[coyotos-dev] OSDoc issue

[Jonathan S. Shapiro]

On Thu, 2008-03-13 at 07:57 -0600, John B Halleck wrote:
> Much web based documentation I read hides the MathML by having something
> preprocess
> it into an image and text and puting the image with an alt tag giving
> either the
> MathML or a crude text rendering.  The screen reader gives out the alt
> tag, and sighted people see the image. (I think wikipedia does this?
> You might check to see what they
> use for tools.)

I considered this, and I came to two conclusions about it:

  1. The end effect of this is to transform something that might,
     with difficulty, be "read" by a sightless person (the mathml
     markup) into something that certainly cannot (the image).

  2. It complicates the tool chain significantly.

In and of itself, complicating the tool chain wouldn't be such a big
deal in my mind, but complicating the tool chain AND losing the
possibility of reading the stuff seems like a really bad thing.

But once again, I would really like input from people who live with the
issue. I can try to anticipate and respond to the issue, but it is not
possible for me to fully share their perspective.

> Thank you for the thoughtfullness of noticing that there was a handicap
> accessiblity issue here.  I spend too much of my life fighting to keep
> sites accessible.

May years ago I had an office mate at Bell Labs who was legally blind. I
think she could get four characters at a time using serious
magnification. She was quite possibly the best coder we had in that
department -- certainly among the top two or three, and it wasn't
exactly a group of nitwits. My older relatives have dealt with loss of
vision, and at one point we had a blind reader on one of the EROS lists
who was very helpful to me around these issues.

And of course, being diabetic myself, there is some reason for concern
that my own turn may come in time...


shap