<trackbot> Date: 14 March 2012

<pchampin> stuck in my previous meeting, can not dial in for the moment, sorry

<danbri> regrets from me; on a trip

i'll scribe.

<scribe> scribe: alexhall

Admin

guus: PROPOSED to accept the minutes of the 7 Mar telecon

RESOLVED to accept the minutes of the 7 Mar telecon

Action Item review

<sandro> http://www.w3.org/2011/rdf-wg/track/actions/open

sandro: nobody here has open action item
... (aside from guus and sandro, and these aren't done).

guus: reminder, 1 hour earlier for europeans next week, then back to regular schedule

Turtle

<ericP> ISSUE-67: \xxxx escaping in prefixed names

<trackbot> ISSUE-67 \xxxx escaping in prefixed names notes added

guus: would like to close issue 67 before moving on

<gavinc> http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#sec-escapes

eric: text has been committed to hg
... believe this text addresses that by saying \uxxxx is a numeric escape, and that these are allowed in IRIs but not in other places

<AndyS> Strings?

guus: add link to minutes where we discussed this in the issue notes, close the issue

gavin: we do allow numeric escapes in strings, this disagrees with the text

eric: history is that richard didn't want escapes in prefix names, claimed it needlessly complicated things

<AndyS> SPARQL allows \u in strings (and local parts for that matter). Must keep strings.

eric: thought we agreed to eliminate escapes from prefix names to align with SPARQL

andy: unicode numeric escapes allowed in IRIs, strings, local part of prefix names in SPARQL
... they are needed in IRIs and strings to account for N-Triples

<gavinc> AlexHall: I thought the original disagreement was to allow numeric escapes in prefix names. I thought we agreed not to do that, but allow the reserved char escapes.

<AndyS> As of today's grammar update \u is not permitted in SPARQL local names.

<gavinc> AlexHall: I thought we'd agreed that numeric escapes should not be allowed in local names

andy: preference is to allow numeric escapes in IRIs and strings for SPARQL

<gavinc> PROPOSAL: http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#sec-escapes should be reflected in the grammar

sandro: are qnames/prefixed names allowed in n-triples?

gavin: no, they are not, but part of the grammar is in there as bnode labels
... use cases for having numeric escapes in IRI/string literals is to support people with ASCII-only editors
... and to support existing documents to be used unchanged

<sandro> I'm surprised to see \\ is not a reserved character escape sequence.

<PatH> Apologies for lateness, dialling in ASAP.

alexhall: the turtle grammar does not match the sparql grammar for prefixed names

<sandro> PatH, no worries, we just talking about character escaping in Turtle at the moment.

<PatH> Oh then I might just go back to bed...

<sandro> +0 either way is fine with me

guus: think we've talked enough about this, can we go ahead and make a proposal?

<gavinc> PROPOSAL: http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#sec-escapes should be reflected in the grammar. Eg, remove \u escaping from Prefixed Local Names

<sandro> PatH, I think we're almost done, the'll we'll be ready to argue with you.

<sandro> +0 either way is fine with me

<ericP> +0 want consensus

<gavinc> +1 for text agreeing with grammar

RESOLVED01: http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#sec-escapes should be reflected in the grammar. Eg, remove \u escaping from Prefixed Local Name

<gavinc> ACTION gavinc to fail to update Turtle/N-Triples grammar to not use \u escaping in the local part of prefix names

<trackbot> Sorry, couldn't find user - gavinc

<gavinc> ACTION Gavin to fail to update Turtle/N-Triples grammar to not use \u escaping in the local part of prefix names

<trackbot> Created ACTION-150 - Fail to update Turtle/N-Triples grammar to not use \u escaping in the local part of prefix names [on Gavin Carothers - due 2012-03-21].

guus: moving on to issue 74, that seemed to be mostly editorial?

eric: yes, andy gavin and i talked and came to consensus

<AndyS> I changed SPARQL draft today. (Please check.)

<gavinc> ACTION Gavin to update issue 74 with new escaping rules

<trackbot> Created ACTION-151 - Update issue 74 with new escaping rules [on Gavin Carothers - due 2012-03-21].

guus: next issue...

<gavinc> ACTION Gavin Create new issue for :'s in the local part of prefix names

<trackbot> Created ACTION-152 - Create new issue for :'s in the local part of prefix names [on Gavin Carothers - due 2012-03-21].

