XHTML2 WG FtF, Venice, Italy, Day 1

18 Feb 2008

See also: IRC log


Rich, Roland, Steven, Yam, Alessio, Shane, Gregory, Gerrie
Tina, Mark
Steven, Roland




RDFa last call

<scribe> Scribe: Steven

Rich: I see that ODF 1.2 will be using RDFa

Steven: I heard that they are changing some of the attribute names
... but that is not yet final

Roland: Well then they aren't using RDFa then

Steven: Well, they are, but with a differen syntax :-)
... I think they are inporting the attributes into their own namespace


Roland: Does the current spec allow this?

Steven: Yes, I think all our modules are chameleon; it is the driver that adds the namespace

Roland: I think this is good. It may help clean up some of the namespace mess

Steven: I announced the RDFa last call at HCG last Friday. The dates were announced as 6 weeks ending April 4, but that means starting 22 Feb, which I am not sure we will hit.
... I see a new, potentially final pre-last call draft was published this weekend
... so there is still a reasonable chance to get it out this Friday
... The problem with the spec is that it is for implementors and not authors

Roland: What we really need is an RDFa cookbook, that gives lots of examples that you can just copy
... exemplars

Steven: Which examples do you think are needed?

Roland: Geo coding, people, ...

Steven: Events

Roland: W3C should be doing this much more centrally organised
... and host the namespaces and so on

CURIE last call

Steven: We got the OK on the dates at the HCG meeting



Steven: The current draft that is up is 22 Jan, so the last changes have not been made
... I know Shane has been having trouble uploading to the W3 site

Roland: So we are OK to go?

Steven: HCG has OKd it, we can publish asap

Roland: So we just need Shane to upload the draft

<scribe> ACTION: Shane to upload final CURIEs draft [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action01]

<scribe> ACTION: Steven to Draft last call announcement [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action02]


Roland: We are ready to go to second last call
... but I wonder if we should let it wait a bit so that it runs behind CURIEs, which depend on it

Rich: In ARIA some of the standard roles in the Roles spec like note and definition, they are a subclass of 'region'
... and I would not do that


Rich: I could tweak the examples in the appendix to fix that
... Why are things in orange?

Steven: [laughs] That is for all SHOULDs and MUSTs
... CSS style

Roland: But we don't say that, we should

<scribe> ACTION: Shane to add text explaining why some text is in orange [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action03]

Rich: Oh wait, I already did that change

Roland: We need a role cookbook as well

Rich: My change is in section C -


Rich: I have to say that the taxonomy was one of the best pieces of work we (WAI) has done
... since it allowed us to check correctness really easily
... OWL is a fantastic, and simple, tool for that

Roland: Such taxonomies should be dereferencable via a URI, and not just an appendix of a doc
... so appendix C should be a separate document that people can download

Steven: Does OWL have a Media Type?

Rich: No


Steven: Why is that? Oh I see, it is just RDF/XML
... Maybe an interesting experiment to represent OWL in RDFa
... We need to decide on dates for second last call

Roland: So we publish CURIEs tomorrow, let's say, done 18 March
... and look at a role last call starting 18 March
... unless we have a lot of CURIE push back
... we should try to avoid a third last call!

<Roland_> one last role issue from Shane : http://htmlwg.mn.aptest.com/cgi-bin/xhtml2-issues?findid=8001

Steven: So he wants the roles to be machine discoverable
... which is a good idea
... it seems like a good opportunity to use RDFa:
... when you go to the URL you get an XHTML document that is readable for the human, explaining the values
... but is also machine readable for UAs
... You do want UAs to be extensible for new role values
... otherwise new role adoption would be very slow

Rich: But it may be in the middleware where the presentation of the role is made
... it may not make it down to the UA

Steven: But from the user's perspective middleware is just part of the UA

Rich: We don't have a method at present of specifying priorities of roles
... there is no semantic information that says that secondary is of lower priority to primary
... at least with RDF you can always update it later

Steven: I don;t think we disagree with his points here

Rich: No not at all
... though I doubt you can specify the end rendering
... So how should we respond to him? Firstly create a URI to the roles
... the OWL document
... Steven could do an RDFa version

