W3C

- DRAFT -

RDF Working Group Teleconference

08 Jun 2011

See also: IRC log

Attendees

Present
+1.540.898.aaaa, Guus_Schreiber, +3539149aabb, SteveH, mischat, davidwood, Peter_Patel-Schneider, cygri, ww, AZ, +1.443.212.aacc, AlexHall
Regrets
Chair
SV_MEETING_CHAIR
Scribe
SteveH

Contents


<trackbot> Date: 08 June 2011

<cygri_> ah well, that was unnecessary

<cygri> i hear that "trackbot, start meeting" magically invites zakim and rrsagent

<mischat> hello all

<cygri> zakim aabb is me

<scribe> scribenick: SteveH

<scribe> scribe: SteveH

Guus: minuites
... any objections....
... resolved, accept minutes of last meeting
... no actions pending review, open action items:
... options for issue 15

cygri: it's related to graphs stuff, we should refactor it
... start progress over again

Guus: it's on an agenda item
... lets close this action, and see
... 3rd action is on Sandro "start conversation on reservings"/

[it might be .well-known]

cygri, it's whether we approach the IEFT now, or wait

Guus: can someone add a note saying what it means

davidwood: I'll add something

<cygri> http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html#section-Graph-Literal

cygri: ...about literals ^

<davidwood> SteveH: Thanks. My recent status change seems to have left me unable to edit the wiki.

Guus: nearing time when europeans will go on holiday
... several ways - we can tke a break, or meet every week with a small group, or do telecons every 2 weeks over summertime
... happy to accept other points

<cygri> trackbot, close ACTION-25

<trackbot> ACTION-25 Write up the different options re ISSUE-15 closed

davidwood: we have one week where we know lots of people will be absent

Guus: does 2 weeks sound fine?

<ww> +1 every two weeks

<cygri> SteveH: sparql keeps running through the summer, lots of americans on the group

<davidwood> +1 to 2 weeks

+1

Guus: suggest we do every 2 weeks, back to normal on 3rd week of aug
... I will propose a schedule

<scribe> ACTION: Guus to propose schedule [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action01]

<trackbot> Created ACTION-55 - Propose schedule [on Guus Schreiber - due 2011-06-15].

ACTION-55: schedule for meetings over the summer that is

<trackbot> ACTION-55 Propose schedule notes added

<PHayes> Um..sorry Im late...why are we changing the schedule?

Guus: SPARQL last call WD

<AlexHall> PatH, because Europeans are about to go on holiday.

<PHayes> Ah.

Guus: decided that we will have personal reviews from members + review on behalf of RDF WD
... actions were not recorded

pchampin: haven't had time to look into it

Guus: it's proper behaviour for us to respond quickly

<scribe> ACTION: pchampin to review SPARQL LC WD document [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action02]

<trackbot> Created ACTION-56 - Review SPARQL LC WD document [on Pierre-Antoine Champin - due 2011-06-15].

<scribe> ACTION: Guus to contact Yves R. re. SPARQL reviews [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action03]

<trackbot> Created ACTION-57 - Contact Yves R. re. SPARQL reviews [on Guus Schreiber - due 2011-06-15].

Guus: Lee F. suggested we organise a short telecon to discuss graph terminology

davidwood: could be in the contxet of coord group

Guus: message of 16th May
... 15th May in US

<scribe> ACTION: Guus to organise telecon with SPARQL WG on graph terminology [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action04]

<trackbot> Created ACTION-58 - Organise telecon with SPARQL WG on graph terminology [on Guus Schreiber - due 2011-06-15].

<pchampin> of course

Guus: pchampin, would be nice if you could take into account discussion of string literals

Status of documentation

Guus: concepts document, it's in mercurial
... I assume that most of the respec problems have been fixed

<cygri> RDF Concepts, editors draft: http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html

Guus: I suggest to reuse old templates

cygri: one way to start would be take a copy of HTML files, especially header
... you have to make some obvious changes
... then insert the current content as published

Guus: I did that already for the primer
... would be best is we started adding docs to repo

<ww> aside: i tried writing a spec with respec.js attempting to put the vocabulary in rdfa inside it. didn't work very well...

