[html4all] 5 gears in reverse - anne v k enters the alt attribute debate

[Charles McCathieNevile]

On Fri, 21 Sep 2007 13:37:26 +0200, Steven Faulkner  
<faulkner.steve at gmail.com> wrote:

> http://annevankesteren.nl/2007/09/alt
> Anne has done a post on the alt attribute
> It is interesting to note what he writes:
>
> "I am currently following HTML5 (omitting alt) as it wasn't really clear  
> to me what would be a better solution given the single constraint I
> have: not finding it necessary to provide replacement text for all those
> images.

I think this is a fair point. There is an interesting discussion to have.  
FWIW I think what this is correct in terms of using the technology (and  
correct according to ATAG), but I don't mind if it would not be considered  
valid HTML.

Frankly, validity is not important for itself, and having missing alt text  
flagged as a mistake that is different from alt="" is valuable. It means  
that nobody has thought about the question for this image - and while sad,  
it is better to know that than to pretend something else and then make  
decisions or interpret the content based on a false assumption.

> This would take too much time for little benefit."

It's sad that Anne doesn't see the benefit of deciding whether his images  
need some funtional replacement or are just additional to the text, but  
this is a common decision :(

It's not very difficult to imagine how to add something useful though. And  
a lot of people have solved that problem before.

> His statement appears to widen the debate and put forward what i have
> suspected is one of the key motivations for making the alt attribute
> optional, namely where an author does not *find it necessary* to make the
> images accessible.

Authors are not, in my experience of the web, good judges of whether it is  
necessary to have an alt, nor of the benefits. Nor, for that matter, of  
how to make god alt text - a lot of it is pretty ordinary at best.

On the other hand, nobody has come up with an alternative that can replace  
it (the figure stuff if worked out right might replace it *in some  
circumstances*) as far as I can see. Some people getting it right is  
better than none...

> He also misrepresents the meaning of alt="" :
>
> "it tells the user agent that the image means *nothing*. Nothing at all."
>
> when it can mean
> a. the image has no meaning (e.g. spacer gif)
> b. is decorative
> c. has meaning, but the meaning is provided by text elsewhere in the
> content.

Indeed. It means (when used correctly) that no text provides a functional  
replacement for this image - because for some reason it doesn't need one.  
It makes no claims as to why such text is not needed.

> he also misunderstands how AT UA's treat images without alt
> attributes, generally by default (unless the image is the sole content  
> of a link) they treat them the same as alt="", that is they ignore them.
>
> any thoughts?

This tool is pretty basic, and it makes assumptions that its user is too.  
If it is a bepoke tool for one person then the assumption might be valid.  
Whatever we think that says, with the context given by those assumptions  
it does the right thing. This example tool isn't the kind of thing I would  
show anyone as a real tool, except to ask for help in improving it, and I  
don't see how anyone can say it is any good for accessibility.

But this tool is almost irrelevant to the debate. The answer to the real  
question, which is whether  comes down to minimising the damage.

If alt is optional in HTML 5, but required by WCAG, then people who care  
about accessibility will do the 10 minutes work usually required, and it  
will be clear whether the rest of the people did that or just stuffed some  
random attribute in.

If alt remains required by HTML, we will see a small amount of  
auto-generated content (which is harmful) and some more people deciding  
that if you have to do it you might as well do it right. If, in addition,  
WCAG says that missing alt is less bad than wrong alt, then it will be  
more like if alt is optional.

I don't *know* what gives the best outcome :( It's pretty much a crappy  
choice. I don't see that auto-generated alt text is any better now than it  
was 7 years ago, while I do think that there is a lot more useful alt text  
around than I used to see 7 years ago - and I suspect in that time that  
the proportion of text-only browsers has gone down. So I don't *believe*  
that insisting on a tool putting "something" there is that helpful. I do  
believe that any tool that doesn't allow a user to say "leave blank" for  
alt="", something useful, or just skip it altogether and not add an alt,  
and make it easier to skip than to add a blank, is actively harmful to  
accessibility, but because alt="placeholder" is as bad as anything else  
possible, I am not convinced that tools must, in general, insist on the  
user putting something there since they will end up puttting their own,  
slightly harder to detect, dummy text.

cheers

Chaals

-- 
   Charles McCathieNevile, Opera Software: Standards Group
   hablo español  -  je parle français  -  jeg lærer norsk
chaals at opera.com   http://snapshot.opera.com - Kestrel (9.5α1)