Steven: This is a problem with OWL not having a media type: you can't use content negotiation to specify that you only want the OWL version

Roland: So should we make it a SHOULD or a MUST that people make dereferencable definitions of their role values?
... Is it fair not to specify which representation of RDF is required?
... otherwise the UA must be able to handle all sorts of serializations

Steven: Triples, Turltle...
... But it's not nice to exclude future possible representations of RDF either

Roland: It is a problem for interoperability though

Steven: We need to specify the exact algorithm to get to the machine readable definitions

Roland: Where do we specify the 'vocab' bit of our role namespace

Steven: Section 3
... http://www.w3.org/1999/xhtml/vocab#
... http://www.w3.org/MarkUp/2008/ED-xhtml-role-20080128/#s_role_module_attributes

[Discussion of problems with IE's CSS and namespaced values]

RESOLUTION: The creator of a role vocabulary SHOULD make a derefencable machine-readable definition of the vocabulary

Roland: and one of the available RDF representations SHOULD be RDF/XML

RESOLUTION: The creator of a role vocabulary SHOULD make a derefencable machine-readable definition of the vocabulary and one of the available representations SHOULD be RDF/XML.

Roland: Going back to the issue we are discussing


STeven: We can't do the primary/secondary importance at present

Roland: That is being considered by UWA at present

Steven: So we could say that that is work in progress at the moment

Rich: It is beyond the scope of this release of role

Steven: So who's going to reply?

<scribe> ACTION: Roland to reply to issue 8001 [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action04]

XHTML Basic 1.1

Steven: Here is an icomplete implementation report, that needs some details filling in; maybe we can do it on the fly:


Yam: I believe that Openwave has an implementation, but I am not sure
... Netfront definitely has


Steven: So what do we do? Wait for a second implementation, or argue that it is a hint, and therefore any implementation conforms

Roland: So when will OMA XHTML MP 1.2 be ratified?

Yam: Soon, it is in CR now

Steven: With inputmode?

Yam: Inputmode is in MP 1.3

Steven: The third option is to drop inputmode

<Tim> hi

Roland: If OMA is happy with only one, if we could get MWI's statement that they are happy with one implementation, then we could move forward that way

Steven: So we go with option 1.

Yam: So you have to contact Mike Smith

Steven: I will try and contact Mike Smith, but he is in Tokyo, so it is a bit late now

(sent mail)

CSS Namespace

Steven: I see CSS has published a last call on namespaces in CSS


Steven: We should review it

[ADJOURN for lunch]

<scribe> ACTION: Steven review CSS Namesapces module [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action05]

About to reconvene


<ShaneM> /whois r*

<ShaneM> hi - alarm malfunction

I didn't know you could wildcard whois Shane


Hi Shane

We've already done quite a lot

see the minutes


What is the prob with the W3C site Shane?

<ShaneM> the "2008" folder fell out of my jigedit tree again


<ShaneM> annoying

but you did fix it the first time?

<ShaneM> they fixed it on weds. broke on fri or sat.


Do you know what caused it?

<ShaneM> not a clue. just went away. dont let me distract from the meeting tho plz. I will work this in parallel


we're just restarting after lunch

We are in a beautiful old Venetian Palace

with a view on the canal

It was a little chilly this morning, but we have the sun on our backs now

Want to skype shane?

<oedipus> GJR: good afternoon/morning/[your time of day here]!; will be participating in this session in a minute -- going over the IRC log and minutes

<alessio> hi shane, gregory

<oedipus> steven, i hope i didn't muck things up by trying to fix the datestamp on the wiki!

<oedipus> aloha, alessio -- it's good to hear you, but i wish i were doing so in person (and with a speech synthesizer)!

No Gregory, I hadn't noticed

was that that it said 2007 instead of 2008?

