Re: [html-techs-tf] caption vs alt

Hi Alastair,

 > Alastair Campbell wrote:
> Steve Faulkner wrote:
> "The situation is that in some circumstances the ability to provide a text alternative using the recommended method  (alt attribute) is not available to the author. This is often the case on photo sharing sites (for example). Quite often authors do have the ability to provide a caption, in this case provision of a text alternative via the caption mechanism is a suggested workaround. It is not ideal, but its the best that an author who wants to provide a text alternative can do under the circumstances."
>
> In that scenario we can assume the person adding the photo doesn't know what alt text is (or at least 99.x% will not).
> In which case, the text is unlikely to be an alternative. It is likely to be whatever witty thing the uploader can think of to spice up the picture, or it might be a label (e.g. the name of the people in the picture, like people used to write on the back of paper photographs).

Right. So the @alt text should provide an equivalent for the image - 
then the witty observation/comment can go in the <figcaption>

It is very unlikely to be something you could consider an alt text, 
which is why I'm struggling to see how a technique for fulfilling SC 
1.1.1 would work in practice, except as a get-out clause.
>
> Loretta wrote:
> "I have also seen situations where the caption for a figure says everything that a text alternative would say, e.g., a screen shot with the caption describing the significant content of the screen shot.
> Not every figure caption is an appropriate text alternative, but when one is, why repeat the information?"

Agreed, or if the ability to add @alt just isn't available then the 
<figcaption> may well end up being a combined description of the image, 
or parts or it - or a brief gestalt view - combined with witty comment 
and so on.

> I would agree, I just haven't seen that situation happen. Was it a particular scenario? I can imagine that happening in work-oriented documentation perhaps, or manuals, although I can't think of an example.
>
> Also, if the caption were suitable, wouldn't you still need to use a null alt text? (Or "See figure caption" as Gregg suggested.) Otherwise AT cannot know that the caption is equivalent and will probably read out the filename.

True, but that really is a user agent issue and as the use of 
<figcaption> gains traction new screen readers could suppress/modify the 
heuristic that will read the file name, in the presence of a populated 
<figcaption> element.

HTH

Josh

Received on Monday, 13 January 2014 11:04:46 UTC