RE: CFC re ISSUE-31 Missing Alt

Actually a crowdsourcing mechanism is probably outside of the scope of the HTML specification. What this is about is defining the opportunity for such a mechanism to exist by defining a mechanism for stating @alt-not-asserted. That’s all we need to add to the spec if we deem that it is acceptable for AT to try by whatever means to infer an alt text.

I guess a browser or AT would be at liberty to scour the image metadata as much as it would be to apply a crowdsourced definition, or both, since the AT knows that the author didn’t apply the alt, it can inform the user as to the potential deficiency in the located text(s).

You missed the point of the example, what I was trying to get at in that case was that the alt is missing. But depict a case where the image metadata contains a better source of a replacement, being as it was supplied by the photographer or agency who knew the context of the image capture; than a crowdsourcer who only has the information in the image to go on.

-----Original Message-----
From: Laura Carlson [mailto:laura.lee.carlson@gmail.com] 
Sent: Wednesday, April 28, 2010 1:32 PM
To: Sean Hayes
Cc: Janina Sajka; Matt Morgan-May; Dave Singer; Gez Lemon; HTML Accessibility Task Force; Gregory J. Rosmaita
Subject: Re: CFC re ISSUE-31 Missing Alt

Hi Sean,

> I have a concern about the crowdsourcing idea, I think it suffers from 
> some of the same issues that chaals objected to in using metadata from 
> the image. If the author doesn’t include the contextual reason the 
> image is included in the page, then a 3rd party has to second guess 
> the authors intent for that image use. There may be many 
> interpretations of that image, but only the author can know the 
> 'right' one.

Thank you very much for expressing this concern.

I agree that the author has the best know chance of knowing the true intent of why an image is used. But sometimes they even they don't know. I teach a module on writing text alternatives. Many don't have a clue of why they are including an image. I have to drag it out of them. Crowdsourcing may help to teach the masses how to write text alternatives.

What happens when an author refuses to supply any text alternative after being prompted? ...After a conscious decision has been made by the author to deliberately publish images without text alternatives?
The crowdsourcing idea is a repair technique to help mitigate damages for when the author did not (and does not intend to) provide any text alternative. Ian calls it "when images are not known" in the spec.

Gregory, a blind photographer, has tried to crowdsource text alternatives on a number of his photos on his blog. A built in HTML5 mechanism would facilitate this.

I guess the question we need to answer is:

"Is crowdsourcing a text alternative after all else has failed, better than no text alternative?"

> If it's OK for a 3rd party Crowdsourcer to add "picture of black curly 
> haired dog on white background" to an image catalogue, when the 
> authors intent was "President Obama's dog Bo getting his first taste 
> of snow on the Whitehouse lawn", then it should be OK for the image 
> metadata to suffice too, since the metadata of an AFP/Getty image for 
> example is far more likely to contain the latter than the former.

The missing/crowdsourcing mechanism wouldn't apply to text alternatives that are not missing.

Best Regards,
Laura

--
Laura L. Carlson

Received on Wednesday, 28 April 2010 12:45:51 UTC