<oedipus> yeah - the files themselves were datestamped 2007 so i set redirects to the same content with a 2008 datestamp

<alessio> I know gregory...

good work; we wondered who had done that, but were glad


<oedipus> just glad i didn't screw things up -- and VERY glad that the wiki is a mediawiki

Roland: So at the last FtF we decided that since there is now HTML5 we can be more radical and achieve more advantages to meet our aims

Rich: UWA are trying to do the same things. Maybe we can pool

Roland: I suggest we focus on the markup side of what gets authored
... UWA focusses on adaptation
... but we need to make sure they gel
... but there is a good separation of work

Yam: I have two views
... XHTML2 is far more new generation markup
... but lessons from the past: we need to work faster
... and changing path now might make more volatility
... XHTML1 is more conservative than HTML5 in some respects
... in M12N we don't specify semantics, leaving HTML4 to specify the semantics
... we need to cut and paste the semantics somewhere, and decalre HTML4 is the past
... we can use some parts of HTML5, such as <section>
... with existing browsers we can easily represent XHTML2

Steven: Well, the semantics *have* already been moved to the XHTML2 spec
... I've never really understood the objections to XHTML2, since it pretty much all can be done in existing browsers already
... my conclusion is that namespaces are the root of the complaints

Yam: I just don't understand why people have problems with namespaces. You really need them, and they are an underpinning of W3C

Steven: I think in the end that HTML5 will discover that they need something like namespaces anyway, even if they call it something else

Roland: We need to find a way to make authoring with namespaces easy

<ShaneM> The problem with namespaces is the ability to arbitrarily change prefixes throughout a document stream. Overengineered. You just need the ability to establish a prefix binding for scoping.

Roland: Maybe we could move everything into a single namespace

Yam: I don't really see the problem

Roland: Our design should minimise the need for namespaces in authoring

Yam: We need to make XHTML2 real as soon as possible

Steven: There is an open source javascript library for XForms
... we could expand that to do the rest of XHTML2, so it would work in browsers now

Roland: So what is it that we are doing new?

Steven: Device independence, accessibility, internationalisation
... authoring

<oedipus> GJR: plus 1 to Roland's "move everything into a single namespace"

Alessio: If you want video or audio, you should use SMIL content for that

Roland: If I want to compose different content from different languages I should be able to.

<oedipus> GJR: SMIL Timesheets - a good model offering user-end control a l� stylesheets; role for OBJECT also good idea rather than VIDEO, AUDIO, etc.

Alessio: I strongly believe that we should use CML markup to represent the blocks that are combined in a page

Steven: What you want to do for documents ROland is what modularisation does for languages

Roland: Yes
... you want to compose documents from document blocks
... 'islands'


Steven: This is why all elements have a src attribute
... to allow you to import that content, but if the link fails or media type doesn't match, you still have some content

Roland: I think the content should be able to be changed on the fly, at runtime

[discussion of widgets, and their definition]

Yam: What is the difference between widget and application?

Alessio: none

Yam: From my point of view, it's all applications

Roland: I think it's just terminology

<oedipus> GJR: that's why we need the semiotic web, not just the semantic web!

Roland: it is independently written, it publishes its events, and defines what it will do in the page context
... One of the current problems with 'widgets' is the security issue
... We should address that
... people like openajax are working on it

<oedipus> GJR: notes that PF just dealt with a potential scripting trojan horse delivery method by removing aria's secret property, due to possibility of author abuse

Roland: but we need to mark the security boundaries in documents

Alessio: Aggregation of documents is very important, thart HTML5 is missing

Steven: Part of the problem is that they didn't design composability into XML, look at unique IDs

Roland: When XML Schema started I asked for an 'XML Document' datatype, and they said "We can't do that"

<oedipus> roland, who said "we can't do that"?

Gregory, XML doesn't allow composition of documents; it was the schema WG who said that

<oedipus> thanks for the clarification -- that's what i concluded, but wanted to be sure

Roland: We need to define event boundaries when including other 'applications' using src on an element

Yam: We definitely need some guidelines

Roland: It may depend on particular events, some may be allowed to go over boundaries
... and some not

<ShaneM> Have you considered http://www.w3.org/TR/2008/WD-access-control-20080214/

Roland: Google maps sends events into the containing page

<gshults> Gerrie Shults here. What phone # and code do I call?

