Re: Table accessibility (was Re: headers attribute)

On Jun 2, 2007, at 1:48 PM, Maciej Stachowiak wrote:

>
>
> On Jun 2, 2007, at 2:18 AM, Sander Tekelenburg wrote:
>
>>
>> Yeah, would be wonderful to have actual data available.
>>
>> I've had a few opportunities to test with two blind Jaws users.  
>> For them,
>> headers and scope didn't help one bit. Probably because Jaws  
>> appears to take
>> IE's output as input, so that's useless by definition. But I've  
>> heard claims
>> that newer versions of Jaws consume the actual HTML. If so,  
>> there's hope yet.
>>
>> Note though that, at least for these users, configuring Jaws was  
>> an immense
>> challenge. So they may well have unknowingly had it configured to  
>> ignore
>> useful things. With my own eyes I've only seen Jaws 4 (quite old).  
>> Its
>> configuration area is pathetic. You actually need to be sighted to  
>> make sense
>> of it :( (This may be better in current versions of Jaws, I don't  
>> know.
>> Upgrading tends to be extremely expensive, because it will often  
>> mean also
>> needing a newer Windows versin, PC, speech synth and braille reader.)
>
> Finally, an actual test result! I also checked what VoiceOver (the  
> Mac OS X built-in screen reader) does, and it doesn't support  
> either scope or headers. I started this wiki page to record test  
> results. <http://esw.w3.org/topic/HTML/TableAccessibility>. So far  
> we have identified no screen readers with actual support for the  
> headers attribute, though I'm sure there are many more we could test.

With further research, I found more info about commonly used screen  
readers, added here: http://esw.w3.org/topic/HTML/TableAccessibility

It sounds like to do effective screen reader compatibility testing  
we'd need to check multiple versions of JAWS for Windows, HAL, and  
Windows-Eyes, and at least the latest version of VoiceOver. VoiceOver  
is free with the operating system, but the others are all very  
expensive so it may be difficult to get adequate testing.

I'm not sure what other screen readers are used widely enough to need  
testing.

Overall, it seems that testing support level of various HTML features  
in screen readers will be much more difficult than testing in  
standard desktop browsers or even in mobile browsers.

Regards,
Maciej

Received on Saturday, 2 June 2007 21:14:35 UTC