Re: Investigating the proposed alt attribute recommendations in HTML 5

On Aug 29, 2007, at 21:48, Steven Faulkner wrote:

> Investigating the proposed alt attribute recommendations in HTML 5 - 
> http://www.paciellogroup.com/resources/articles/altinhtml5.html

Great work. Thank you. Based on your testing, it is clear that the  
current state of JAWS is so bad that indeed any generated placeholder  
alt text (even the empty string) is better than omitting it.

Back when I advocated for allowing the omission of the alt attribute  
when the markup generator does not have a textual alternative  
available, I based my argument on the behavior of Lynx (at least some  
version with some settings). That behavior is that alt='' suppresses  
the image altogether but the omission of the attribute causes a  
bearable placeholder to be presented so that the user knows that  
there's an image.

When making Web pages today, catering to today's JAWS, which  
apparently has unbearable placeholders, makes sense. It doesn't  
*necessarily* follow, though, that writing the spec to *require* (as  
opposed to *allow*) catering for the flaws of today's version of JAWS  
makes sense considering the entire life span of the spec.

What, in your opinion, is the outlook on JAWS ever getting fixed? (By  
"fixed" I mean to have image place holders that give a better user  
experience than alt="" or alt="image" or page content duplication in  
the case of a non-decorative image.) Should this WG expect that 7  
years from now, the market leader in voice browsing still hasn't  
evolved to have better heuristics to such extent that J. Random Web  
app developer can do better by putting together *some* generated alt  
text (even alt='', alt='image' or duplicating other data already on  
the page)?

(This is not a flame. This is an honest question. I admit that I  
don't understand the competitive landscape of voice browsing. I'm in  
awe that a product behaving like JAWS can be the market leader.)

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Tuesday, 11 September 2007 08:23:42 UTC