W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 2003

Re: Untestable success criteria

From: Doyle Burnett <dburnett@sesa.org>
Date: Mon, 24 Nov 2003 10:39:33 -0900
To: <gdeering@acslink.net.au>, Joe Clark <joeclark@joeclark.org>, W3C Web Content <w3c-wai-gl@w3.org>
Message-ID: <BBE79065.14B0%dburnett@sesa.org>

To The Group -

As an individual with color deficiency issue - red/green, blue/yellow and
some other visual conditions that limit color combination identification -
I'd like to chime-in on this topic.  I believe full-well that authors can
come close to  determining color combinations that work "okay" for most web
users but also believe it will NEVER be perfect.  Color values can be
machine testable but it seems unlikely that running such a test will solve
ALL the probable vision issues related to color and color combinations.

Then there is the issue of how colors render with different computer
systems, monitors, etc.  I recall a Photoshop expert saying, how different
colors were on Macs compared to PC's  - had to do with light intensity as I

I am only saying what I am because of my own experience with color
deficiency and the fact that these issues are so specific to each individual
that science DOES NOT have all the answers.  We need to remember, it's NOT
just about color - it's about the overall ability to perceive visually of
which color is a part of that perception.

I am not agreeing or disagreeing with anyone - only stating my views as a
person who has SIGNIFICANT issues related to color issues (again, not JUST
the color aspect). 

Doyle Burnett

Doyle Burnett
Education and Training Specialist
Multiple Disabilities Program
Special Education Service Agency

On 11/23/03 3:39 PM, "Geoff Deering" <gdeering@acslink.net.au> wrote:

> Joe Clark wrote:
>>> All I am doing is making a point that colour combinations are machine
>>> testable.
>> Not without having the author manually type in every colour value on a
>> page.
>> You are working from a serious misconception, and, like Ridpath, this is
>> a bone you seem unwilling to drop from your tightly-clenched jaws.
> There is no misconception here, and you do not need to type in values
> manually.  You can parse both style sheets and pages.  With style sheets
> you can identify the color and background-color and compare their
> compatibility.  If you have been informed otherwise you have been
> misled.  If you have a programming background you would know that this
> is not difficult.  Trying to apply it to HTML soup is another question
> all together, but to well formed markup it is not a problem.
> There is software that detects skin tone within images using this approach.
> Geoff
Received on Monday, 24 November 2003 14:37:47 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:46 UTC