W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 2001

RE: Cognition Simulation

From: Charles F. Munat <chas@munat.com>
Date: Fri, 24 Aug 2001 14:56:39 -0700
To: "WAI GL" <w3c-wai-gl@w3.org>
Message-ID: <LHEGJAOEDCOFFBGFAPKBIEEKCJAA.chas@munat.com>
This is an interesting exercise. One problem: by using a multiple choice
test, you lead the user to the answer. A better test would be to simply ask
"what is the above guideline telling you to do?" Leave the answer blank. At
the bottom, you can provide the actual text of the guideline so people can
check their answers.

Without previous knowledge or answer prompts, I might have answered thus:

What does Guideline 3.4 ask you to do?

Go to the left if you're a man and to the right if you're a woman.

or maybe

Go to the left if you have square shoulders and to the right if you have bad
posture.

The point is, illustrations can confuse the issue as easily as they can
clarify it. Adding the audio equivalent is a much better idea. I would guess
that all hearing, English-speaking readers would get the question right
after hearing the audio equivalent (assuming they could make out Charles'
funny accent).

Having a checkpoint about adding illustrations makes little sense to me
unless there is also a checkpoint saying that we should make illustrations
(and other "text equivalent content") clear and simple. Why is it only text
that needs to be clear and simple?

One benefit of illustrations is that they can be of help to people who don't
speak the language. But this is not as simple as it sounds! I think that
telling people to add illustrations, etc. to their sites willy-nilly without
serious instruction on HOW to do it will probably result in LESS accessible
pages (especially given bandwidth restrictions -- there is nothing less
accessible than a page you never see because you couldn't wait for it to
download).

Like Kynn's informal bias experiment, this test is dangerous because it may
lead us to believe that we understand something that we really have no clue
about. These sorts of tests/examples should really be formulated by
researchers who are a) trained to do this sort of work, b) following an
accepted procedure, and c) testing, testing, testing.

Chas. Munat


> -----Original Message-----
> From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-request@w3.org]On
> Behalf Of Anne Pemberton
> Sent: Friday, August 24, 2001 1:47 PM
> To: Charles McCathieNevile; WAI GL
> Subject: Cognition Simulation
>
>
> Folks,
>
>          The first rough and unpolished version of a Cognition Simulation
> is now available at:
> http://www.erols.com/stevepem/Guidelines/Cognition.html
>
>          Of course, y'all will recognize Chaals' illustration in it ....
> interesting .... it's isn't possible to right click on a *.png image and
> download it. I had to link it to Chaals' site ...
>
>          Come to think of it, I should link the sound file to it
> as further
> illustration to help make the point.
>
>          Anyone else have suggestions, comments, etc? I'm most curious if
> some of you think it does or doesn't do what it's intended to do
> - simulate
> what it's like to be faced with an incomprehensible page.
>
>                                          Anne
>
> Anne Pemberton
> apembert@erols.com
>
> http://www.erols.com/stevepem
> http://www.geocities.com/apembert45
>
>
Received on Friday, 24 August 2001 17:54:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:12 GMT