Thread Subject: Re: Users AT Definition

Note

This archival content is maintained by WebAIM and NCDAE on behalf of TEITAC and the U.S. Access Board . Additional details on the updates to section 508 and section 255 can be found at the Access Board web site.

From: Tom Brett
Date: Mon, Jun 18 2007 1:50 PM

The Section 508 coordinator's in the Government look at products and ask
were the made/coded to the Section 508 standards (are they conformant).
There are some coordinators who will actually look at the coding and perform
evaluations without the use of any AT. This type of evaluation determines
if the vendor has inserted code that force the product to work with only 1
type of AT. When it has been determined that the product does conform they
can state that the product meets the Section 508 Standards. The next level
of testing is done with AT. Many cases the AT needs to be tweaked to get it
to work with the product.



If the product conforms with the Section 508 standards, which are considered
to be technical, and the AT does not work with the product a conclusion that
can be drawn is that there is something wrong with the AT. I agree that
there are no AT-IT interoperability standards but AT follows documented
accessibility features provided by the platform.



Tom Brett



_____

From: = EMAIL ADDRESS REMOVED =
[mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of Randy Marsden
Sent: Saturday, June 16, 2007 10:52 AM
To: TEITAC General Interface Accessibility Subcommittee
Subject: Re: [teitac-general] Users AT Definition



Peter:

You wrote:

"... forced to do some amount of AT testing"

" ... safely blame the AT"

" ... conjure up a workaround"

I think you may be overlooking the option of actually working together with
AT vendors to ensure compatibility.

As far as your proposed definition goes, it pre-supposes a standard exists -
which it doesn't. If you say "code to the standards" - what standards do
you mean? Currently, there are no accepted AT-IT interoperability standards
- only 508, which is not a technical standard, and can't be "coded to".

-Randy
------------------------------------------------
Randy Marsden, P.Eng.
President & CEO, Madentec Limited
ATIA Global Policy Chair

780-450-8926 ext. 223
= EMAIL ADDRESS REMOVED =


From: Peter Wallack < = EMAIL ADDRESS REMOVED = >
Reply-To: TEITAC General Interface Accessibility Subcommittee
< = EMAIL ADDRESS REMOVED = >
Date: Fri, 15 Jun 2007 13:36:07 -0700
To: TEITAC General Interface Accessibility Subcommittee
< = EMAIL ADDRESS REMOVED = >
Subject: Re: [teitac-general] Users AT Definition



I'd propose the following wording to replace 'Users AT':

AT that is itself coded to be inter-operable with products coded to these
standards.

As for 'coding to the standards' vs. actual AT testing, that is indeed the
$1 million issue. Of course we want to simply code to a standard, just as we
do for any API. But the Functional Performance criteria require us to supply
a mode that actually works, so we are forced to do some amount of AT
testing. But of course we cannot test every possible AT. And if it doesn't
work, we then must make a judgment call: can we safely blame the AT (and be
prepared to explain it over and over and over), or should we conjure up a
workaround. The same issues occur when we test and find problems with the OS
and browser, but its a bit different because they are much rarer, and the
interfaces tend to be very well defined 'industry standard' APIs.
Peter Wallack
Accessibility Program Director
Oracle Corporation


Tom Brett wrote:

[teitac-general] Users AT Definition
The use of term such as reasonable, current and common in the standards seem
to be appropriate. I am concerned that the conversation tended to talk
about using 1 type of AT to test applications. Testing with AT is a good
final step in validating a product but the developer needs to be able to say
that they have followed the standards. I am aware of products that work
with Jaws but will not work with other screen readers or speech to text
technology. The vendor developed Jaws scripts to allow the product to work
with the product instead of following the Section 508 standards and then
says the product meets the Section 508 standards because it works with Jaws.




In addition, the AT vendors have a certain amount of responsibility in this
also. These vendors need to develop AT that will use commonly available
features in the operating system and programming language. For example, if
AT does not read a properly coded alternative value of an image (alt=) the
problem is with the AT not the EIT. The EIT has been written to the
standard. There is nothing more that the developer can do.



There are cases when the AT just does not work. About 3 years ago the makers
of a popular screen reading software made a mistake and removed the support
for JAVA in their product. The screen reader application did not work with
Java applications. The problem was not with the software application
because, presumably, it was written to the Section 508 standards.



If applications and systems follow the technical standards but the AT does
not work the problem is with the AT not the product. It may be necessary to
tweak the AT to make it more compatible with the network architecture and or
the product. Using any type of AT, screen reader, head mouse, blow tube,
speech to text, etc. to test a product to determine if it conforms to the
Section 508 standards is not good practice. This is particularly true of
products that are available to retrieve public information by members of the
public. The Government cannot say to a member of the public that the AT they
are using was not used in the testing of the product so you will need to
change.





Tom Brett




_____


From: = EMAIL ADDRESS REMOVED =
[mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of Randy Marsden
Sent: Tuesday, June 12, 2007 10:04 PM
To: TEITAC General Interface Accessibility Subcommittee
Cc: Jessica M. Brodey; David Dikter
Subject: [teitac-general] Users AT Definition



Hi General Folks:

I apologize neither Jessica nor myself were able to attend yesterday's
General Subcommittee meeting to represent ATIA. I have read the transcript
and see it would have been nice to have had one of us there for the
discussion about defining what "Users' AT" means.

Here are the points I would have made, had I been on the call:

1. I hear what Phil (and others) are saying about wanting to be able to
design to a defined interface, and then testing with select AT to make sure
it works. Then, if it doesn't work with other AT, the responsibility should
be on that other AT as IT must surely have done their job by writing to the
spec and conducting limited successful testing. (I hope I have paraphrased
his comments correctly - if not, please let me know).


MY RESPONSE: unfortunately, no such interoperability specification exists
today. It would truly be nice if one did; life would be easier for both IT
and AT, and almost certainly better for the end-user. But it doesn't exist
yet, so I don't see how we can make it part of what we're doing now.
Secondly, even if such a standard did exist, I would be very hesitant to say
all IT has to do is meet it and they're good to go. In practice, there is
almost always "tweaking" that is required for true compatibility - even when
interoperability standards and API's exist. (This is true with any standard
- for example, we have found a wide range of different current capabilities
on USB ports of different computers, notwithstanding a well-defined
specification).

If the IT company tests with an AT product and finds it to work, and then
tests with a second AT product which doesn't, it doesn't necessarily mean
the second AT product has done something wrong. Using my example, perhaps
the second AT product requires a higher current (but still within spec) than
the first AT product, which the IT product is failing to supply. Or
perhaps the second AT product really does have a problem. It could be due
to any one of a number of things - all of which would likely be much easier
to fix with collaboration between the IT and AT companies.

For many years, AT operated pretty much in a "collaboration vacuum" with IT
(with notable exceptions). One of the primary benefits of 508 is that it
encourages AT and IT collaboration (like what we're doing right now). As
ATIA, we are leery of anything that may take away that benefit (such as
saying all IT has to do is meet a yet-to-be-defined interoperability
specification, and they're done).


As far as the definition goes, I still think we have to keep it vague, and
then provide ancillary notes to help provide direction. We talked about
using words like "reasonable", "readily available", "current", and "common".

-Randy

------------------------------------------------
Randy Marsden, P.Eng.
President & CEO, Madentec Limited
ATIA Global Policy Chair

780-450-8926 ext. 223
= EMAIL ADDRESS REMOVED =






_____

WebAIM is an initiative of:
Center for Persons with Disabilities (CPD) Utah State University