RDFa Working Group Teleconference -- 17 Feb 2011

<trackbot> Date: 17 February 2011

<tinkster1> I might need to disappear at short notice, but other than that, yes.

<manu1> scribenick: tinkster

Default Profile URLs

<ivan> http://www.w3.org/profile/SDFGHJ

<ivan> http://www.w3.org/profile/rdfa-1.1

<ivan> http://www.w3.org/profile/xhml-rdfa-1.1

<ivan> http://www.w3.org/profile/html-rdfa-1.1

Ivan: W3C will hopefully allow us to use profile URIs like the ones I'm posting.

<manu1> PROPOSAL: The two default profile URLs should be http://www.w3.org/profile/rdfa-1.1 and http://www.w3.org/profile/html-rdfa-1.1

<ivan> +1

<webr3> +1

<manu1> +1

<Knud> +1

<Benjamin> +1

<Steven_> +1

<ShaneM> +1

<manu1> RESOLVED: The two default profile URLs should be http://www.w3.org/profile/rdfa-1.1 and http://www.w3.org/profile/html-rdfa-1.1

<manu1> ACTION: Ivan to setup the default RDFa profile URLs with the systems team [recorded in http://www.w3.org/2011/02/17-rdfa-minutes.html#action01]

<trackbot> Created ACTION-67 - Setup the default RDFa profile URLs with the systems team [on Ivan Herman - due 2011-02-24].

Steven_: do we need the "1.1" in the URI?

manu: we need a version number in case we want to remove terms/prefixes in future versions.
... and creating an "rdfa-latest" profile would create incompatibilities in documents when it changes.

Ivan: for most people, who will not express the profile explicitly, this breakage will happen anyway as parsers will start using a new profile by default.

Steven: just wanted to make sure this had been thought through.

manu: personally I want the version number in there.

Hypertext Coordination Group Participation

manu: we'd like to participate in the co-ordination group.

(people discuss Peter Mika's RDFa usage data)

manu: huge sample size; lots of non-trivial uses of RDFa.
... at least 430,000,000 pages using RDFa.

<manu1> ack [IPcaller]

Last Call Review

manu: any issues we've missed?

Toby: I've not yet provided responses for mine (draft for one, no draft for the other yet)

<ivan> ISSUE-70?

<trackbot> ISSUE-70 -- RDFa Core 1.1 LC comments about versioning from Jeni Tennison -- open

<trackbot> http://www.w3.org/2010/02/rdfa/track/issues/70

ShaneM(?): we've not discussed issue 70 yet.

manu: mostly seems to be editorial/won't fix.

ISSUE-70: RDFa Core 1.1 LC comments about versioning from Jeni Tennison

<manu1> scribenick: manu

<manu1> Ivan: I think we discussed some of this before

<manu1> Ivan: I think what she's asking for is that processors should be able to recognize whether or not they're recognizing a 1.0 vs. 1.1 document.

<manu1> ack [IPcaller]

<manu1> Nathan: As far as I can tell, I think she's saying that if she uses the same processor for a RDFa 1.0 vs 1.1 document there's no way for her to tell the difference between the two?

<manu1> ShaneM: There is a @version attribute in 1.1 - but only in XHTML

<manu1> Manu: ... explains what he thinks that Jenny wants ...

<ShaneM> Isn't this what she wants? http://www.w3.org/2010/02/rdfa/sources/rdfa-core/Overview-src.html#a_history

<ShaneM> we said "There SHOULD be a @version attribute on the html element with the value "XHTML+RDFa 1.0""

<manu1> ShaneM: @version is required for XHTML+RDFa 1.0 documents.

<manu1> Ivan: We don't say that

<manu1> ShaneM: I didn't mean to imply that - that's how my processor works.

<manu1> Ivan: My processor doesn't do anything w/ @version attribute

<manu1> Ivan: I don't see any way to allow RDFa 1.0 processors to process RDFa 1.1 documents w/ @prefix etc.

<manu1> Ivan: I think what she's asking for is guidance, not a new processor feature that ensures that RDFa 1.0 processors can process RDFa 1.1 documents w/ RDFa 1.1 attributes.

<manu1> scribenick: tinkster

manu: does anyone believe that @version will help processors?

ivan: @version is useful for Jeni's first question, but useless for the second.

ShaneM: should we document techniques for processors to conform to 1.0 and 1.1?

ivan: we could mention @version as a switch for modes?

ShaneM: we could also use <!doctype>

ivan: in many XML processors, it's simpler to get the attribute than to find the doctype.

manu: i don't think we should change processing rules. we could add a section on compatibility though.
... overall i think our goal should be to tell people to just use 1.1.

for a conIvan: formant 1.1 processor, if i get a 1.0 document, what's the proper behaviour?

manu: process as 1.1.

<Zakim> ShaneM, you wanted to discuss 1.0 support

ShaneM: if I see "1.0", then surely the correct behaviour is to treat it as 1.0. otherwise we're as bad as HTML5. ;-)

ivan: it's quite simple to switch.

manu: it would be strange if we suggested @version to detect but didn't define a @version attribute in 1.1.
... we'd have a big fight with HTMLWG if we added @version.

<manu1> ack [IPcaller]

toby: points out that HTML4, XHTML 1.x, HTML 3.2 and HTML 2.0 all had @version.

