Re: Is there any guidance on CSS keywoard 'float'?

[distribution note: this reply drops w3c-wai-cg to delve into the 
substance of the issue.]

At 8:19 AM -0500 5/10/05, Jon Gunderson wrote:
>Peter,
>The web content accessibility guidelines (WCAG) would not deal
>with the browser functions for caret navigation of content,
>unless the caret was being controlled by a web application
>creating its own user interface.  I beleive you are looking
>for guidence in the behavior of caret navigation when some
>elements are being graphically positioned using either CSS
>position or CSS float properties. 
>
>Recommendation
>
>In this case my recommendation is that the caret should follow
>document order.  The caret may seem to "jump" to another part
>of the document using document order if the author CSS is
>enabled, but I think it provides the best compatibility since
>some users will turn off stylesheets or apply their own
>stylesheet.
>
>Coordination
>
>But I will bring this up at our next WAI Coordination Group
>(CG) meeting to see what WCAG and PF have to say about the
>issue.  I beleive the next CG meeting is May 18th.
>
>Jon

While moving through all the text in the textual order in the
HTML page is sufficient to give some means of access to all
the text, it doesn't seem like what the mobility-impaired user
looking at the layout would most want.

Let's take a step back and let me try to understand the issue.

Caret navigation has up-a-line, down-a-line motions that
relate to the graphic text as laid out, not the abstract text
structure pre-layout in the HTML.  Maybe moving the caret
in and out of a floated box should take some note of the
layout, too.

How does navigation relate to frames at this point?

In other words, to move from frame to frame, can you do this within
caret navigation, or is another kind of navigation required?

The reason I ask is that a floated text box is probably logically a
sidebar, or at least a distinct functional unit with respect to any
text in the enclosing box.

On the hypothesis that the text in the floated box is a sidebar, and
that sidebars are concurrent, that is to say have no logical order
relation with regard to the main flow, but are accessed first or
second at user option; it would seem to make the most sense to treat
a floated or absolute-positioned box as a strong cell wall and
require special action to jump that wall,  The caret would normally
give precedence to following the text flow on one side or the other
of the wall.

I would like a more extensive description of the functionality that
is not present when there is a floated box with text in it.  Are there
clauses in the requirements document that fail to be satisfied in
some way for text in a float-positioned box?

Does the caret move across the box boundary when you don't want it
to? Does the caret never make it into the box?

[if the latter, it sounds like you are not getting the resolved
values of the bounding box for the float after positioning. This is
not a problem of policy for UI behavior but data access witin the
browser.]

I would think that the requirements would at least be to follow the text
within a flow.  *A float is not in the flow[1].*  There should also 
be a requirement
to navigate among distinct flows, but that may be in the UI but not
part of what you are calling caret navigation.

[If it is not and there is some indirect way of entering the float, such
as a graphical move of the cursor, it could make sense that this
verb include a snap-caret-to-cursor constraint.]

Please give a clearer picture of what functions you think should be
supported that are not supported, either with the current implementation
or with the current design.

