W3C home > Mailing lists > Public > w3c-wai-ua@w3.org > October to December 1999

Re:Re: System API checkpoint issue

From: <schwer@us.ibm.com>
Date: Sat, 4 Dec 1999 10:28:47 -0600
To: menovak@facstaff.wisc.edu (mark novak)
cc: w3c-wai-ua@w3.org, ij@w3.org, jongund@ux1.cso.uiuc.edu
Message-ID: <8525683D.005AE3B3.00@d54mta08.raleigh.ibm.com>



>MN2:  my comment is correct at stated.  I felt your reply to Ian was
misleading,
>if Mozilla is truely using "standard" controls when on the windows
platform,
>that information is available thru MSAA, and *does not* require an
>Off-Screen Model.   MSAA can also provide more information than just the
>"text" for standard controls.  As for what is defined by "enough", I think
>that depends entirely upon a user's needs.

I am aware of the other features MSAA provides and from a purely academic
sense your comments are justified. However having actually written and
developed screen reader and screen magnifcation products I can safely say
that it is not enough. An OSM *is* required for non-standard controls.
Every screen reader out there today needs an OSM to provide access to
Windows and Windows applications. These comments are not ment to disparage
MSAA either, but they are to state facts. I would like to see MSAA be
extended to address the needs. Your comment on using a DOM is well
received. Microsoft's work on the DOM client access has come a long way.

Regarding custom controls you need to remember that I am not just talking
about simple file dialogs, menus, buttons, and list boxes, but I am also
addressing the entire client area which could be anything from a complex
doument to a spreadsheet.

As for the encouragement of cross-platform application developers to use or
not use custom contols, I think that from a developers perspective it is
unrealistic to expect them to use MSAA on Windows if they could provide a
cross-platform DOM-based solution designed to address common contols
access. Furthermore, we need to be aware of the fact that developmment
organizations are increasingly developing heterogeneous solutions. There
are tremendous business opportunities on Linux, Solaris, AIX, the Mac, and
pervasive devices that do not specify MSAA. As a result, groups like
Mozilla are creating cross-platform infrastructures. This alone is a major
development effort that to make extensive changes to their code base to
support MSAA is unrealistic. We need to ensure that they can make
cross-platform accessible solutions. If it is not there we need to build it
like we did with Sun on Java.

However, Java and Windows are not be-all and end-all solution. W3C WAI
design specifications need to be created to accomodate this. Basic access
system requirements like keyboard, mouse, and standard system API
mechanisms for drawing text and generating audio can be supported much more
easily.

I hope at this point I am preaching to the choir, and I apologize for
getting on my soap box, but in order for accessibility to successfully
incoroporated into the devlopment fabric we need to accomadate developers
business needs as well. Heterogeneous environments need to be addressed and
after all this is really what the WAI group has to address anyways. I hope
this clears up where I might have mislead people.

Rich



Rich Schwerdtfeger
Lead Architect, IBM Special Needs Systems
EMail/web: schwer@us.ibm.com http://www.austin.ibm.com/sns/rich.htm

"Two roads diverged in a wood, and I -
I took the one less traveled by, and that has made all the difference.",
Frost


menovak@facstaff.wisc.edu (mark novak) on 12/03/99 07:23:05 PM

To:   w3c-wai-ua@w3.org
cc:   ij@w3.org, jongund@ux1.cso.uiuc.edu, Richard
      Schwerdtfeger/Austin/IBM@IBMUS
Subject:  Re:Re: System API checkpoint issue




see comment at MN2 below:

>MSAA *does* rely on an offscreen model to provide access to text in
>documents and many other custom contols. For simple buttons and menu items
>it does allow you to get access to the text but that is not enough by a
>long shot.

MN2:  my comment is correct at stated.  I felt your reply to Ian was
misleading,
if Mozilla is truely using "standard" controls when on the windows
platform,
that information is available thru MSAA, and *does not* require an
Off-Screen Model.   MSAA can also provide more information than just the
"text" for standard controls.  As for what is defined by "enough", I think
that depends
entirely upon a user's needs.

Now, If you want to change that to "custom" controls on the windows
platform, I already
stated in an earlier email, that is a more general accessibility problem,
and
one which I hope *someone* is encouraging Mozilla development team to
avoid.  If the Mozilla development team is going to provide custom controls
that are not accessible to MSAA on the windows platform, then by all means
I
think it is also their responsibility to make sure those controls are
accessible.

As for accessing the text in doc., I wasn't implying to use MSAA to do
that.  I would
continue to use the DOM, for UAs, as you've stated.


