FtF XHTML2 WG, Cambridge, MA, USA

8 Nov 2007


See also: IRC log


Steven, Rich, Roland, Raman, Tina, Shane, Oedipus, Alessio




<scribe> Scribe: Steven


<Tina> Good [insert suitable time of day here], everyone.

Hi there Tina

We are just starting

Agenda topics:

* Strategy

* Meetings

* Documents


Roland: The group is called XHTML2, and our primary delivery is XHTML2
... what are we trying to achieve?
... I want to suggest we focus on authoring, and not rendering
... and look at a pipeline from authoring to rendering on lots of devices
... with personalization, device independence
... etc
... being applied from the authored materials to the user experience

Steven: Now that we have the split and the chartering behind us, it gives us the opportunity now to focus on the h\igher-level stuff
... and HTML can still remain the assembly language of the web
... and we don't have to worry about it

Raman: Like we did with XForms

Roland: We don't need to think of the page as the unit of a document either
... with mashup ideas in mind

<Tina> I believe we should focus on creating a language with which structure and semantic interpretation can be encoded - rendering ought be a non-topic in achieving this goal.

Raman: I would like to avoid SOAP and WSDL if possible in the solution space

I agree Tina

Raman: So let us focus on authoring, delivering clean stuff

Roland: If we consider our world supporting mashup styles, whatever our unit of work is, we can consider the broader ecosystem such as security
... It allows us to focus on *intent*, and therefore we can ditch h1-h6 for example, and just use the context
... and remove some of the clutter that is still in the XHTML2 spec for historical reasons

Steven: Sounds excellent

Roland: But we are still talking about compound documents
... if we need mathml, svg, it is still composable
... and one other thing, I would like us to try and deal with the namespace ugliness

Raman: Given that XML Schema doesn't define the root element, we could actually define some new root elements
... such as <web>
... we can avoid XML Schema
... we can alias names where needed
... to make authoring easier

Roland: This is a distinct subject in its own right

Raman: We need to protect the author from the ugliness

Rich: I was talking with Dave Raggett about components

Roland: We need to talk about levels of abstraction

Steven: I think we need to consider XBL at some level as part of the infrastructure for XHTML2

Welcome Shane

skype available if you want SHane


Raman: If we do the authoring layer cleaner, then we can later use XBLn to implement

Rich: Can you still render to the web?

Steven: Yes, in fact ebay.co.uk are already soing this. There are at least two companies using XHTML2 to adapt to devices

Rich: And can we add time?

Steven: I went to a meeting with SYMM this week, and they demo'd a page with XForms and SMIL working together with the Forms driving the SMIL presentation
... that's the advantage of the W3C architecture
... SYMM are the domain experts in time, and we can use thier stuff with ours

Rich: THis makes our stuff much more useful
... because it makes it much easier to write to


Steven: Mark Birbeck has recently been suggesting making an XHTML version of XBL, with xhtml as the container

Raman: Then we could use XPath selectors

Steven: Which fits better into our architecture, since we use XPath already

Roland: So we are going to get over our bipolar disorder

Steven: Well we have been trying to make everyone happy in a bimodal world
... and you end up making everyone unhappy
... now we can concentrate on being clean

Roland: We need a technical strategy, but also a communication strategy

Rich: Change naming?

Roland: Maybe in time

Rich: We need to talk to companies to ask what they need

Raman: We must avoid making it too big
... talk 1-1 with them is fine though

<ShaneM> are there people who think that the current XHTML 2 strategy is NOT sufficient? XForms + RDFa + cleaner, more consistent markup?

<ShaneM> as far as I am concerned, XHTML 2 in its current form is ready to go.

I think we can publish

<Tina> We must also be careful regarding talking to companies. Listening to their ideas, certainly, but not to include willy-nilly whatever they might desire.

but we are discussing cleaning it up even more than it is

+1 on that Tina

<ShaneM> okay. I ahve no problem with that.

<ShaneM> sorry I missed the earlier discussion. my week is pretty messed up

We are going to talk about publishing strategies later

<Tina> Will we include name-concepts in that discussion? It really is needed.

Rich: Why did EBay go with this?

Roland: Single authoring

Steven: Time magazine use XHTML2 as well for similar reasons
... they have one central format that they can extract the information from to deliver in as many other forms as they like

Raman: Let's take some use cases, and look to see if anyone thought we needed to add anything, but also what we can now safely take out

Rich: There is also Dita

<ShaneM> lots of people have asked for hr back

Roland: A publishing format

<ShaneM> and br, fwiw.

Name concepts Tina?

