Re: 48-Hour Consensus Call: InstateLongdesc CP Update

Hi Silvia,

SEO, repeating the mistaktes, old aldvice doesn't die,

On Sun, 23 Sep 2012 02:41:56 +0200, Silvia Pfeiffer
<silviapfeiffer1@gmail.com> wrote:

> On Sun, Sep 23, 2012 at 5:51 AM, Laura Carlson
>> longdesc is experiencing increased usage in the wild
>> http://lists.w3.org/Archives/Public/www-archive/2012Sep/att-0025/comments-mt.html#increasedusage
[...]
> Interestingly, you data actually confirms the message that I just sent
> on another thread [1] about moving to @aria-describedat. I would think
> that all of these institutions that are supportive of accessibility
> and are using @longdesc in the correct way for this purpose would
> accept moving to use @aria-describedat if we encouraged them to do so.
> In this way, the vast pollution of @longdesc values that we see in the
> wild would be replaced by only clean and accurate use of
> @aria-describedat . It would make it easier for tools to identify
> sites that have appropriate and usable long text alternatives, since
> you can just search for @aria-describedat and know you will get good
> results, rather than searching for @longdesc and having to wade
> through vast numbers of misuses.

Actually, I doubt that will be true. There are several factors that I  
believe will lead to aria-describedat having significant pollution:

1. It is invisible metadata. While this is only an issue for content  
producers whose workflow isn't sufficient for quality output, there are  
many of those. It seems unlikely this problem will completely disappear.
2. SEO is big business. Many people would rather use accessibility  
features for SEO than for improving accessibility. As an example,  
http://dilbert.com seems to do this with the alt attribute for the comic  
of the day - and I believe this is still a common practice.
3. Longdesc isn't terribly complicated. The original description of  
longdesc for img begins:
"This attribute specifies a link to a long description of the image. This  
description should supplement the short description provided using the alt  
attribute..."
While it isn't a perfectly clear specification, I found it difficult to  
understand how reading it could lead to any major confusion. Yet  
demonstrably intelligent people who did read it have reported being  
confused, and it seems almost certain that some people used the attribute  
(and possibly even wrote guidance on using it that others read and  
believed) without actually reading it at all. These mistakes are eternal,  
and while they can be reduced with better naming, clear educational  
efforts, etc, it is very difficult to eliminate them.
4. The old guidance, and wealth of discussion, will hang around.  And  
people will find it, and follow it. Or continue relying on what they have  
already learned. Which will lead to some people using longdesc instead of  
describedat or whatever a replacement is called. While this isn't  
pollution per se, it is reducing the signal:noise ratio for describedat,  
by reducing the signal.

All of these factors can be mitigated somewhat, and I expect enhancing  
longdesc (whether we change the name or not) will lead to an increase in  
the signal:noise ratio. But I believe that there will be a lot of noise -  
just as there is with alt.

The alt attribute is almost certainly the best-known  
"accessibility-specific" (at least mostly, in popular perception) feature,  
and one of the most generally applicable to web content. After more than a  
decade of massive education, regulation and motivation for improvement, I  
would make a large bet that the majority of Web content has badly polluted  
alt attributes. In addition, many people (including me) believe that it  
has some nasty limitations that cause real-world problems. Yet I hear no  
serious suggestion to replace it, and believe there is strong consensus  
that it is very valuable despite the misuse and lack of use.

cheers

Chaals

-- 
Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
         chaals@yandex-team.ru         Find more at http://yandex.com

Received on Monday, 24 September 2012 10:59:34 UTC