Weekly RDFa Teleconference -- 24 Jan 2008

<Steven> -> Previous http://www.w3.org/2008/01/17-rdfa-minutes.html

<Ralph> Steven: note that the just-published HTML 5 WD proposes a registry for @rel values

<Ralph> ... I think we need to keep an eye on this and seed the idea of using URIs as @rel values, perhaps shortened in some way :)

<Ralph> HTML 5 WD proposed mechanism for registering @rel values

<mhausenblas> http://www.w3.org/2001/01/cgi-irc

<Ralph> scribenick: msporny

<Ralph> Scribe: Manu

Agenda

ben: chaining, we should be voting on the whole model
... do people feel like we should be voting on the individual items?

michael: do I have to vote? vote yes or no? abstain?
... I feel like I don't have a strong enough opinion on some issues.

ben: fair enough, let's see when we get to voting.

ralph: voting is meant to be only when we're at an impasse.

ben: last item on list is @src attribute.
... anything else?

mark: there are some items that should be voted on separately.

mark: gives examples of some issues.

ben: any particular order you want?

<scribe> ACTION: Ben followup with Fabien on getting his RDFa GRDDL transform transferred to W3C [recorded in http://www.w3.org/2007/11/15-rdfa-minutes.html#action01] [CONTINUES]

<scribe> ACTION: Ben to add status of various implementations on rdfa.info [recorded in http://www.w3.org/2007/10/04-rdfa-minutes.html#action06] [CONTINUES]

<scribe> ACTION: Ben to respond to comment on follow-your-nose [recorded in http://www.w3.org/2007/12/06-rdfa-minutes.html#action02] [CONTINUES]

<scribe> ACTION: Ben to set up a proper scribe schedule [recorded in http://www.w3.org/2007/10/11-rdfa-minutes.html#action01] [CONTINUES]

<scribe> ACTION: Michael to create "Microformats done right -- unambiguous taxonomies via RDF" on the wiki [recorded in http://www.w3.org/2007/08/23-rdfa-minutes.html#action06] [CONTINUES]

<scribe> ACTION: Ralph followup with Dublin Core on what's going on with their namespace URI [recorded in http://www.w3.org/2007/12/06-rdfa-minutes.html#action01] [CONTINUES]

ben: nice to know Dublin Core is paying attention to RDFa.
... also nice that people are looking into trackback.
... as a use for RDFa.

non-prefixed @rel/@rev and reserved words.

ben: we need to understand the end-effect.
... I think we are in agreement at this point.
... @rel/@rev, when it is a reserved XHTML word, it should generate a proper triple.
... non-prefixed XHTML words should be ignored.

<benadida> PROPOSAL: @rel/@rev non-prefixed values other than reserved keywords are ignored. Reserved keywords resolve to vocab#[keyword].

steven: with the release of HTML5, they have new @rel values

michael: can we decouple that?

ben: they do have their own DOCTYPE, so will probably have to tweak a few things.

<mhausenblas> +1

RESOLUTION: @rel/@rev non-prefixed values other than reserved keywords are ignored. Reserved keywords resolve to vocab#[keyword].

<scribe> ACTION: Ben to update tracker with todays resolutions. [recorded in http://www.w3.org/2008/01/24-rdfa-minutes.html#action07]

<Ralph> scribenick: ralph

test cases

<msporny> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2008Jan/0252.html

-- test 76

Manu: test 76 checks every one of the @rel reserved values that we currently know

<msporny> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2008Jan/0252.html

RESOLUTION: test 76 approved

Manu: the Syntax doc mentioned 'start' but no 'end' or 'last'
... is that ok?

Mark: I think it would be handy; does it hurt?
... also consider putting the GRDDL value in

Ben: but the GRDDL relation is in a different namespace

Mark: oh, right.
... should we test for unrecognized values not generating triples?

Manu: we don't test for spurious triples; we've said those don't go into the default graph

<Steven> I agree with including 'last'

Michael: we do have a way to say that the 'pass' result is that the query returns no result

Ben: I thought we'd decided to ignore extra triples

Mark: during that discussion, Ben made a strong case for "these triples and only these triples"
... so we came up with the "default graph" notion
... the draft was reworded from "no triples shall be generated" to "no triples shall be added to the default graph"

<Steven> I oppose nofollow

Ben: propose adding another test for non-reserved values; this would be a negative test
... 'nofollow' would be in the negative set

Manu: should we add 'last'?

