Re: HTML Action Item 54 - ...draft text for HTML 5 spec to require producers/authors to include @alt on img elements.

Hi Andrew,

On May 13, 2008, at 12:53 PM, Andrew Sidwell wrote:

>
> I have reservations about adding to this thread, but so it goes...   
> (I fear I'm making the same point Dave Singer rather excellently  
> made.)
>
> Steven Faulkner wrote:
>> dear maciej,
>>> It is not a matter of opinion. Making a use-case non-conforming is  
>>> by
>> definition not handling it for purposes of document conformance. It
>> may be a >conscious choice to reject a use case, but it is not
>> support.
>> you assume without justification, that it is desirable or right to
>> accommodate use cases that  result in important data and/or data
>> relationships not being provided. i don't subscribe to this.
>
> I would be happy if someone (or several someones) in favour of  
> making alt mandatory in all cases would answer very simply: How does  
> a blind photographer mark up a photo, which is known to be critical  
> content, but which she herself cannot describe?

According to the new draft section, the alt attribute is not to be  
used for description of photographs that are critical content. You're  
thinking of the current editor's draft that attempt to expand the alt  
attribute to cover many more accessibility functions. However,  
according to best practice recommendations descriptions — including  
descriptions of photographs — should be handled through other means.  
For a blind author they would likely not even think in terms of using  
an image of rich text or an iconic image or a chart to convey their  
meaning. Rather they would work in the opposite direction which many  
here could benefit from doing. In this case, they would start from the  
issue of conveying meaning with text. A graphic designer might then  
take that recurring text and apply an icon or rich text to it. A blind  
graphic designer might draw from a library of existing icons  
(previously described and stored in the files metadata). Likewise a  
blind graphic designer might make the text into rich text beyond the  
capabilities of present-day (or legacy) CSS. However, the necessary  
alt text is already available to the blind author and the blind  
graphic designer author because the meaning is conveyed in the the of  
the document: it merely needs to be moved to the alt value of an img  
element that embeds the rich text or icon.

On the other hand, the description of the image is not a requirement  
nor a proposed requirement here. The description falls in the SHOULD  
or MAY language (perhaps "authors SHOULD provide a descriptive title  
and MAY provide further subject and visual descriptions of images  
referenced by the longdesc attribute).

So to summarize critical content text alternative is not a description  
of an image. It's the necessarily brief text that would be required  
for a user to comprehend the document in the absence of the image.


> Is it:
> <img src="photo">
> <img src="photo" alt="Photo">
> <img src="photo" alt="Exposure 2s, f/12">
> or something else?

Something else (a photo will rarely require anything but null alt):

<img src='photo1' alt='' longdesc='descriptions#photo1' >

or something else (for an icon that draws the visual user's eye):

<img src='important-item-icon' alt='important item' >

or something else (for rich text such as text written on a path or  
text painted with a color gradient or text with an affine  
transformation):

<img src='InvitationSalutation' alt='You are cordially invited' >

or something else (a chart intended to visually convey the breakdown  
of electricity production among various methods):

<figure>
<legend>The breakdown of electricity production in Illinois</legend>
<img src='InvitationSalutation' alt='Nuclear 50%, Coal 30%, Natural  
Gas 15%, Other 5%   ' >
</figure>

In all of these cases the blind author has all of the information  
necessary to construct conforming alt text (according to the newly  
drafted section).

The rest of your questions I'll leave for you to answer. As you can  
probably see now, they were based on a fundamental misunderstanding of  
the alt attribute.

Take care,
Rob

> Furthermore, should a blind photographer who cannot provide alt text  
> be able to author conforming HTML5?  (I ask this without prejudice;  
> maybe the answer should be no.)
>
> This is essentially what "handling the use case" would consist of.  
> Guidance needs to be given in cases like these-- "cases that result  
> in important data and/or data relationships not being provided",  
> whether you want to accommodate them within the spec or not.
>
> Why?  Because there *is* content with no textual alternative and  
> there will be whether you say there should be or not.  The  
> interesting question is not "should alt be mandatory?" but "how do  
> you mark up critical content with no known alt text?".  Saying "but  
> alt text should be available" does not change the fact that  
> sometimes, it is not.  With me?  Just because you set up your  
> goalposts such that the interesting question is outside them does  
> not remove the question.
>
> As a photographer, I do not have the time or the money to pay  
> someone else to have the time to provide anything resembling  
> alternate text for 99% of my photos.  I am quite willing to produce  
> non-conforming HTML5 if alt was to became mandatory in all cases,  
> but the question remains-- how should I mark up the photos I don't  
> have the resources to provide alt text for whilst still saying that  
> they are critical content?
>
> (Call me someone with disregard for disabilities if you like, but I  
> see putting photos online for those who can view them analogously to  
> showing my friends prints of photos in real life.  IRL I don't write  
> alternate text for every print I make in case someone who can't  
> adequately view the photo wants to know what it's about, and so I  
> don't expect to do the same online.  I would be happy to describe a  
> photo in such an instance, but it's something I would do on-the-fly,  
> and not at upload-time/print-time.)
>
>
> Andrew Sidwell
>
>

Received on Tuesday, 13 May 2008 13:53:43 UTC