[whatwg] ALT and equivalent representation

[Benjamin Hawkes-Lewis]

Shannon wrote:
>  To make
> matters worse some browsers display the alt tag while waiting for images 
> to come from the server and this creates visual artifacts that designers 
> and clients generally consider undesirable.

That's a feature not a bug. Many users are on slow connections.

> The end result of this is that alt tags tend to be seen as a burden by 
> the majority of web designers I've met. 

Yes, web designers generally have very little attachment to quality.

> Also these are not 
> government sites or contractors with mandated accessibility, and as far 
> as I know there is no law requiring corporate sites to provide 
> alternative text for blind users.

IANAL, but laws aimed at businesses tend to talk in terms of prohibiting 
discrimination rather than drilling down to the techniques used to 
provide equivalent access (which is a good thing, since it allows 
techniques to improve). Whether there's any sort of legal obligation to 
provide text equivalents depends both on your jurisdiction and how 
crucial the alternative text is to the general accessibility of the 
service provided by the site. See for example:

http://www.jimthatcher.com/law-target.htm

http://www.jimthatcher.com/law-target2.htm

http://juicystudio.com/article/web-accessibility-dda.php

> The ONLY "business" justification I have for using alt tags is that a 
> w3c valid site REQUIRES them and this may increase the sites Google rank 
> (which is just speculation really).

I'd have thought an art gallery would benefit from getting their images 
into image search engines.

http://googlewebmastercentral.blogspot.com/2007/12/using-alt-attributes-smartly.html

> I think this is a case where logic must give way to corporate 
> consideration, as public and charitable sites would probably use alt 
> tags without being told, but 95% of the mainstream internet will not - 
> given half a chance.

The rationale for making alt optional in certain cases is to increase 
accessibility in those cases. Now I don't really buy the reasoning for 
those cases, but I do not think the spec should sacrifice the 
accessibility of those cases (however it's best provided) on the altar 
of evangelising providing text equivalents through a validator.

--
Benjamin Hawkes-Lewis