<ShaneM> why is it a bad idea to give people the opportuntity to tell a processor what version of RDFa is in a document they write? why?

<manu1> because they won't use it

<manu1> and it becomes useless very quickly if it's optional

<manu1> and if it's not optional, their documents break - same issue w/ @version in HTML4 vs HTML5

webr3: we need to not just announce which version of RDFa is being ignored, but whether RDFa is being used at all.

manu: to completely ignore any document that doesn't have @version will not work - google, yahoo, etc want to get as much information as possible, whether or not document authors stuck to the rules.

ivan: agrees

We can require @version to be conformant for authors without forbidding consumers from being liberal in what they accept.

ivan: my approach - an RDFa 1.1 processor that sees a 1.0 @version attribute SHOULD/MUST attempt to maintain backwards compatibility.

<ShaneM> Note that I do not think we need to specify any of this in RDFa 1.1. We said it in 1.0. incompletely. that horse has left the barn.

ivan: and thus we should allow people to use version="RDFa 1.1" for those people who want to target a specific version.

<ShaneM> FYI @version in XHTML+RDFa has a value of 'XHTML+RDFa 1.1"

ShaneM: we could say this in core. @version is still in the spec, we just don't say authors SHOULD use it any more.

manu: OK, we need to come up with @version values for HTML4+RDFa, HTML5+RDFa.

<manu1> version="HTML+RDFa 1.0"

<manu1> version="HTML+RDFa 1.1"

<manu1> version="HTML4+RDFa 1.1"

<manu1> version="HTML5+RDFa 1.1"

<webr3> RDFa 1.1

<manu1> version="RDFa 1.1"

<ShaneM> version="RDFa11" ?

Toby: I think I look for version =~ /RDFa\s*([0-9\.]+)/i

<webr3> rdfa="1.1"

Ivan: Nathan suggests rdfa="1.1", on the root element. Not such a bad idea.

<webr3> if you see @version = 1.0 then infer rdfa="1.0"

<Steven> I would rather argue for the use of an existing attribute

ShaneM: Ben would say that many authors can't access <html> - they're using wikis and CMSes. What are they supposed to do?

<Knud> the default version is the latest available RDFa spec?

<ShaneM> knud: yes

Toby: Steven's idea would require parsing RDFa to know how to parse RDFa?

<webr3> i have to go for 10 minutes sorry, can return

manu: We're past the hour now, so bye bye to those who can't stick around.

<Steven> Sorry I can't stay, have to get kids

<manu1> ShaneM: There is an issue w/ requiring @version

<manu1> ShaneM: If we want to introduce an rdfa attribute in the root element in rdfa core and assign it the value of "1.1" - I don't know about the HTML WG folks.

<ShaneM> http://www.w3.org/TR/2002/WD-SVG11-20020108/#version-att

<manu1> Nathan: RDFa is seen as an extension of HTML, not a part of HTML - if it's an extension, do we need a version for the extension?

<manu1> Nathan: In some cases it'll be seen as part of HTML in others it'll be seen as an extension - @version doesn't apply to everything.

<manu1> Nathan: No way to detect a version when people want to just place snippets in there.

<manu1> version="*RDFa 1.0*

<manu1> version="*RDFa 1.1*

<manu1> version="*RDFa 2.0*

<ShaneM> PROPOSAL: introduce @version in RDFa core with a value of RDFa 1.1. If it sees the 1.0 string, then it is processed as 1.0. If it sees 1.1 it processes as 1.1. if it sees nothing, it will be treated as the latest version of RDFa (currently 1.1).

-1 (SVG compat)

<ShaneM> so we are back to the 'rdfa' attribute?

<webr3> does svg not use a doctype?

<manu1> PROPOSAL: Re-introduce @version in XHTML+RDFa 1.1. If an RDFa Processor sees the string "RDFa 1.0", then it is processed as 1.0. If it sees "RDFa 1.1", it is processed as 1.1. If the processor does not see @version in XHTML+RDFa, the latest processing rules are used. If a Host Language doesn't allow specification via @version, the latest processing rules are used.

<ShaneM> +1

<ivan> +1

<Benjamin> +1

<manu1> +1

<manu1> RESOLVED: Re-introduce @version in XHTML+RDFa 1.1. If an RDFa Processor sees the string "RDFa 1.0", then it is processed as 1.0. If it sees "RDFa 1.1", it is processed as 1.1. If the processor does not see @version in XHTML+RDFa, the latest processing rules are used. If a Host Language doesn't allow specification via @version, the latest processing rules are used.

<webr3> +1 (but recognise caveat that RDFa always has to maintain backwards compat the same way HTML does w/ different rules for processors, and XHTML essentially needs it's own per version strict rules modes)

Last Call timeframe

<manu1> Manu: We'll try for new Editors draft documents in 2 weeks

<manu1> Manu: 2nd Last Call publication in 3-4 weeks from now.

- DRAFT -

RDFa Working Group Teleconference

17 Feb 2011

Attendees

Contents

Default Profile URLs

Hypertext Coordination Group Participation

Last Call Review

ISSUE-70: RDFa Core 1.1 LC comments about versioning from Jeni Tennison

Last Call timeframe

Summary of Action Items

Scribe.perl diagnostic output