W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > July to September 2003

Re: JAWS for Windows for FREE? Lets give a try!

From: Matt May <mcmay@w3.org>
Date: Tue, 2 Sep 2003 19:29:19 -0700
Cc: 'W3C-WAI-IG List' <w3c-wai-ig@w3.org>
To: Phill Jenkins <pjenkins@us.ibm.com>
Message-Id: <6BAECFE4-DDB6-11D7-9785-000393B628BC@w3.org>

On Tuesday, September 2, 2003, at 03:50  PM, Phill Jenkins wrote:
>> As long as these folks don't consider what goes into their
>> text-to-speech engine to be a trade secret, I think this
>> is a reasonable (and minimally-invasive) request to make.
>
> Matt, a screen reader is made up of several main parts, namely the
> text-to-speech engine, the screen reading engine and the user 
> interface.
> What the screen reader does to create the string of text that goes to 
> the
> text-to speech engine is the key value of the screen reader.  It is the
> largest part of the value of the product.  It is the part that does 
> have
> many patents - not the actual text, but the algorithms that process the
> DOM, MSAA, and/or off-screen model and then creates the strings of 
> text.

All I was asking for was the actual text, formatted in HTML, for 
debugging purposes. Effectively, a screen shot for speech. That would 
help most of the problems authors would encounter, understand, and be 
inclined to fix: alt text and other equivalents, table linearization, 
forms rendering, frame interaction, and scripting faults (if 
applicable).

Over and above that, authors should be going out and testing with 
actual users of AT, not going it alone. If they don't know how people 
use an AT, JAWS et al. will be useless to them, whether they bought it 
or got it for free.

> All of these strings of output also depend on the user's input, which
> navigation keys they just pressed, what mode or configuration they are 
> in,
> where did they just come from, etc.  So it's not as simple as your post
> implies.  Yes we should consider making the text-to-speech engine 
> optional,
> but the user interface and screen reading engine are required.

Clearly it's not that simple. I can't design the system; I can only 
frame the problem.

And yes, configurations are different from system to system and install 
to install. But only the most dedicated of designers is going to test 
more than one AT, much less various configurations. They won't know 
enough about how the ATs are configured to do meaningful testing work 
anyway.

When it happens that a given design doesn't work with a given AT, the 
failure is not necessarily on the part of the designer, particularly 
when the designer followed WCAG in good faith.

>> But I can assure you none of them are going to give away thousands of
> copies of their software.
>
> Actually the screen reader vendors do basically that, in time-out 
> versions.

That's not the same thing. It's good that you all do it, but it's not 
what the petition is asking for, which is to write off the whole 
package as goodwill, rather than provide a time-limited demo.

> What I think is needed more than free screen readers are example web 
> pages
> (test suites) that screen reader vendors, authoring tool manufactures, 
> web
> accessibility checkers/repairers, and web developers can compare 
> against.
> The techniques are a start, just too small of fragments.  We need page
> examples of a few different versions of navigation, forms, tables and
> combinations of these that make up web applications used on today's 
> sites.
> What I don't see is a model (business incentive) for anyone to produce 
> or
> contribute it to the W3C.

I'm not concerned with the business model for producing a test suite. 
If the AT vendors will agree to implement to it, I'd be inclined to 
write it myself.

For the time being, they have the User Agent Accessibility Guidelines 
test suite[1] (which covers assistive technologies) to check their 
performance against HTML 4.01. There are hundreds of tests to run 
through already.

[1] http://www.w3.org/WAI/UA/TS/

-
m
Received on Tuesday, 2 September 2003 22:29:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:14:10 GMT