<PHayes> I have to say, this whole process is utterly alien to me and I really have not even begun hjow to install the necessary software. As I have no idea what it is doing, I dont know how to know if I ge it right.

PHayes: I'll learn how to do it, but it will take me a while

Guus: Richard sent a doc with shortnames for docuemnts, seems obvious
... but why is it turtle, not rdf-turtle

cygri: either would be ok

Guus: we have rdf- infront of all of them

davidwood: I propose to make that change

<AZ> \me +1 to rdf- for all documents

cygri: we should have a page (on the wiki) about the documents
... I could create

<scribe> ACTION: cygri to create page on wiki about documents and editing [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action05]

<trackbot> Created ACTION-59 - Create page on wiki about documents and editing [on Richard Cyganiak - due 2011-06-15].

Guus: will try for early Turtle draft, relatively little work, but work needs to be done

ACTION Guus to discuss Turtle doc schedule with ericP

<trackbot> Created ACTION-60 - Discuss Turtle doc schedule with ericP [on Guus Schreiber - due 2011-06-15].

Guus: will attempt to report back next week

<PHayes> +1 to david

davidwood: can we leave the telecon slot open

[general agreement]

Guus: content issues
... as far as I can see the main changes to concepts are graphs TF issues, have to reach consensus, but lots of open issues
... wondering if its useful to do review next week
... is someone willing to prepare that discussion

<davidwood> +1 to refocus discussion on graphs

Guus: about 10 issues open, propose we start discussing next week
... re. concepts doc

<PHayes> Unfortunately this particular week is impossible for me, or I would volunteer. Good idea.

<pchampin> +1

Guus: issues are well documented, so should review issues, and assign actions

<cygri> +1 to reviewing the issues + deciding on actions

Guus: one issue is being tacked by telecon
... we were close to consensus in last discussion
... next week 30 mins minimum for review of status of graphs issues
... last 5 postponed issues
... ISSUE-58
... david proposes we should close it as it's archaic

<pfps> +1

+1, close it

<PHayes> agreee close

<pchampin> +1

<AlexHall> +1

<AZ> +1 close

<cygri> +1 close

RESOLUTION: close ISSUE-58

by consensus

ISSUE-59

Guus: "RDF XML syntax can't represent arbitrary graphs"

<pfps> +10 to not upgrade RDF/XML to do this

+1 to close

<pfps> +1 to *close*

<AZ> +1 to close

<AlexHall> It's already noted in the specs

<AlexHall> +1 to close

<PHayes> propose we leave this open for now, until we consider rdf/xml. No need to close it.

<davidwood> +1 to close

<cygri> ISSUE-59?

<trackbot> ISSUE-59 -- Revisit "The RDF/XML syntax can't represent an an arbritary graph structure" -- raised

<trackbot> http://www.w3.org/2011/rdf-wg/track/issues/59

<pfps> I don't see a possible future in which RDF/XML changes to represent all graphs.

+1

<PHayes> OK, 0 from me.

<MacTed> silly phone system....

<pchampin> g-

<PHayes> richard has a good point. THis may be a non-issue due to an old clerical error.

<cygri> issue description here: http://www.w3.org/2000/03/rdf-tracking/#rdfms-syntax-incomplete

<scribe> ACTION: pfps to check whether ISSUES-59 is still pertinient (may be obsolete) [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action06]

<trackbot> Created ACTION-61 - Check whether ISSUES-59 is still pertinient (may be obsolete) [on Peter Patel-Schneider - due 2011-06-15].

<ww> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jul/0163.html <-- actions to add nodeID recorded

ISSUE-60?

<trackbot> ISSUE-60 -- Revisit "Defining the interpretation of fragment identifiers in RDF embedded in other document formats" -- raised

<trackbot> http://www.w3.org/2011/rdf-wg/track/issues/60

Guus: propose to continue

<cygri> ISSUE-37?