<gshults> OK. It'll take me some time to get Skype installed and configured

sorry about that Gerrie

<gshults> Some things just can't be helped/avoided.

<oedipus> q cite="http://www.w3.org/TR/2008/WD-access-control-20080214/#security"

<oedipus> "Application authors should be aware that content retrieved from another site is not itself trustable. Authors should protect themselves against cross-site scripting attacks by not rendering or executing the retrieved content directly without validating that content. Authors sharing content with domains that are on shared hosting environments should ensure to not allow access from arbitrary ports on those domains."

<yamx> Thank you for the reference.

Roland: So one of the core issues we should deal with is composition
... And make sure that composite objects are first class citizens, with security, events, etc all taken care of
... we need to work with other groups, and we take care of the markup bit
... Many of the security problems are the fault of security experts being unable to explain the issues to the outside world
... people we need to work with iunclude UWA, Web API, Security

<Roland_> Web Security Context Working Group -- http://www.w3.org/2005/Security/wsc-charter

Roland: Are there any other issues we need to address in XHTML2?

<ShaneM> I think we need to publish an updated draft RIGHT NOW. how can we do that?

Steven: We have been waiting for RDFa, role to be ready to plug them all together

Roland: Should we be creating legacy modules?

<scribe> ACTION: Steven to organise a new draft of XHTML2 [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action06]

<ShaneM> feel free to comment on the editors draft. its current.

Steven: The question of new modules relates rather to M12N V2

<oedipus> shane, ok - that's what i was checking over the weekend - i should check the issues list before posting anything, though...

Rich: Then there is the naming problem

<markbirbeck> I have to make still more changes to the RDFa syntax document :( so can't join yet.

Alessio: XHTML2 SP1

<markbirbeck> But...can't we produce a modularised version of XHTML 2 using the current M12N?

<markbirbeck> Would give XHTML 2 a little momentum if it was simply a set of modules/extensions to XHTML 1.1.

<ShaneM> no you cannot

<oedipus> GJR: what would give XHTML2 momentum would be to go to TR before HTML5

<ShaneM> I dont htink you all can hear me on skype... steven can you check your mute or whatever?

Shane: ... crackle... splut...

Now you realise Mark???

<ShaneM> I will figure it put

<ShaneM> I will stop call and calibrate microphone

we will stop to get a coffee

Jersey??? What are you doing there?

<oedipus> er, new jersey, not the isle of...


<oedipus> but i consider myself a jerseyman above all else!

Shane: A question
... we should put a draft out of XHTML2 now, and we could do a M12n 2 at any time we wanted

Roland: Will we be able to mix and match M12N2 and M12N 1.1 modules?

Shane: No real need since every module in 1.1 is also in 2.0

Roland: Can we use the headers module from 1.1 for instance

Shane: There isn't one, the text module is rather large in 1.1

<gshults> How do I call in with Skype?

<gshults> thx


Steven: Just got a reply from Mike Smith that he is OK with us moving forward with one implementation

Shane: What are the three things we have to do right now to get out a new XHTML2 WD?
... What is new: rdfa stuff, which is out of synch

Steven: But we can point to the RDFa spec now
... like we do with XForms

Roland: And there is CURIEs and role

Shane: Not sure we can do it with RDFa
... because it is an M12N 1.1 module

Steven: But XForms is modularised at all (yet)
... The syntax is trivial
... we refer to RDFa for the semantics

Roland: Access is in XHTML2 as well
... so we can refer to those too

<oedipus> http://www.w3.org/MarkUp/2007/ED-xhtml2-20071024/mod-access.html#s_accessmodule


Roland: So sections 23 and 24 can largely be deleted

Steven: And PLEASE let us drop the style attribute

<oedipus> 23 and 24 are the 2 metadata sections, right?

Roland: We can move them to legacy modules

Shane: I think it is worth keeping the style attribute

<oedipus> GJR: style as in 29, not 28? or both?

