Cross cutting ARIA property issue discussion
17 Oct 2007

See also: IRC log





Can we agree on a way to adopt role between xml markups that will also be consumed by HTML and how would this be consumed by SVG?

How can we support base ARIA role values without requiring a namespace (a request of the HTML working group)?

Would the use of aria-property be acceptable for xhtml and html and how would this be consumed by SVG?

How do we support role values in xml-based markup and html based mark up?

<hsivonen> I got the note

<scribe> scribe: Rich

use of aria-property

Doug: can use aria namespace to reference aria

henri: authors need to know how to implemnt aria support in svg browsers/plugins

Doug: the place to state this is in ARIA and this could be coordinated with SVG

henri: If svg agreed that all aria properties are defined in the aria spec. then are ok.

Doug: by putting aria- as the prefix you are actually name spacing
... we don't need to import the complexity of namespaces in an aria spec to address namespaces
... there is no complexity that is solved

Anne: You can use aribitrary prefixes

Aaron: you cannot use colons for attribute selectors in Internet explorer

Doug: I just want to establish that we are using namespaces no matter what. Can we agree on that?

Henri: I think you are confusing the issue.

Doug: I am clarifying it

Rich: you are just saying that we are using namespacing by even using aria- even though we are not using the namespace syntax

Doug: yes you are avoiding using the namespace syntax.
... the only thing old validators know about are using the : prefix
... anything that is w3c specified would have a problem w/o the :

Anne: you said it would break all user agents if you used aria-.

Doug: we would require buy in by all the vendors. Not saying we can't but it is a big job.

Henri: I can support either the : or - syntax in xml. But if are using svg in text/html using the : way and addressing all the namespaces that come with it we will have a problem since the - or : is invalid

Doug: I did not mean validators. There are other things, besides validators, which would also need to come into line.
... There are things out there that generate SVG and integrate SVG into other namespaces. For example, docbook uses SVG. SVG is included in othter documents like military drawings.
... it is not a simple matter of these things changing.
... we just state that we ignore unknown attributes.

Henri: if the SVG tools followed the rules we are discussing this would not be an issue.

<hsivonen> (that was Anne)

<hsivonen> (I agree, though)

thanks henri

Doug: I can see where it is possible to include aria- as a special way of namespacing going forward.
... to be clear, there was a statement before that arbitrary namespaces will work in svg and that is not true.

Henri: we should be using relax schema technology

Doug: I just want to make sure that it was not a correct statement about namespaces.

Anne: svg 1.2 does not use dtds
... browser vendors don't use dtds

Doug: someone made a statement about the flexibity of the - syntax over the : syntax allowed you to process arbitrary strings and that is not true
... it is stated in the xml namespaces spec. that you can use the : prefix.

Anne: you can fix it in the dtd and it is pre-declared.
... the whole point of namespace is that you can use whatever you want
... using a - instead of a : makes it easier to integrate in html going down the road.

Henri: because we want the html parser to parse existing content and we want new stuff to not break in new browsers we cannot make : act any different in current browsers
... if we are going to use svg in html in IE it will be much easier to do it without :

Rich: so the problem is that IE does not handle the :

Henri: IE does something special with the : in that the conformant DOM. It is a mess.

Aaron: I just wanted to use aria: and I could not use a CSS rule to address namespaces
... you can use | instead of : in CSS to reference prefixed properties in IE. The hyphen works fine in CSS, the content, everywhere.
... there is a problem which CSS introduced and it is in a problem with IE

Doug: hyphen is not a good delimiter as it is used throughout SVG.

Aaron: where is - a problem in SVG

Doug: |-width
... you can use presentation properties.

Henri: this is not a problem as we are not trying to introduce it as a namespace delimiter.

Doug: there is a lot of working going on with other web application formats
... we want for othter groups to import aria into their specs.
... my point is we want open and closed formats to be able to use.

Aaron: I think Doug is saying that he is open to using hyphen but he wants to go through all the thoughts about it.

Doug: yes.
... I am open to using a - and I am still not convinced that : can be used in html
... does aria work now in IE?

Aaron: with a - you can can trigger dynamic updates but without (:) you can
... you have to work around IE problems when you develop cross-browser script
... you don't get anything automatically and people describe what they are doing.
... aria does not drive the behavior

Doug: The - syntax is problematic with multiple user agents as it should be more generic than just aria

Rich: what about underscore?

Doug: I proposed if SVG were to import role we will accept this. but we were then split

Aaron: please propose aria- or aria_ as a compromise to SVG


Doug: SVG does not care if role is in or out of the xhtml namespace.
... we don't want to get attacked on role
... the people who are implementing svg tiny would be a problem. adding role would require going back to last call
... we could add role as part of the core language in a future release.

Henri: both role and aria need to be allowed on all elements

Doug: I think these are an orthogonal issues

<hsivonen> (that was Anne again :-)

henri: I think versioning could address how aria is done

<hsivonen> rather, I think versioning should get in the way of doing aria

<hsivonen> shouldn't

<hsivonen> (I don't really believe in Web language versioning)

Doug: if we had a chart syntax ... these charts have these values is great

aaron: lets work together and get compatible.

Doug: having fixed namespaced prefixes with _ is a good way to move forward

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2007/10/17 14:34:18 $

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/puttin/putting/
Succeeded: s/henri: using a - instead of a : makes it easier to integrate in html going down the road. /anne: using a - instead of a : makes it easier to integrate in html going down the road. /
Succeeded: s/deliminter/delimiter/
Succeeded: s/this is/these are/
Found Scribe: Rich
Inferring ScribeNick: Rich

WARNING: No "Present: ... " found!
Possibly Present: Aaron Doug Rich aaronlev anne henri hsivonen
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

Got date from IRC log name: 17 Oct 2007
Guessing minutes URL: http://www.w3.org/2007/10/17-aria-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]