<trackbot> ISSUE-37 -- Handling of fragment identifiers in RDF embedded in other document formats -- raised

<trackbot> http://www.w3.org/2011/rdf-wg/track/issues/37

cygri: things like RDFa make this question more important, so this should be considered

Guus: we have an issue already, so we can close 60, redir to 37
... can someone add a link to 37, and close 60?

<pfps> I'll do it, instead of my other action.

<cygri> +1 to close and redirect to ISSUE-37

<PHayes> FWIW, re. issue 59, the 26 july 2000 wg minutes say that this issue is "removed from the WG's issue list", not "postponed".

RESOLVED by consensus to close ISSUE-60 and redirect to ISSUE-37

<pfps> go it

<pfps> got it

<PHayes> OK

Guus: looks like ISSUE-59 was an admin error

ISSUE-61?

<trackbot> ISSUE-61 -- Revisit "An XML literal without markup, e.g. "foo" should denote the same thing as the plain literal "foo"" -- raised

<trackbot> http://www.w3.org/2011/rdf-wg/track/issues/61

<pfps> I don't think that Issue-12 talks about XML literals now.

I don't believe that "foo" is a legal XMLLiteral, is it?

"<foo />" is legal, I think

<PHayes> +1 Steve.

Guus: prefer not to close it with the current text, quite sure that we will close it with the statement that it's misguided

<PHayes> I dont understand this issue? Was it to make "foo"^^^rdf:XMLLIteral be identical with something else? If so, what?

<AZ> Close it but do not mention Issue-12

<PHayes> +1

Guus: propose to close the issue stating that the statement is not true

<AZ> +1

<davidwood> +1

<AZ> "foo"^^^rdf:XMLLIteral owl:differentFrom "foo"

Guus: more discussion?

propose to close issue-61 stating that the answer should be no

<pchampin> +1

+1

<AZ> +1 use pfps proposal

<pfps> +1

<davidwood> +1

RESOLUTION: close ISSUE-61 stating that the answer is "no"

<PHayes> Why do I keep thinking 'augean'?

ISSUE-62?

<trackbot> ISSUE-62 -- Revisit "The test cases manifest format has a semantic error" -- raised

<trackbot> http://www.w3.org/2011/rdf-wg/track/issues/62

Guus: propose to continue this issue, and look again when we're working on testcases
... leave it open

<PHayes> +1 to doing nothing.

+1

<AZ> +1

ISSUE-12?

<trackbot> ISSUE-12 -- Reconcile various forms of string literals (time permitting) -- open

<trackbot> http://www.w3.org/2011/rdf-wg/track/issues/12

Guus: there's a thread about this, it appears we're close to consensus
... would like discussion about things we don't have consensus about
... plan for resolution next week

<PHayes> I think we need a better name for rdf:LanguageTaggedLIteral

PHayes: I think we're close to consensus, what about alt. proposal about using datatypes
... about only remaining thing we're still debating is what we're calling this datatype, and how best to explain it so it doesn't sound complicated

Guus: it's important to spend time on naming

PHayes: there's on more issue, there's 2 ways to present it, the new datatype

<ww> so "chat"@fr -> "\"chat\"@fr"^^rdf:LTR ??

PHayes: one is to retain the current model sctrictly, but we have to use PlainLiteral device in abstract syntax, to include both parts in one string, easy, but ugly(?)
... or would could bite the bullet and treat it like its a datatype, but it takes a pair, the extension is trivial, some people things it's complex, but I don't agree

cygri: there are several ways to handle connection between abs. syntax and semantics, one was is to leave PLs as they are, and say that rdf:LTS is not actually a DT, but a class
... of all <string, langtag> pairs
... worht considering, or do people object

PHayes: that's a viable option too
... if you look how they get used, it's only used as a classname
... or a token
... we can just say that (something) without saying it's a DT

cygri: it seems that it's the least painful way
... it might be a bit cleaner to do it with a DT, but we still have three things
... seems to cause oposition from implementors