> A miracle has not happened here yet.
>
>Rich

>menovak@facstaff.wisc.edu (mark novak) on 12/03/99 12:54:24 PM
>
>To:   w3c-wai-ua@w3.org
>cc:   Richard Schwerdtfeger/Austin/IBM@IBMUS, ij@w3.org,
>      jongund@ux1.cso.uiuc.edu
>Subject:  Re:Re: System API checkpoint issue
>
>
>
>
>see one comment at MN below:
>
>
>>I will check into this some more, but I am told that the DOM API
>>implemented by the Mozilla group is the W3C DOM Specification (at least
as
>>much as the IE is and perhaps more).
>>
>>However, I would strongly disagree that if a company is working to
develop
>>a consistent cross-platform solution based on W3C specifications that
they
>>are not compliant. Especially if they can document how an assistive
>>technology can connect with the user agent.
>>
>>If Mozilla is willing to create a cross-platform accessible solution
based
>>on W3C standards we should support them. The other problem you have is
>that
>>MSAA is still reliant on an offscreen model. A DOM-based solution should
>>not be.
>>Rich
>
>MN:    I need to do a lot of email reading to catch up to the root or
>source of
>this discussion, but one thing that caught my attention directly is the
>comment
>that "MSAA is still reliant on an offscreen model".
>
>Please note, MSAA *does not* require an Off-Screen Model for the
>information
>it provides.
>
>
>
>>Ian Jacobs <ij@w3.org> on 11/21/99 01:18:36 PM
>>
>>To:   Richard Schwerdtfeger/Austin/IBM@IBMUS
>>cc:   Jon Gunderson <jongund@ux1.cso.uiuc.edu>, w3c-wai-ua@w3.org
>>Subject:  Re: System API checkpoint issue
>>
>>
>>
>>
>>schwer@us.ibm.com wrote:
>>>
>>> Jon,
>>>
>>> I had a brief meeting with the IBM web browser team and we discussed
>>> Mozilla accessibility. Mozilla is designed as a cross-platform solution
>>> even though it is compiled for each platform. It turns out that all
>>> components including the application chrome can be accessed through
>their
>>> DOM. Our guidelines state that we need to use the system-provided
>>> accessibility features like MSAA however they also require that we use
>>the
>>> DOM albeit for the actual document.
>>
>>"Their DOM" is not the same as "The DOM" (meaning the W3C DOM).
>>Therefore,
>>despite the good design idea of making the system platform independent,
>>by not using system conventions or a recognized API, the design causes
>>assistive technologies to lose. Consistency among the interfaces
>>offered by the particular user agent across different platforms may
>>be less important than consistency among different pieces of software
>>on a given platform.
>>
>>Perhaps the checkpoints are flawed or behavior in the case of
>>overlap is underspecified. Consider these three requirements:
>>
>>  1) Implement system conventions (checkpoint 5.2 of [1])
>>  2) Implement the W3C DOM (checkpoint 5.6)
>>  3) Implement your own, accessible and open API (checkpoint 5.1).
>>      NOTE: I'd like to review what 5.1 means exactly.
>>
>>Mozilla seems to be doing 3 at the expense of 1. Is there a
>>way to map Mozilla's API to MSAA on Windows?
>>
>>This seems like a real issue where developer and AT input would
>>be very useful.
>>
>> - Ian
>>
>>
>>[1] http://www.w3.org/TR/1999/WD-WAI-USERAGENT-19991105
>>
>>> The conclusion of the meeting was that the User Agent guidelines should
>>> allow for cross-platform accessibility through DOM 2 as a minimum since
>>> this will utimately be a W3C standard providing the solution provider
>can
>>> clearly define how an assistive technology would interact with the DOM
>to
>>> provide an accessible solution. I believe that cross-platform
>>accessibility
>>> is a more important issue given that it can enable assistive technology
>>> solutions on other OS platforms and devices. Support for device
>>> independence and standard I/O API and all other requirements would
still
>>> apply naturally.
>>>
>>> I would like to raise this issue for the next meeting.
>>>
>>> Rich
>>>
>>> Rich Schwerdtfeger
>>> Lead Architect, IBM Special Needs Systems
>>> EMail/web: schwer@us.ibm.com http://www.austin.ibm.com/sns/rich.htm
>>>
>>> "Two roads diverged in a wood, and I -
>>> I took the one less traveled by, and that has made all the
difference.",
>>> Frost
>>
>>--
>>Ian Jacobs (jacobs@w3.org)   http://www.w3.org/People/Jacobs
>>Tel/Fax:                     +1 212 684-1814
>>Cell:                        +1 917 450-8783
>>
Received on Saturday, 4 December 1999 11:35:17 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:49:25 UTC