411914 - Please implement RELATION_FLOWS_TO/RELATION_FLOWS_FROM

Status: RESOLVED WONTFIX

(Core :: Disability Access APIs, defect)

Description

[Joanmarie Diggs]

Examine the first section of the following page in Accerciser:
http://www.mozilla.org/projects/firefox/3.0b3pre/whatsnew/ 

The following is the accessible hierarchy with the spatial order in parentheses:

I. Heading (Upper left)
        A. Link
II. List (Lower right)
        A. List Item (fifth item)
                1. About Link
        B. List Item (fourth item)
                2. Developers Link 
        C. List Item (third item)
                3. Store Link 
        D. List Item (second item)
                4. Support Link 
        E. List Item (first item)
                5. Products Link 
III. Form (Above the list)
        A. Section (The items in the form are in expected order. Yea!)
                1. Label
                2. Entry
                3. Push button

When attempting to present the current line to a user, it is helpful to have some way of knowing what text "comes next" without having to examine the extents of each accessible.

Comment 1

[alexander :surkov (:asurkov)]

Would be nice to update link on testcase ;)

Comment 2

[Joanmarie Diggs]

Shall a open a new bug on that one Alexander? ;-) ;-) ;-)

I've since come across another example of the same issue but which perhaps highlights the problem/need a bit better:  If you take a look at http://www.bbc.co.uk/radio4/ you will see "Programme A-Z" followed by a map with links from A through Z.  Unfortunately, due to the markup of the page, the hierarchy gives us the links as Z through A.  

Comment 3

[alexander :surkov (:asurkov)]

Yes, please, mark this one worksforme and file another one.

Ok, I got you. Since iirc our point is do not broke DOM hierarchy (have accessible tree similar with DOM tree rather than frame tree) then I think we should expose relations.

Comment 4

[alexander :surkov (:asurkov)]

Attachment: testcase (based on bbc.co.uk)

testcase based on bbc.co.uk, contains two areas 'a' and 'b'.

Comment 5

[Aaron Leventhal]

This is going to be a guess on our part right?

Is there a conflict between guessing and the exactness of using aria-flowto?

Comment 6

[alexander :surkov (:asurkov)]

(In reply to comment #5)
> This is going to be a guess on our part right?

Yes by analyzing frame tree.

> Is there a conflict between guessing and the exactness of using aria-flowto?
> 

aria-flowto should overrite it I think

Comment 7

[David Bolter [:davidb] (NeedInfo me for attention)]

Mass un-assigning bugs assigned to Aaron.

Comment 8

[Raphaël POITEVIN]

What is the impact for a blind user? Is it just performances problem or do we really see inverted lists?

Comment 9

[Alex ARNAUD]

Dear all,

Sorry for the wrong need info request.

We ask for help of Joanie to know what is the outcome of this bug.

Best regards.

Comment 10

[Joanmarie Diggs]

For now I'm going to close this as WONTFIX. Pretty sure I'm working around it in Orca. And if I'm not, I can reopen when I have time to provide a patch for consideration.