+1

scribe: maybe it would be a good option

<Guus> +1 for Richard's option

<PHayes> As long as it can be treated as a 'type' in SPARQL :-)

scribe: two more things, may still be disagreement, there's ....

ww: this proposal seems like a half measure, intriduces an extra 3rd thing, and cant use rdf machiney to model langs, which we might want to do
... we should leave open the possibility

<PHayes> Which proposal is Ww referring to?

ww: having a tuple-space with datatype means we cant do that

cygri: i'm confused

ww: dt with string,lang pairs means the lang is disconnected
... there should be a DT for LTS, with subtypes, for every language
... leave the door open for modelling that
... abolish langtags

<PHayes> q

<PHayes> +q

cygri: can't follow that
... what is the proposal

ww: langtags abolished, strings are strngs, subsets of the sets of all strings that are strings in particular languages, subtypes of the string datatype

<PHayes> I dont htink that our users will tolerate our bainishing lang tags on literals.

<pchampin> that does not work; language are orthogonal to strings

PHayes: there is a sizeable user population that demanded them with passion, can't get rid of them
... inc. the 23c i18n group

ww: not saying remove the function, just make it a kind of DT

[sounds like ww is describing langtags as datatypes option]

<ww> SteveH: yes

ww: get rid of langtags yes, but map them to dts(?)
... want to make languages a tree of datatypes

PHayes: any proposal that removes langtags from syntax of RDF wont''t fly

+1

ww: will write proposal to list

cygri: some discussion is needed re. preference of different contrcete syntax forms
... e.g. in NTriples would now have two options, "foo", "foo"^^xsd:string
... there are different tradeoffs in different formats
... in NTriples is good that there's not much syntax variation
... would make things harder if I find both in the wild
... should we say that one SHOULD, MUST or SHOULD NOT use one of these forms
... or allow each spec to do it's own thing
... I think I disagree with AndyS about some format issues

[what]s AndyS's position?}

<PHayes> +1 to getting all this VERY CLEAR, for sure.

Guus: shortest form is usally preferable

+1 to VERY CLEAR

cygri: AndyS says that authors SHOULD use the shortest form, in SPARQL results I would really like to be able to know whether the strings are going to have the DT or not

+1 to cygri

<PHayes> Richard, you are shooting Andy in the foot here.

cygri: so a SHOULD isn't strong enough to me
... in turtle I don't see the need
... would like to see a stronger statement

<PHayes> Everyone wants the query language to be both semantically transparent and also sensitive to the smallest lexical detail. Cant have it both ways, guys.

pchampin: I agree for need for regularity, but maybe there are differences
... NTriples I see 3 alternatives
... allow both, bad idea
... keep shortest one, best of three, but some iregulariy, string literals must be treated in special way

<PHayes> How much legacy RDF is there out there that uses one and not the other? Do we ahve a choic eot not allow both?

<pfps> I'm feeling very weirded-out by all this SPARQL stuff. RDF is supposed to be about *meaning*, not syntax, not even abstract syntax!

pchampin: enforce xsd:string, but breaks a lot of existing NTriples
... for the sake of back-compat we have to keep shortest form

<PHayes> Yes, pfps, but querying is all about syntactic matching. You betcha.

Guus: users typically use the shortest form, but sparql query uses the DT form

<pfps> Well, not as far as I am concerned. Querying is about retrieving meaning. (As opposed to straight entailment, which is simpler.)

