RDFa Working Group Teleconference -- 10 Feb 2011

<trackbot> Date: 10 February 2011

(ipcaller isn't there, that's me)

I /was/ ipcaller, i hung up, ipcaller name stayed..

ISSUE-73 and ISSUE-78: RDFa Default Profile

http://www.w3.org/2010/02/rdfa/track/issues/73

http://www.w3.org/2010/02/rdfa/track/issues/78

<manu> Proposal: http://lists.w3.org/Archives/Public/public-rdfa-wg/2011Feb/0099.html

ivan: choice for URIs for profiles is important, dated or not
... I'll discuss with people in W3C about this

<manu> ACTION: Manu to contact SemWeb Coordination Group to discuss default RDFa Profile URLs [recorded in http://www.w3.org/2011/02/10-rdfa-minutes.html#action01]

<trackbot> Created ACTION-59 - Contact SemWeb Coordination Group to discuss default RDFa Profile URLs [on Manu Sporny - due 2011-02-17].

ivan: technical question: I read your mail as if we have a profile valid for RDFa Core, and a different one for HTML+RDFa etc, is that correct?

manu: I don't think we can have just one profile, for instance HTML will have custom terms

ivan: that creates some issues.. it means that any host language has the right to define a default profile
... are there any restrictions on host languages?
... or perhaps we say we have to default profiles, a core one for everyone, and some host languages can add a host specific one..
... not only W3C can control host languages, other standardization bodies could do other RDFa host languages, sem web activity lead can't constrain what they do

manu: host languages could override default profile..

ivan: yes they could do that anyway by redeclaring all terms in default core profile
... I am leaning that we have two, core profile for all RDFa, and a language specific one

<manu> Nathan: Core profile for all of RDF? Or core profile for all of RDFa?

ivan: we can't define for RDF, we are RDFa working group, RDF WG will need to decide that

<manu> Manu: Do we expect default profiles for RDF/XML or Turtle?

ivan: it may happen, it does not seem very likely for rdf/xml and turtle at least

<manu> The profile for RDFa - http://www.w3.org/2011/profiles/rdfa

<manu> The profile for (X)HTML+RDFa - http://www.w3.org/2011/profiles/htmlrdfa

Manu: we would be defining two profiles, RDFa default, profile for (X)HTML+RDFa

Ivan: Next Point, still related, we had a standing issue, if we do that, how do I know that I am managing XHTML+RDFa? (to get profile)

Manu: I was thinking we could trigger off (some element, <html> etc)

Ivan: first question is, is this something we need in the document?

Manu: yes..

<Zakim> ShaneM, you wanted to discuss announcement

Ivan: i don't like this sniffing to much..

Shane: I thought we debated and agreed 3/4 weeks ago that the only thing we'd put in the document was that we detect on media type

Manu: anybody disagree with triggering of media type to get the correct profile?

<manu> Manu: So we trigger off of "application/xhtml+xml" or "text/html" or "application/xml" (or whatever the proper mime-types are)

general agreement heard (nobody disagrees)

Ivan: I am uneasy with the text: "Prefixes and terms MAY be /updated/ if the new meaning of the term or prefix is semantically backwards compatible with the previous term or prefix."

Manu: example for this is dublin core, dc11 vs dcterms - dc tried to make sure terms were "supported" in new vocab
... you could change vocab uri and have same meaning

ivan: no that's wrong, the two are very different, from rdf point of view

manu: i said semantically because the meaning hasn't changed domain + range setting still means property means "title"
... how are they not the same?

ivan: because RDF and other tooling sees things differently, takign in to account these new statements
... as a rule we should say, prefixes and terms should nto be changed

<Steven> Why commit ourselves one way or the other?

shane: I don't know if I agree w/ ivan, let's consider og: (opengraph for a while...) [considers]

<manu> Only because we're trying to give guidance to vocabulary authors, Steven.

shane: og may change data at their uri, updating it - if that's correct why isn't the other?

ivan: that's correct, this is a problem at sem web level and RDFa should not even attempt to solve

<manu> Steven, I don't think we're trying to commit ourselves - just give guidance? (but we may be accidentally committing ourselves)

ivan: what manu proposes is that the triples will be different (differen uris)

<Zakim> webr3, you wanted to agre w/ ivan

nathan: also people hard code against URIs, we can't have them changing

<ivan> 6. Vocabulary maintainers SHOULD include an 'Expires' header in the HTTP response when a profile is dereferenced via HTTP. RDFa processors MAY use this header to implement local caching of the profiles.

<manu> Nathan: We should also probably put "Last-Modified", etc.

<manu> Nathan: "E-tag" if possible, "Cache-Control"...

<manu> Ivan: Concerned that that's going to be difficult to implement - too complicated.

<manu> Nathan: The server should generate everything correctly.

