it's only 134 words...
<JosD> found http://filext.com/detaillist.php?extdetail=SRQ and should mean "Unprocessed Microsoft Server Request"
straw poll: who wants to resolve XML Results mime-type today?
0 want to delay
PROPOSED: that http://www.w3.org/2001/sw/DataAccess/rf1/mime.txt (delegating choice of file extension to the editor) addresses issue resultsMimeType, contingent on review by Elias
RESOLVED, OBJECTIONS: 0, ABSENTIONS: 0
<Zakim> DanC, you wanted to note impact on protocol examples
<DanC> ACTION: KendallC to update protocol spec w.r.t. results mime type [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<DanC> How about changing sparqlResults to
<DaveB> cvs 1.45 of rf1 now has mime type section 5
<kendall> I prefer dash to camel case, but don't care about the month droppage.
DanC: Bjoern: '-' separated
... ... also '-' separated namespaces are more consistent with other W3C namespaces
7-ish prefer '-'
<DanC> PROPOSED: to change the namespace name to http://www.w3.org/2005/sparql-results (and update rf1 and rq23)
<kendall> ACTION: KendallC to Check whether the results namespace is in protocol draft; if so, update. [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<scribe> ACTION: AndyS to fix rq23 to reflect new results namespace [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<scribe> ACTION: EricP to ask for namespace approval and put a document there [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<DanC> ACTION: DaveB to update rf1 with new namespace [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<DanC> resolution is contingent on director's approval
<DaveB> rf1 cvs 1.46 now has the new namespace
EliasT: atom [#-]-less namespace is ugly when concatonated
DanC: i know of no intention to use with RDF
<DanC> (in irc, who prefers http://www.w3.org/2005/sparql-results# ? )
<kendall> i prefer # to bare "sparql-results"
who prefers a hash?: 3 or 4
who's prefers none?: 0
<DanC> PROPOSED: to change the namespace name to http://www.w3.org/2005/sparql-results# (and update rf1 and rq23)
<DanC> RESOLVED, JeenB abstaining
<scribe> ACTION: DaveB integrate advice on ordered and distinct and propose last call candidate for results set spec [DONE] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<_> ACTION: LeeF, SteveH and Jeen to review XML results set format [DONE] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<_> Jeen's review
<_> Lee's review
DanC: DaveB has integrated ordering and limits
LeeF: i *think* all outstanding points have been addressed
DanC: I imagine a test case with an attribute from a randing namespace
DaveB: text says "this element here, and there are these attribtutes there" but the schemas could be normative (striking the caveat "these are informative " sentence)
<DanC> "Normativeness of the XML schemas. Pick one?"
<kendall> (I'm planning to use the RelaxNG one in my web framework to validate XML stored in a db...FWIW)
<LeeF> similarly, we have as of yet uncoded plans that will likely use the XML schema
DanC: who wants to make the XSD schema normative?
<DanC> PROPOSED: to note that the .xsd is derived from the .rng (and therefore, as far as we know, they mean the same thing) and make them both normative
<howardk> daveb: did you get my original email posted to the group?
<DanC> so RESOLVED
DanC: is there a term for "SPARQL Results Format document"?
<kendall> "serialized result set"?
DaveB: could add to section 1 or 2
<DanC> "sparql results document"
<DanC> ACTION: DaveB to choose a term [for "sparql results document"] and define it [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<kendall> I'm happy following Dave's lead here.
<kendall> ACTION: KendallC to use Dave's name for a results set doc in the protocol draft... [recorded in http://www.w3.org/2005/07/26-dawg-irc]
DanC: i can now test an instance with
a schema validator
... i can now test an instance with a schema validator
... yes or no, take the XML Results Format + actioned cahnges to last call
... LC to end at the same time as SPARQL Query
all are happy with it
<DanC> PROPOSED: to take v1.46 + edits per actions today (plus SOTD) to last call.
<DanC> RESOLVED. Action EricP
<DanC> critical path people are: Dave, DanC, EliasT, EricP
<scribe> ACTION: ericP to publish rf1 [recorded in http://www.w3.org/2005/07/26-dawg-irc]
kendall: worried about optimizing for one particular tool (Axis)
AndyS: have eliminated all Axis-isms
<DanC> ACTION: KendallC to add POST binding to protocol doc [recorded in http://www.w3.org/2005/07/26-dawg-irc]
kendall: post binding is number one on my protocol doc list
<DanC> ("'patch' is invoved making the xsd from relaxng..."? hmm.)
<DanC> ACTION: KendallC to consider flattening rdf-dataset a la http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0087.html [recorded in http://www.w3.org/2005/07/26-dawg-irc]
kendall: I added rdf-dataset only to mirror the prose
<EliasT> re: xsd/relaxng patch. Does it matter if both are independently normative?
<kendall> AndyS: Sorry to have not ACK'd publicly on that message. It's useful, I just lost track of it.
<DanC> ACTION: DanC to investigate having CVS commits send to the WG list [CONTINUED] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<scribe> ACTION: DanC to write SOTD; work with EricP to publish [DONE] [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<DanC> SPARQL Query Last Call
<AndyS> Kendall - no problem
DanC: I asked sysreq to do it. they said "done". i haven't seen it work
<kendall> AndyS: yr point 2/ is also on my TODO list already, so ACK that. I'll think about 3/ & 4/
DanC: we now have 4 open issues (from the comments)
<kendall> (My only comment about 5/ is that I wish we had xml serialization of queries so we could POST *that*, but -shrug-)
<AndyS> Err - that's different - I'm asking for a POST HTML form
DanC: editorial comments included ref to ABNF
<DanC> ACTION: DaveB respond to "sparqlResults namespace" comment http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Jul/0039.html , after rq23 is updated [recorded in http://www.w3.org/2005/07/26-dawg-irc]
DanC: we owe Bjoern an "are you happy?" response
<kendall> AndyS: well, yes, gotcha. We have to POST something, and the only thing we can post now, since it's the only thing we have predefined, is application/x-www-form-urlencoded.