<sandro> ericP, the link you want now works: http://www.w3.org/2011/rdf-wg/meeting/2012-03-14#resolution_1

guus: revisiting issue from last week (n-triples)
... andy proposed a way forward last week

gavin: most of this proposal is reflected in the current editors draft, so i like it

<gavinc> http://dvcs.w3.org/hg/rdf/raw-file/c43ac715dde2/rdf-turtle/index.html#sec-triples

guus: would like comment from the oracle people. have they had a chance to review it?

zwu2: have been busy with customers, no chance to review it

<scribe> ACTION: zhe to review Andy's email from last week re N-Triples and comment on it. [recorded in http://www.w3.org/2012/03/14-rdf-wg-minutes.html#action01]

<trackbot> Created ACTION-153 - Review Andy's email from last week re N-Triples and comment on it. [on Zhe Wu - due 2012-03-21].

<zwu2> thanks guus

guus: let's look at planning for Turtle LC

<pchampin> what's the deadline?

guus: andy indicated willingness to review Turtle draft for last call. any other volunteers?
... editors, what is a reasonable deadline for reviewers?

eric: Trying to think of changes. There's removal of \u escapes from prefix names

gavin: change to reference ourselves for n-triples definition
... change grammar to reflect text per today's resolution
... editorial changes that Andy collected in an email
... do we need issues for all of these?

eric: think we can edit as needed to address these items before review, once it's reviewed we need wg consensus

<PatH> what?

gavin: another issue is case-insensitivity of boolean true/false

<PatH> ah.

<gavinc> ACTION Gavin Figure out what happened with case-independent true/false

<trackbot> Created ACTION-154 - Figure out what happened with case-independent true/false [on Gavin Carothers - due 2012-03-21].

guus: don't see this noted anywhere, need to track it

gavin: can somebody confirm that we've covered all feedback from comments mailing list?

guus: i think we have tracked all that, we can go back and verify in parallel with group review

gavin: what are we writing changes relative to?

guus: since no previous standard, use original wg note as a reference point

eric: did we decide that true/false are case-insensitive?

gavin: sparql says that they are case insensitive, but grammar isn't

<PatH> sound of head exploding?

<ericP> sound of infinite recursion?

<PatH> lol

guus: i'm happy with anything you guys agree on

<AndyS> skype bug is attempting to take over all my machine

gavin: trying to agree on grammar keyword, but in either case they will map to canonical (lower-case) "true" or "false" for the actual literal value

eric: correct

guus: is 2 weeks realistic for a review?

gavin: 2 weeks should be ok

guus: have draft available in 2 weeks, then give reviewers another 2 weeks to respond after that
... sound ok to editors?

<ericP> AndyS1, would you prefer that the keywords were case-sensitive or insensitive?

<pchampin> ok, count me in as a reviewer

guus: let's make this our tentative schedule. mar 28 draft available, discuss reviews at apr 11 telecon
... andy and pierre-antoine, would be best if reviews available by apr 9

<AndyS> Not sure it's a big deal - compatibility for SPARQL is insensitive but not sure it happens in the wild much (indeed even using true/false at all)

<scribe> ACTION: gavin to prepare Turtle LC draft by March 28 [recorded in http://www.w3.org/2012/03/14-rdf-wg-minutes.html#action02]

<trackbot> Created ACTION-155 - Prepare Turtle LC draft by March 28 [on Gavin Carothers - due 2012-03-21].

<scribe> ACTION: eric to prepare Turtle LC draft by March 28 [recorded in http://www.w3.org/2012/03/14-rdf-wg-minutes.html#action03]

<trackbot> Created ACTION-156 - Prepare Turtle LC draft by March 28 [on Eric Prud'hommeaux - due 2012-03-21].

<scribe> ACTION: andy to review Turtle LC draft by April 9 [recorded in http://www.w3.org/2012/03/14-rdf-wg-minutes.html#action04]

<trackbot> Created ACTION-157 - Review Turtle LC draft by April 9 [on Andy Seaborne - due 2012-03-21].

01ACTION: pchampin to review Turtle LC draft by April 9

scribe: discuss testing approach in the coming weeks
... hope to have document published by end of april

<scribe> ACTION: pchampin to review Turtle LC draft by April 9 [recorded in http://www.w3.org/2012/03/14-rdf-wg-minutes.html#action05]

<trackbot> Created ACTION-158 - Review Turtle LC draft by April 9 [on Pierre-Antoine Champin - due 2012-03-21].

guus: SPARQL wg coordination -- think we're ok on this

named graphs semantics

<PatH> im getting bad sound here.

guus: pat sent out note on contexts, can we take a few minutes to ask clarification questions?