Manu: probably want to write or point to some good guidance on caching

Ivan: yes everything is set automatically, but Expires needs set specifically normally
... last thing is dated vs non dated uris for profiles
... I want to suggest we have non dated URI and provide dated too so people can reference explicitly

<Zakim> manu, you wanted to discuss URI dating

Manu: I'm fine with versioned URI, issue i have with unversioned is that we could never remove prefixes or terms from it

<manu> Nathan: We also have to keep /terms/ in mind.

<Zakim> ShaneM, you wanted to disagree with ivan's URI propsoal

Shane: I agree w/ you manu - we debated this the other day, it was a light discussion
... if we have a /tr/ space URI then it means we can never remove a prefix

ivan: but that's what caching is for

<manu> What happens when somebody does this in 2011: profile="http://w3.org/profiles/htmlrdfa"

shane: but as soon as I update to the latest profile, every document stops working

<manu> http://w3.org/profiles/htmlrdfa

ivan: but the proposal of manu is to say that we will have a new dated profile where it's lost anyway

(scribe misses what manu says, he seems to be going against the idea of unversioned uris)

ivan: we removed any way to tell what version of RDFa is being used

<manu> Manu: I think that we need to either have a versioned or dated URI

ivan: maybe we need to have version in there again (?)

manu: we tell people not to use profiles if they want interop, if they do want to use a profile they should mention which profile specifically
... and our fallback is, if they haven't done any of it, we'll use the latest default profile to try and get some triples out
... so we have levels of protection

ivan: okay, i see what you mean...
... I think it's fine then, let's just not use "2011" in the uri

<manu> What about http://w3.org/profiles/rdfa11 and http://w3.org/profiles/htmlrdfa11 ?

manu: k, mmm, k

<manu> What about http://w3.org/profiles/rdfa1.1 and http://w3.org/profiles/htmlrdfa1.1 ?

<manu> What about http://w3.org/profiles/rdfa1 and http://w3.org/profiles/htmlrdfa1 ?

nathan: remembers Ivan is goign to speak to w3c for guidance on this

manu: any other concerns?

<manu> Nathan: I think it's about as good and close as we can get - I don't like default profiles, but I don't think others agree with that viewpoint, so this is as good as we get.

<ivan> ISSUE-73?

<trackbot> ISSUE-73 -- The RDFa WG needs to determine how each RDFa Profile document is managed -- open

<trackbot> http://www.w3.org/2010/02/rdfa/track/issues/73

manu: ivan you had 73? can you type up a response LC one?

ivan: it was raised by manu sporney, do we need to reply to him?

<ivan> issue-78?

<trackbot> ISSUE-78 -- Should we have default prefixes and terms for host languages -- open

<trackbot> http://www.w3.org/2010/02/rdfa/track/issues/78

[general conversation about who LC replies to 78]

<manu> ACTION: Manu to write up RDFa Profile Management on RDFa Wiki [recorded in http://www.w3.org/2011/02/10-rdfa-minutes.html#action02]

<trackbot> Created ACTION-60 - Write up RDFa Profile Management on RDFa Wiki [on Manu Sporny - due 2011-02-17].

manu: I'll put a draft response on wiki, Ivan will clean up and respond properly
... I'll do the response to 73

<manu> ACTION: Manu to respond to ISSUE-73 [recorded in http://www.w3.org/2011/02/10-rdfa-minutes.html#action03]

<trackbot> Created ACTION-61 - Respond to ISSUE-73 [on Manu Sporny - due 2011-02-17].

Ivan: next step will be at some point, what will be the initial content in these profiles?

manu: later, lets do it later please

shane: it's a last call issue to define these as part of the docs going out
... we can't do it once it goes out, contents of default profile is tied to the spec, we need to get consensus, lc-wise, to it

manu: that effectively says we shouldn't change the default profile after it goes to rec..
... we could just define the minimum possible..

ivan: i think we could do the following.. there is a core set of terms and prefixes we just need, we have to pop them in and collect them
... what we should set up is the mechanism whereby the resolution will happen
... let's only put in w3c well knows first, rdf: for example
... then open up mechanism after rec to add new things

<Zakim> webr3, you wanted to ask if any other spec references a "moving part"

nathan: does any other rec have this moving part to it?

ivan/steven: yes, plenty xml related to

manu: okay so we don't think this will hurt our LC
... we need terms and prefixes to add

<manu> ACTION: Nathan to put together list of prefixes and terms for default profiles [recorded in http://www.w3.org/2011/02/10-rdfa-minutes.html#action04]

<trackbot> Created ACTION-62 - Put together list of prefixes and terms for default profiles [on Nathan Rixham - due 2011-02-17].

nathan: I'll do that

manu: top of the hour

ivan: how about edits?

shane: how should i do them, edit or propose edits?

manu: let's not get stuck behind the bike shed
... my preference is just to edit the spec

shane: that's fine w/ me, that's my preference too

[general agreement]

meeting ended

tracker, make draft minutes

- DRAFT -

RDFa Working Group Teleconference

10 Feb 2011

Attendees

Contents

ISSUE-73 and ISSUE-78: RDFa Default Profile

Summary of Action Items

Scribe.perl diagnostic output