RE: alt use cases

David -

I think that is an excellent example, but I *believe* that it still falls
under what I am calling "CASE #5" (where the system does not provide an
@alt, or @alt is blank). But you do bring up a good point too, in that
sometimes, providing @alt is actually WORSE than not providing it at all.

I would really appreciate any feedback from others in the group regarding
this. Right now, all I have seen are a ton of examples bandied about, but no
actual analysis of those examples or distillation of them into a set of
abstraction that a spec can be written to. It's not like the spec is going
to say, "if the image a CAPTCHA, the @alt should indicate that it is a
CAPTCHA and how to use a non-visual version; if the image is a scanned in
document, @alt should be a full transcript", etc.



-----Original Message-----
From: [] On
Behalf Of Dailey, David P.
Sent: Monday, May 05, 2008 2:56 PM
To: Justin James;
Subject: RE: alt use cases

Justin James wrote:
>I think we need to take a step back, look at the use cases for alt, and see
where we need to go from here. 
While I probably don't agree with some of the comments you follow up with, I
do rather like the methodology you propose. It seems as though the arguments
have reached some sort of impasse -- if so this usually is a signal that a
new perspective is needed.

So this is not merely a "+1" sort of remark, I did have another use case for
leaving out alt tags for possible consideration:
the jig saw puzzle:
In either HTML or SVG, one generally (in the pre <canvas> world) uses a
stack of identical images relocated through script and overlaid so that
their clipping regions are mutually exclusive. If one were to appropriately
assign unique alt's to the various images as differentially clipped (for
example echoing the stacking order of the images inside the alt tags), one
would be providing a solution to the jigsaw puzzle (as in a paint-by-number)
which to sponsors of a friendly competition might be seen as an unfair
Such overlays of multiple instances of pictures might be found not only in
jigsaw puzzles, but also in interweavings of multiple pictures, morphings
between pictures, affine distortions of pictures,  indeed in many
script-driven image analytic fragmentations of images.  One can argue that
the failure to provide unique alts to each of a series of programmatically
overlaid images is  consistent with authorial intent rather than through
authorial neglect. Whence, if modern browsers choose to punish such pages it
will have been through their neglect rather than through intent of the spec.
In cases such as the above, it seems as though the meaningful place to put
an alt attribute would be in the superordinate heading (the <g> or the
<span>) containing the pile of otherwise identical images -- though in truth
I'm not sure what a screen reader would do with 1000 images stacked atop one
another, with different chunks of pigment revealed through clipping paths.

Received on Monday, 5 May 2008 21:18:51 UTC