I would be -1 to SPARQL using the long form, that's a lot of bytes

pchampin: I would be in favour of MUST for NTriples and SPARQL res, but not others, but not sure which form is best
... both would break some existing data, most reg. form is with the datatype

<PHayes> pfps, so listen to Richard. He wants to make queries which distinguish a from b when a = b is *necessary*. Any why not? Hos code has to handle the suyntax, not the meaning.

pchampin: explicit is better than implicit
... there are a lot of plain literals out there

<ww> less typing, more clarity and consistency - make developers lives easy as possible.

cygri: for back compat we have to keep both forms valid

<ww> +1 cygri

cygri: we cant say that any forms would now be invalid in NTriples

+1

<Guus> +1 for not using MUST

scribe: when parsing both forms are valid, but when serialising, only use one form,

<pchampin> +1, enforced regularity would break backward compatibility

<pchampin> +1 about distinguishing old stuff/new stuff

PHayes: I agree with Richard, there's so much stuff out there, can't make it illigal
... one meaning can be expressed two different ways, the tool should treat them as equivalent

<Guus> 2 min left

PHayes: is results sensitive to the way the query is stated

<ww> +1 for tools treating them equivalently (and probably normalising them to w/ datatype internally)

<ww> +1 for fewer bytes on the wire

<PHayes> +1 to SteveH.

SteveH: the long form is less efficient
... even though it's easier to canonicalise to

<pchampin> well, you are trading bandwith for (slight) code complexity

<pchampin> you're just moving the inefficiency somewhere else :)

Guus: is someone willing to look at cygri's proposal

<PHayes> LOL

<ww> pchampin: not only bandwidth - developers who look at it don't want to see extraneous cruft

Guus: next week re restart graphs discussion
... hopefully can set a dte for graphs naming discussions

<cygri> ww, pchampin: by same argument, N-Triples should write 6 for "6"^^xsd:decimal

<AZ> bye

<pchampin> +1 cygri :)

<pchampin> bye all

<ww> cygri: that wouldn't be the end of the world, but agree there is a slippery slope

trackbot, end meeting

<davidwood> trackbot, end meeting

Summary of Action Items

[NEW] ACTION: cygri to create page on wiki about documents and editing [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action05]
[NEW] ACTION: Guus to contact Yves R. re. SPARQL reviews [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action03]
[NEW] ACTION: Guus to organise telecon with SPARQL WG on graph terminology [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action04]
[NEW] ACTION: Guus to propose schedule [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action01]
[NEW] ACTION: pchampin to review SPARQL LC WD document [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action02]
[NEW] ACTION: pfps to check whether ISSUES-59 is still pertinient (may be obsolete) [recorded in http://www.w3.org/2011/06/08-rdf-wg-minutes.html#action06]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2011/06/08 16:19:28 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.136  of Date: 2011/05/12 12:01:43  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/would/could/
Succeeded: s/string enough/strong enough/
Found ScribeNick: SteveH
Found Scribe: SteveH
Inferring ScribeNick: SteveH

WARNING: No "Topic:" lines found.


WARNING: Replacing list of attendees.
Old list: +1.540.898.aaaa Guus_Schreiber +3539149aabb SteveH mischat davidwood Peter_Patel-Schneider cygri ww AZ +1.443.212.aacc AlexHall [Sophia] pchampin PatH MacTed
New list: +1.540.898.aaaa Guus_Schreiber +3539149aabb SteveH mischat davidwood Peter_Patel-Schneider cygri ww AZ +1.443.212.aacc AlexHall

Default Present: +1.540.898.aaaa, Guus_Schreiber, +3539149aabb, SteveH, mischat, davidwood, Peter_Patel-Schneider, cygri, ww, AZ, +1.443.212.aacc, AlexHall
Present: +1.540.898.aaaa Guus_Schreiber +3539149aabb SteveH mischat davidwood Peter_Patel-Schneider cygri ww AZ +1.443.212.aacc AlexHall

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Found Date: 08 Jun 2011
Guessing minutes URL: http://www.w3.org/2011/06/08-rdf-wg-minutes.html
People with action items: cygri guus pchampin pfps

WARNING: No "Topic: ..." lines found!  
Resulting HTML may have an empty (invalid) <ol>...</ol>.

Explanation: "Topic: ..." lines are used to indicate the start of 
new discussion topics or agenda items, such as:
<dbooth> Topic: Review of Amy's report


[End of scribe.perl diagnostic output]