<PatH> sound ok now

guus: in the 3 cases, particularly case 2 where dataset has time of context
... my reply was that this is orthogonal to case 3. is this correct?

pat: they can't coexist casually. if you think of the 4th column IRI as a label in a graph store, in case 3 the IRI denotes the graph, in case 2 the IRI denotes the context

<PatH> i forgive you already

guus: other questions about this?
... if not, would like to give pat the floor to discuss latest message

<AZ> Pat's last long email: http://lists.w3.org/Archives/Public/public-rdf-wg/2012Mar/0086.html

pat: the latest email was a sort of epiphany about this whole thing -- whether the scope of IRI labels should be local to a graph or global to the web, etc.
... both sides seem blindingly obvious to their proponents
... think of RDF as have always been a context logic all along, where the context was implicit
... i.e. there was a global RDF context but we never actually acknowledged it
... suggest that we make this explicit in RDF, with the assumption that if a context is not explicitly named it's assumed to be the global RDF context
... allows authors to fix their data in a hierarchy of contexts
... this suits the cyc approach, which has been doing this for a while

<zwu2> I like the idea of importing another context

pat: also works well with owl:imports, sort of importing another context
... can add this machinery to RDF by modifying the semantics and syntax, with datasets as the syntax
... allow people to grow this global network of contexts
... supports andy's notion of islands and allows people to explicitly link their islands to participate in other contexts
... seems to give us a good story to tell on how it all links together
... and is an extension of the current RDF, not a modification of it

az: pat, you say we could integrate in RDF semantics the notion of context. couldn't we just keep RDF as it is, a context-less logic & data model, and define another contextual data model
... people would explicitly opt into it if they don't want the context-less RDF

pat: yes, we could do that. would like to give people a way to publish data in a context and provide the machinery for people to explicitly declare their context
... could certainly be done in two layers as you say

az: if it's all one layer and you publish something in a context, don't you run into issues of contexts of contexts, quints, etc?

pat: had the same worry, but there's a way to stop that from happening.

guus: if i look at your trig, everything except the first line is the same as what others have suggested
... but what's the point of the rdf:DatasetMeta in the first line?

pat: the IRI identifies a context. an IRI can both name something but also reference a context.

<Arnaud> did he just say "meta context"??

pat: rdf:DatasetMeta is a special IRI that allows us to denote the actual context, and refer to it in other places as the actual context

<AZ> this makes sense to me

pat: addresses the issue where people want to use the 4th column as a context but have it denote something else (a person, etc)

sandro: so far this sounds encouraging. a lot more details to work out.

guus: yes, many more details. would like to work out some examples.

sandro: the term that you use for importing something, is this actually asserting something to be true?
... can you use this to make statements about the truth of the stuff in a context?

<LeeF> consistent

<LeeF> ?

<zwu2> compatible contexts?

pat: i think of this more as an endorsement or agreement.

<Arnaud> re: meta context, I hope we're only talking about one level of context

pat: phrasing it in terms of absolute truth is a bit misleading. seems like you want to claim that something is globally true, which seems counter to the point of all this

sandro: but if i'm in a context and i say something is true, the truth of that can't be any stronger than my own context

pat: exactly. we want to avoid the perception that we're trying to state something as globally true.

sandro: what about the school of thought that if you're using a URI, you're accepting the document named by that URI? is that OK to do here?

pat: I'm aware of that convention. but it often doesn't work. you could think of it as an implicit version of what i'm trying to make explicit here.
... you could encourage as good practice that if you use a URI, you're accepting as true the contexts mentioned by that document.

guus: at the end of the telecon time. would like to see more email conversation on this, especially about how it relates to the example use cases.

<zwu2> bye

guus: adjourned.

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)

Found Scribe: alexhall
Inferring ScribeNick: AlexHall

WARNING: No "Present: ... " found!
Possibly Present: AZ AndyS AndyS1 AndyS_ Arnaud IPcaller ISSUE-67 LeeF MacTed NickH OpenLink P38 P41 P5 P7 PROPOSAL PatH ScottB SteveH SteveH_ Tony aaaa aabb aacc alexhall andy cgreer danbri eric ericP gavin gavinc guus joined manu manu1 mdmdm mischat pat pchampin pfps rdf-wg sandro trackbot yvesr zwu2
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy


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

Found Date: 14 Mar 2012
Guessing minutes URL: http://www.w3.org/2012/03/14-rdf-wg-minutes.html
People with action items: andy eric gavin pchampin zhe