Roland: There are two directions, more radical, and less radical.
... The less radical is XHTML 1,2, which is the combination of what we have now,
... allowing XForms in 1.1 for instance
... rdfa, role etc

Rich: I'd like to show these slides on Dita (which is being standardised by OASIS)

[no URL I'm afraid, sorry people off line]

<Roland_> http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=dita

Rich: They have topic maps in the language

Raman: We need to keep our focus sharp, so we shouldn't add topic maps, but let people import them if they want

Steven: But looking at this slide, I see something that you can do with RDFa

<Rich> Topics:

<Rich> <topic id=�abc�> <title>My topic</title> <body><p>And so on</p></body></topic>

<Rich> Maps:

<Rich> <map id=�xyz�> <topicref href=�abc.dita#abc�> <topicref href=�def.dita#def�/> </topicref></map>

<Rich> Specialization to create new types of topics and maps:

<Rich> <topic> ? <task>

<Rich> <map> ? <bookmap>

<Rich> Defaulted class attribute maps specialized elements to ancestor ones:

<Rich> <context class=�- topic/section task/context�>Here�s why</context

Steven: a span with rel="topicref" and the href

<scribe> ... done deal, that's the glory of RDFa

Welcome oedipus

<oedipus> just got your invitation -- thanks!

<oedipus> thanks

you might want to see the minutes to catch up

<Tina> Steven: yes. At the moment there exist confusion as to the *name* of the thing - aka XHTML 2. Sorry I'm late.

<oedipus> that's what i'm doing

Yes Tina, that is a future discussion, but we're leaving it as-is for the time being

<Tina> That's fine, as long as the discussion is had. The way it is going now will lead to some chaos, if earlier experiences is any indication.

[Rich's slides even mention RDFa as a potential application]


Welcome Alessio!

<oedipus> shane, reading minutes shouldn't the horizontality or verticality of a separator be a styling issue / controlled by stylesheets

<oedipus> no need to bring back HR

<alessio> hi all, I'm on a train :)

<alessio> so connection could be not stable


<ShaneM> oedipus: of coruse it should. people asked about other extensions. Just playing along.

<oedipus> RDFa is very attractive to a few of us in PF (Protocols & Formats), not just RichS


It got some good comments yesterday at the TP as well

<ShaneM> hell, I even have uses for it.

<oedipus> lol

<ShaneM> its a sign of the apocalypse

<oedipus> hope it's only the sixth seal

Roland: Delivery context is an important part of what we need to do for adaptation
... so we need to look at what DIAL has been doing
... CDF framework for eventing

<oedipus> intrigued by ideas expressed about an XHTML version of XBL, with xhtml as the container

Yes Oedipus, a pity that MarkB isn't here

I hope he will join at some point to give us more detail

but the idea is just a mapping using selectors, to other actions

and that is easy to represent using XHTML2

we have all the infrastructire to achieve the necessary effect

Roland: So Raman, you were talking about XPath
... DIAL uses XPath as well

<oedipus> ah... i've already talked with the XBL people about assisting in the XBL2 Primer where there is a placeholder for accessibility enhancements via XBL

Steven: And XForms
... It's a shame that the CSS selectors are not more coherent
... XPath selectors are clearly more consistent

<oedipus> pity, too, that CSS selectors aren't more coherently implemented, either...

Steven: I suggested when I was in the CSS WG that they ought to design the whole set right at the beginning, even if they didn't include them all in the first version

<oedipus> XForms and XPath in an XHTML version of XBL sounds like the best of both worlds, for what that is worth

Offsite people, we are going to talk to UWA people shortly

just so you understand if we go quiet

We have an appointment at 3

(in 6 mins)

<oedipus> camcorder captured ARIA demo for HTML WG at:

<oedipus> quick question to anyone who has an answer: is XProc a viable means of at least managing mash-ups?

Oedipus, can't reach the site

<oedipus> Steven, neither can i - the IP address is that of the router in the HTML WG conference room

<oedipus> actually, since i couldn't make it in person, i've been able to attend multiple meetings simultaneously

good work!

<ShaneM> ROTFL

<oedipus> it literally said "versioning, dropped"

<oedipus> marcos is going to send me a copy of the ARIA demo and i'll put it someplace publicly accessible


We are off for coffee, and then UWA (on #uwawg)

We'll be back later in this channel

<ShaneM> roger

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2007/11/08 20:07:22 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.128  of Date: 2007/02/23 21:38:13  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/iedas/ideas/
Succeeded: s/RO/Ro/
Succeeded: s/TH/Th/
Succeeded: s/TH/Th/
Succeeded: s/RO/Ro/
Succeeded: s/SI/Si/
Succeeded: s/redical/radical/
Succeeded: s/aplicaiton/application/
Succeeded: s/RO/Ro/
Found Scribe: Steven
Inferring ScribeNick: Steven
Present: Steven Rich Roland Raman Tina Shane Oedipus Alessio
Agenda: http://www.w3.org/2007/11/08-xhtml-irc#T18-35-56
Got date from IRC log name: 8 Nov 2007
Guessing minutes URL: http://www.w3.org/2007/11/08-xhtml-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]