[html-techs-tf] caption vs alt

Hi everyone,



I'm looking into an HTML5 accessibility technique [1] for the Figure & Figcaption to meet SC 1.1.1, but I'm not sure it should exist?



I assume the technique should be for the part of the HTML spec "4.7.1.1.16 When a text alternative is not available at the time of publication" [2]



However, if we're writing a technique that is trying to meet SC 1.1.1, I don't think applying alternative text within a figcaption would be suitable. If you are adding alternative text, shouldn't it be on the alt attribute?

I don't see how a technique could be applied to a situation where you can't add something.



Closely related to this are the "HTML5: Techniques for providing useful text alternatives", 4.2 [3], which provides two examples that are intended to be equivalent:



<img src="example.jpg" alt="Shadow like figures and a graffiti tag drawn on the walls of a partially demolished building, illuminated by the light from a street lamp.">



And then:



<figure>

<img src="shadows.jpg">

<figcaption> Shadow like figures and a graffiti tag drawn on the walls of a

 partially demolished building, illuminated by the light from a street lamp (photo). </figcaption>

</figure>



The idea being that the caption is equivalent to the alt, when the alt is missing.



However, I think this is fundamentally at odds with the intent of alt text: to be for people who can't see the image.



A far more realistic example that I could see regular authors doing is assuming that people can see the image:

<figure>

<img src="shadows.jpg">

<figcaption>The local graffiti artists have been busy!</figcaption>

</figure>



But what we would want is examples more like 9.1 [4], which would be something like:

<figure>

<img src="shadows.jpg" alt=" Shadow like figures and a graffiti tag drawn on the walls of a

 partially demolished building, illuminated by the light from a street lamp..">

<figcaption>The local graffiti artists have been busy!</figcaption>

</figure>



I think it comes down to what relationship we think figcaption should have to the image. I had thought it would be a similar relationship to aria-describedby, which adds description but is not the label itself.



I can see figcaption having a technique under SC 1.3.1, but should we have a technique under SC 1.1.1?



-Alastair



1] http://www.w3.org/WAI/GL/wiki/Techniques/HTML5

2] http://www.w3.org/html/wg/drafts/html/master/embedded-content-0.html#when-a-text-alternative-is-not-available-at-the-time-of-publication

3] http://www.w3.org/TR/html-alt-techniques/#m6

4] http://www.w3.org/TR/html-alt-techniques/#hbe