Al

[1] http://www.w3.org/TR/REC-CSS2/visuren.html#floats



>---- Original message ----
>>Date: Mon, 09 May 2005 18:33:03 -0700
>>From: Peter Korn <Peter.Korn@Sun.COM> 
>>Subject: Re: Is there any guidance on CSS keywoard 'float'? 
>  >To: Jon Gunderson <jongund@uiuc.edu>
>>Cc: Ginn.Chen@Sun.COM, Peter Korn <Peter.Korn@Sun.COM>
>>
>>Hi Jon,
>>
>>Quoting Ginn Chen from Sun Beijing from the internal-to-Sun
>bug report on
>>Mozilla difficulties with http://news.yahoo.com/asiadisaster:
>>
>>--------- begin quote ----------
>>
>>The css class .ynelhalfblock is using float:left.
>>We're not supporting Caret Browsing with CSS Keyword 'float'.
>>ginn.chen@sun.com 2005-2-25 03:07:57 GMT
>>
>>In http://news.yahoo.com/asiadisaster
>>'<style type="text/css"><!--@import
>>url("http://us.a1.yimg.com/us.yimg.com/lib/hdr/ygma5.css");
>body{margin:0px
>>4px;}--></style>'
>>In http://us.a1.yimg.com/us.yimg.com/lib/hdr/ygma5.css
>>'#ygma #ygmaproplinks span{padding:22px 0 0;float:left;}'
>>
>>According to Mozilla/Gecko Keyboard Navigation Proposal
>>http://www.mozilla.org/access/keyboard/proposal#X.
>_Unresolved_Issues
>>
>>VIII. Unresolved Issues
>>
>>     1. How to access fixed/floating widget using keyboard?
>>        Some widgets (most of them are ads) can be put
>anywhere in the web page
>>by specifying the absolute position attribute in CSS style.
>Their physical
>>position may be very different from their logical position.
>>        NOTE: We didn't find anything in Web Content
>Accessibility Guidelines
>>2.0 discussing using fixed/floating widget in web pages.
>>ginn.chen@sun.com 2005-03-09 06:10:01 GMT
>>
>>--------- end quote ----------
>>
>>Does this answer your question Jon?  If not, perhaps Ginn can
>answer it more
>>directly.
>>
>>
>>Peter
>>
>>Jon Gunderson wrote:
>>>  Peter,
>>>  I am not sure what CSS keyboard 'float' means. 
>>>
>>>  There is a CSS float property for positiong content, but it
>>>  does not have a keyboard  featue.
>>>
>>>  Could you give me a little more explaination of the
>behavior
>>>  and the code you think is causing the problem.
>>>
>>>  Jon
>>>
>>>
>>>  ---- Original message ----
>>>
>>>>Date: Mon, 09 May 2005 10:04:09 -0700
>>>>From: Peter Korn <Peter.Korn@Sun.COM> 
>>>>Subject: Is there any guidance on CSS keywoard 'float'? 
>>>>To: Jon Gunderson <jongund@uiuc.edu>
>>>>Cc: Peter Korn <Peter.Korn@Sun.COM>
>>>>
>>>>Greetings,
>>>>
>  >>>The website http://news.yahoo.com/asiadisaster (among
>others)
>>>
>>>  uses the CSS
>>>
>>>>keyboard 'float', which Mozilla keynav is unable to deal
>with.
>>>>
>>>>Do you know if the CSS accessibility group provides any
>>>
>>>  guidance on this?
>>>
>>>>
>>>>Thanks,
>>>>
>>>>Peter
>>>>
>>>
>>>
>>>
>>>  Jon Gunderson, Ph.D., ATP
>>>  Director of IT Accessibility Services
>>>  Campus Information Technologies and Educational Services
>(CITES)
>>>  and
>>>  Coordinator of Assistive Communication and Information
>Technology
>>>  Disability Resources and Education Services (DRES)
>>>
>>>  Voice: (217) 244-5870
>>>  Fax: (217) 333-0248
>>>
>>>  E-mail: jongund@uiuc.edu
>>>
>>>  WWW: http://cita.rehab.uiuc.edu/
>>>  WWW: https://netfiles.uiuc.edu/jongund/www/
>>>
>>>
>>
>>
>
>
>Jon Gunderson, Ph.D., ATP
>Director of IT Accessibility Services
>Campus Information Technologies and Educational Services (CITES)
>and
>Coordinator of Assistive Communication and Information Technology
>Disability Resources and Education Services (DRES)
>
>Voice: (217) 244-5870
>Fax: (217) 333-0248
>
>E-mail: jongund@uiuc.edu
>
>WWW: http://cita.rehab.uiuc.edu/
>WWW: https://netfiles.uiuc.edu/jongund/www/

Received on Tuesday, 10 May 2005 22:25:31 UTC