I agree with the key points that access for blind users is necessary (and not
necessarily too difficult), and this would be of benefit for all users. Text
descriptions would have great utility ( whether presented visually, auditory, or in
Braille), however a text only browser is not a must, .
Chris's idea of modifying the parser to output (like Gavin suggested) a simple list
of links and associated text descriptions is right in line with what we have been
thinking of. I would like to discuss the modified parser idea further. This would
quickly provide functional access. Additionally, if the underutilized "Info" and the
new "ASCIIText" nodes could be fed into this list. All of this would serve in the
same way that HTML "ALT" does.
Looking towards 1.1, the introduction of sound will cause a similar problem. If
world designers include information in some audio format, it will again be
necessary to have a text description. Not only would this be of benefit for persons
with hearing impairments, but it would be necessary for people using VRML in
situations such as libraries or noisy environments. So, would it be cleaner for the
spec to have separate text descriptions for visual, auditory, and eventually haptic
elements of a given VRML world. Or, would it be best to have one general textual
description (with some indication of which modality this text is a substitution of)
node, or complementary file type, for all text descriptions?
Chris
ATRC -- University of Toronto
[email protected]