<Steven> I agree with Ben there

Ben: I think this is for the XHTML2 WG to decide, not this TF

[Ralph] I agree, out of scope for us

Shane: any objections here from the XHTML2 WG folk?
... we'll recommend to the XHTML2 WG to add 'last'

-- test 77

<msporny> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2008Jan/0253.html

<markbirbeck> Mark: Pointing out that once these are added, they are usable everywhere, not just in <head>. So they would help in lists of any items, for example.

Manu: test 77 tests the reserved words in @rev

RESOLUTION: test 77 accepted

<msporny> scribenick: msporny

<Ralph> Manu: not sure we reviewed test 17

Chaining

<benadida> --> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2008Jan/0257.html

Ben: There is an e-mail from mark summarizing the issues: http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2008Jan/0257.html
... rel and rev issues, which we just addressed.
... @src/@href/@resource to complete triples.
... should we go in order that they're stated in the e-mail?

mark: if everybody agrees to the generic chaining thing, there is nothing to discuss...
... if everybody doesn't agree, we have a list of individual items we can talk about.
... you could break it down further

ben: lets proceed optimistically, see what the opinions are on the first issue.
... if not, we'll slice it down.

mark: closing statements - we have a general algorithm on the table
... @instanceof always applies to a subject.
... we have things to set the subject, @about/@href/@src.
... based on a @rel, subjects either slide left or right of the @rel.
... if we have general rules, we might have some wierd stuff going on.

ben: I think Mark's proposal is very consistent.
... my least preferred option is to mix my model and Mark's model.
... there is a different model to interpret chaining,
... @href doesn't complete a hanging rel.
... the only time you have it chaining is with @about.
... my worry is that with attributes shifting left/right, and @href becoming the subject of @property
... it will become more confusing.
... any questions?

ralph: wondering if there is a summary in a single place showing side-effects of decisions.

ben: not really, not as complete...
... Mark's e-mail is a fair summary of what is going on.
... do folks feel confident in the understanding of what is going on?

ralph: nope

michael: yep

Manu: yep

ben: let's start with @href/@resource completing hanging @rel.

<benadida> <a href="#mark">Mark</a> and <a href="#ben">Ben</a>.

<benadida> Note how I'm using <a href="http://foaf.org">FOAF</a> here.

manu: I'm for @href/@resource completing hanging @rel.

michael: yes, for it.

steven: I can live with either, abstain.

shane: I'm abstaining.

ralph: No problem with either, abstain.

ben: i would object, but I would be the only one.
... no reason for me to object at this point.
... as chair, I'm very happy.

<benadida> PROPOSAL: @href/@resource complete hanging @rel's

RESOLUTION: @href/@resource complete hanging @rel's.

ben: rel="", should that shut it off?

mark: it works... would we promote it?

ralph: do the implementations naturally do that?

ben: we should tell people to push hanging rel as deep down as possible.
... an extra rel should instantiate a bnode?

mark: what that relies on is that a bnode that doesn't generate a triple shouldn't be used as the object for a higher-triple.

ralph: rel="" is ugly markup...

ben: let's push the issue off until we need it.

ralph: let's not include it.

mark: this was never meant to be part of the issue.

ben: there are different ways to solve the issue.
... anybody that really wants rel=""?
... seeing none, let's move on then.

@href/@resource become subjects?

ralph: what about @src?

ben: any rules that we make about @href/@resource would apply to @src.
... if @src is equivalent to @href/@resource, it will complete triples. Same for @about equivalency.
... two proposals: @src chains like @about, or @src chains like @resource.

ben: the only part where it makes a difference is where you want to relate information for an IMG element.

<Ralph> [I understand "behaves like" to be "chains like"]

mark: is it worth voting now , should setting subject include @src?

ben: do we want to be able to be able to plop in a predicate with @src?

manu:+1 for @src behaving like @about

mark: to clarify - if you put a div above the img you get several types of goodness.

mark: like being able to note a license, or have @instanceof apply to the @src

manu: I agree, @src should chain like @about. +1

mark: +1

shane: +1

michael: +1

steven: I have a small problem with it...
... no problem with this, but we discussed using @src a long time ago - found a problem, but I can't remember what the problem is.
... abstain

<Ralph> we're asking whether the triple <ben.jpg> <xhtml:license> <http://creativecommons.org/licenses/3.0/> ?

ralph: don't quite see the issue... based on how the question is worded.

ralph: I thought we had already decided this once before.

ben: we did, but when Mark explained the chaining model, we could move @src to left.

ralph: so this is a significant change from the previous model?
... both ben and mark seem to be happy with it.

ben: any other comments?

