[turingtest] Inviting discussion on Anti-Robot tests

A white paper (draft) went up on the W3C Technical Reports page Wednesday.

  Inacessibility of Visually-Oriented Anti-Robot Tests: Problems and 
Alternatives
   http://www.w3.org/TR/turingtest/

Kelly Ford summed up the status in this are succinctly when he said

<quote
cite="mid:001301c3a4db$eaf2ce40$1e02a8c0@wingroup.windeploy.ntdev.microsoft.com">

This is worth a read if you haven't already explored the issue.  There's no
perfect solution to this problem today though in my opinion.

</quote>

So given that there is no perfect solution right now, can we do a little 
brainstorming?

Answers in any of the following areas are relevant:  Some points in each
area are already in the Working Draft.  Can you add any?  Use the keywords
in [brackets] below in your subject line for a jump-start framing your
comment.

[turingtest - squeaks]
More examples of where this kind of inaccessibility is creeping into practice?
What is it, precisely, that is problematical?

[turingtest - works]
What is out there and in use that partially works or works, but not well 
enough?

[turingtest - sky]
If you had a fresh sheet of paper and could design a solution from scratch, 
how should it ideally work?

[turingtest - steps]
What incremental changes would make things a bit better?

Al

Received on Friday, 7 November 2003 14:38:54 UTC