<DanC> I'm noodling on distinguishing "are you happy?" responses
DanC later suggested the prefix [OK?], [CLOSED] indicates an ack to an "are you satisfied" ack
DanC: WG members are welcome to act
on any comment by proposing text to the editor via the WG
... some prefix will indicate a request to close
<AndyS> I just want an HTML forms interface!
<kendall> i don't want one. i guess i was suggesting that if we define a way to post stuff, people can do that from HTML forms or from some kind of automated client. i care about the latter, you care about the former. i don't see a problem?
<LeeF> So, application/x-www-form-urlencoded.is fine then, right, AndyS?
<LeeF> I don't think the two of you are disagreeing.
<kendall> i was explicitly agreeing, actually. i guess not explicit enough.
<AndyS> LeeF : yes
<AndyS> Thought you were wanting XML in plain HTTP.
<kendall> nope. (well, yes, i said i do want that, but then I said application/x-www-form-urlencoded was good enough for now. :>)
[DanC summarizes http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0096]
AndyS: I use java.net.uri . i don't think it does normalization.
<DanC> PROPOSED: to add clarify SPARQL QL spec about base IRI normalization and add tests as per http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0096.html
RESOLVED, OBJECTIONS: 0, ABSTENTIONS: 0
<scribe> ACTION: ericP to add "don't normalize" to rq23 (perhaps supplied in 0096) [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<scribe> ACTION: ericP to send [OK?] message to Bjoern [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<DanC> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Jul/0042.html Bjoern Hoehrmann
<scribe> ACTION: EricP to add test in 0096 to rq23 tests. label "approved" and ref this meeting record. [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<kendall> My only concern here is making it possible in the future to have an XML serialization of SPARQL queries. Err, I mean, not making it not possible.
<AndyS> text/ ?
<DanC> mime.txt,v 1.2 2005/07/25 15:41:27
<AndyS> UTF-8 vs UTF-16?
<DaveB> no, not text/
<DaveB> then you *do* get into charset issues
<kendall> (Hmm, I guess my concern is moot, thinking about it some more.)
<kendall> I'd prefer "application/sparql-query" to "application/sparql", I think. More specific, more clear.
<kendall> Also symmetric with "sparql-results"
<LeeF> +1 kendall's suggestion
<howardk> +1 as well
<EliasT> file extension? .sq?
<EliasT> as opposed to ".rq"
-> http://www.w3.org/2001/sw/DataAccess/rq23/mime.txt application/sparql-query mime-type registration
<kendall> ACTION: KendallC to add security considerations section to proto draft, under 4. Policy... [recorded in http://www.w3.org/2005/07/26-dawg-irc]
mime.txt updated, @@'s added
<kendall> Is ".rq" the conventional extension for RDQL queries?
<LeeF> rq = "RDF query" presumably?
<scribe> ACTION: ericP to update rq23 to include the text of rq23/mime.txt reflect security concearns [recorded in http://www.w3.org/2005/07/26-dawg-irc]
<AndyS> We use that in the test suite
<kendall> cool; just wondering
<DaveB> yes all our tests are .rq
<AndyS> It's not in obvious use elsewhere (strangely)
<kendall> i find it a tiny bit confusing, but not a big deal.
<DanC> "SPARQL defines a set of functions & operations (sections 11.1 and 11.2) that all implementations must provide."
<kendall> +1 to DanC's point
<kendall> though I'd slice up the core differently, I think. I don't want to be required to implement all those Fs&Os to be a sparql processor. Am I the only one?
<DanC> DanC: as I wrote earlier, I prefer not to refer to "implementations" at all in the QL spec; I prefer to specify languages, e.g. Core Sparql and extended sparql
11.2.4 Extensible Value Testing
kendall: I don't want to have to implement all (for instance, dates)
<kendall> I *may* just be pointlessly whining here!
kendall: of the functions and
... i'd like the core to be smaller
<DanC> AndyS: if dates are not core, < might not work on them
<AndyS> Integers are harder than dates!
<kendall> but i have clients for integers :>
<howardk> serious andy?
Kendall: dates and strings probably aren't that much work. don't want to make a big deal of it.
<jeen> as a data point: we recently received a code contribution in Sesame for supporting dates. it took the coder about a week to implement (that is, including getting familiar with the sesame code base).
<AndyS> Type promotion needs to be done. 1.5 > 2^^xsd:byte
<DanC> FILTER func:even(?id)
<DanC> 11.2.4 Extensible Value Testing
<AndyS> I used the Xerces code (via Dave Reynolds wrappers) for date comparision, If you have a library, its easy, if not, it is long.
<AndyS> (I was picky and wanted timezone to be preserved - that was my value add - Xerces turns all to Z)
<kendall> or if you have a library and there is semantic mismatch, but -shrug-
<DanC> bug: (section 11.3)
<DanC> " there is an extension mechanism (section 11.3)"
<DanC> (I recommend you use XSLT to make/check your xrefs)
<kendall> it's pretty much 90 minutes, FYI
<DanC> ACTION EricP/Andy: revise rq23 to remove reference to implementations/engines (e.g. 3.3 Value Constraints -- Definition )
<DaveB> the rf1 abstract also uses impl
<DaveB> uhoh, I gotta go rsn
<DaveB> seems .srx isn't used according to filext.com
<DaveB> or file-ext.com
<DanC> (looking at http://www.faqs.org/rfcs/rfc2234.html )