<benadida> PROPOSAL: @src behaves like @about and is overridden by @about

<mhausenblas> +1

manu: +1

steven: abstains

RESOLUTION: @src behaves like @about and is overridden by @about

<Steven> Here is my previous mail:

<Steven> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2006Jan/0018.html

ben: lets go on for another 5-10 minutes, since we're getting close to time running out.

@href/@resource as subject for @property

<benadida> Ivan

ben: when @href/@resource is on an element, should @property apply to that @href?

ben: since we have agreed that @href chains like @about, then we have to agree that @href can set subject.

ben: mark any additions?

mark: +1

manu: +1

ralph: I feel less comfortable about this....

<markbirbeck> I met with

<markbirbeck> the other day.

mark: there is an e-mail to the list with an example...
... we're saying that @href completes hanging triples, it sets the subject for any contained element.
... maybe we should change what we're talking about.
... what we're voting on... if we vote on if @href should set property, everything else falls into place.

ralph: if mark and ben are happy, then I may be happy.

mark: the general thing about all of this is that what happens when you add/remove @href/@about... is it consistent.

ralph: if it is consistent, then that's good.

ben: any other comments?

manu: +1

michael: +1

shane: abstain

steven: don't like it, but don't want to block us... abstain.

ben: +1

mark: +1

ben: my +1 is based on the previous decision, I value the consistency above all else.

ralph: abstain

<benadida> PROPOSAL: @href/@resource without @rel/@rev/@about sets the subject for @property on the same element.

RESOLUTION: @href/@resource without @rel/@rev/@about sets the subject for @property on the same element.

If subject of @property should be @href, shouldn't @instanceof follow?

ben: last issue for the day?

manu: don't like it, @instanceof has different semantics from @property, so it's okay for it to apply to different attributes than @property. My argument, however, is counter to the way we've been voting today. It doesn't seem consistent, and I argue that it doesn't need to be consistent since @instanceof can create a bnode, it isn't like @property in that sense.

ben: it would be wierd if @property applied to @href, but @instanceof didn't.

steven: I need to think about this more.

manu: I think @instanceof should apply to @about and @src only.

<Ralph> Ralph: I think @instanceof should work like @property, so whichever choice makes that work is what I want :)

manu: abstain

michael: +1

steven: abstain
... as long as I can send in last call comments.

ralph: I don't want to revisit @class :)

mark: We need to finish this.

*scuffle*

ralph: If we talk about this more, it should be new data.

ben: in the interest in getting to last call, I'd like to take a vote.
... I'm fine with you bringing up new information at last call.

mark: let's put the vote off...

steven: I don't like @instanceof...

ralph: we can't drop @instanceof at this point...

ben: we've resolved alot of good things at this point.

steven: we need to get to last call, willing to abstain on this issue.

manu: Ben with his Chair hat on, notes that steven may bring up new information at last call.

shane: abstain

ben: 3 abstained from the vote.

<benadida> PROPOSAL: @href/@resource without @rel/@rev/@about sets the subject for @instanceof on same element.

RESOLUTION: @href/@resource without @rel/@rev/@about sets the subject for @instanceof on same element.

ben: can you spend some time on the spec to bring it up to snuff?

mark: yes

ben: I'm very excited that we're getting close to last call.
... I will work on my implementation... any closing comments?
... end of call, let's look at Mark's latest draft when it comes out. Steven/Shane we need to have XHTML working group review before last call.
... draft by tomorrow, mark?

mark: yep

Weekly RDFa Teleconference

24 Jan 2008

Attendees

Contents

Agenda

non-prefixed @rel/@rev and reserved words.

test cases

Chaining

@href/@resource become subjects?

@href/@resource as subject for @property

If subject of @property should be @href, shouldn't @instanceof follow?

Summary of Action Items