Re: [html4all] New issue: IMG section of HTML5 draft contradicts WCAG 1 & WCAG 2 (draft)

Katie Haritos-Shea wrote:
> 
>> There are people who upload literally thousands of photos a week. If a 
>> photo takes one minute to describe, which is probably optimisic, that's 
>> two full days' worth of work per week just to describe the photos. That's 
>> not happening. Even people who upload 10 photos a month don't care enough 
>> to describe their photos.
>>
>> You might be able to get some people to describe some of the most popular 
>> photos, but there's no way that's going to scale to all photos on all 
>> photo galleries, so the problem of what to do with photos that have no 
>> useful alternative text will always exist.
>>
>> -- 
>> Ian Hickson
> 
> 
> Which is a good reason why alt text needs to be required by this spec
> (HTML5).
> 
> People will continue to commit crimes and break the law........but it
> that a reason not to have them?

It might well be a good reason not to have a particular law if it is 
symptomatic of that law being a sub-optimal way of addressing the 
underlying problem that it tries to solve. See also [1] for some more on 
this and apropos examples.

In this particular case it's not entirely clear what effect changing the 
conformance requirements would have. For example, a validation warning 
might be just as effective as an error at reminding people that care 
about accessibility to include alt text, whilst preventing people who 
will just do the simplest thing to pass automated conformance checks 
from including spurious, unhelpful, values. As a side effect UAs 
implementing automated image analysis procedures such as [2] to try and 
fill in the gaps for their users would be more effectively able to key 
off missing alt as a sign that the automated analysis should be 
performed and empty alt as a sign that the image did not require 
alternate text.

I do not claim this is the only possible outcome of the spec text as it 
stand. But it is a fallacy to say that because a certain idea was 
conceived to fulfill a given use case we must not change the feature 
because the use case has not been addressed as successfully as we would 
like.



[1] http://lists.w3.org/Archives/Public/public-html/2008Feb/0255.html
[2] http://pubs.iupr.org/DATA/2007-IUPR-25Jul_1615.pdf

-- 
"Mixed up signals
Bullet train
People snuffed out in the brutal rain"
--Conner Oberst

Received on Friday, 11 April 2008 21:52:50 UTC