Shane: One of the reasons we have everything inline was to make a readable document.
... where everything was defined in place
... I don't mind if we drop that requirement

<oedipus> GJR: would be willing to give up 28 (which is really inline style declarations as substitue for FONT), but not 29

I agree Gregory, but there is no consensus here I am afraid

<oedipus> well, count me as plus one to deleting 28 and retaining 29

Yes, we are talking about 28, not 29

<oedipus> section 28 is perilously close to HTML5's "Presentational markup" - http://www.w3.org/TR/html5/#presentational

<ShaneM> basically we want to chnage section 4 so that some items are now outside the document - then remove the relevant sections.

Roland: Going back to headings and legacy

Steven: Well I think they are two different things. Firstly we need to decide what is in XHTML2, and secondly we need to decide what is M12N v2

Shane: But creating an h1-6 legacy module is fine

Steven: All *HTML* (where the *s are wildcards) should be in a module somewhere

Roland: Is there an M12N v2 document yet?

Shane: no, only in the source tree

Roland: Is there anything else at this stage that needs to go into XHTML2?

Steven: Don't think so



Roland: Why isn't iframe in XHTML2?

Steven: Because everything you can do with iframe you can do with object

Alessio: but object allows fallback

Rich: We should create a taxonomy for all elements like we did for ARIA

Roland: There is a question of familiarity

<oedipus> GJR: i may be mistaken, but with widgets and "Web 2.0" apps, framesets are far less common then they once were...

Gregory, the question was about iframe rather than framesets

but yes, it is true

partly also because the design of framesets is so bad

<oedipus> ah, well, what used to be in IFRAME is increasingly stuffed into a proprietary mime-type

<oedipus> GJR: such as flash, etc.

yes, and using DOM manipulation

<oedipus> GJR: i am in agreement with StevenP and Alessio -- object rather than iframe should be the way to move forward

Roland: I would like to adopt the "sub URIs visible in the top-level URI" feature of XFrames into XHTML2 as a whole

Rich: I agree in simplification (reducing the elements in the language)

Yam: Drop the name Frame

Steven: Yes, maybe it sends the wrong message

Tomorrow's agenda

Roland: XML Events 2

<ShaneM> http://www.w3.org/MarkUp/2007/ED-xml-events-20071114/

<yamx> Thank you, Shane, for reference.

Roland: and its relationship to XBL(2)



We are cold, need to adjourn

Summary of Action Items

[NEW] ACTION: Roland to reply to issue 8001 [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action04]
[NEW] ACTION: Shane to add text explaining why some text is in orange [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action03]
[NEW] ACTION: Shane to upload final CURIEs draft [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action01]
[NEW] ACTION: Steven review CSS Namesapces module [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action05]
[NEW] ACTION: Steven to Draft last call announcement [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action02]
[NEW] ACTION: Steven to organise a new draft of XHTML2 [recorded in http://www.w3.org/2008/02/18-xhtml-minutes.html#action06]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2008/02/18 16:00:11 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.133  of Date: 2008/01/18 18:48:51  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/TH/Th/
Succeeded: s/ST/St/
Succeeded: s/Type/Type?/
Succeeded: s/experiemtn/experiment/
Succeeded: s/toins/tions/
Succeeded: s/available RDF/available/
Succeeded: s/jary/ary/
Succeeded: s/RO/Ro/
Succeeded: s/with/with /
Succeeded: s/WHen/When/
Succeeded: s/tq/t/
Succeeded: s/SH/Sh/
Succeeded: s/There are/What are the/
Succeeded: s/WD/WD?/
Succeeded: s/XFO/XFo/
Succeeded: s/section 29/section 28/
Succeeded: s/kd/ld/
Succeeded: s/suc/duc/
Succeeded: s/TO/To/
Found Scribe: Steven
Inferring ScribeNick: Steven
Present: Rich Roland Steven Yam Alessio Shane Gregory Gerrie
Regrets: Tina Mark
Got date from IRC log name: 18 Feb 2008
Guessing minutes URL: http://www.w3.org/2008/02/18-xhtml-minutes.html
People with action items: roland shane steven

[End of scribe.perl diagnostic output]