WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Minimal style needed to make links accessible?

for

From: Kroon.Kurtis
Date: Aug 26, 2010 2:27PM

For the most part, these tools measure contrast between foreground and
background ... which is less useful when you're comparing two _text_
colors viewed against the _same_ background.

If we rely strictly on the test results, fully-saturated blue (#00f) and
black (#000) fail to provide enough contrast. And that's 100% true:
blue text on a black background (or vice versa) indeed fail to provide
enough contrast, _when one is set on the other_.

Consider, however, that this blue is the "classic web link color" ...
and most web-using folks have no problem recognizing it as link text.
Instead, I propose that we evaluate color _difference_. The W3C
considers this to be just a part of contrast, and have removed it from
their officially-recommended algorithm. As a result, the tools I tested
don't report it. (I can't download software, so I can't test the
Paciello Group's tool ... can anyone tell me -- privately, so we don't
clutter the list -- if it does?)

For myself, I set up a spreadsheet in Excel to calculate color contrast.
I created it before the W3C changed their algorithm, so it reports color
differences. However, it uses the old success criteria, which failed
any score lower than 500: the blue/black example would still fail
because it only scores 255. Not a problem: color difference is no
longer normative, so I can redefine success to mean anything that suits.

I'd say anything with a color difference greater than about 200 should
be okay -- purple (#800080), the "classic visited web link color",
scores 256 ... and if it's darkened to the nearest websafe color (just
for grins), it scores 204.

Just something to consider ...

Thanks!

Kurtis Kroon
Web Business Services
State of California Franchise Tax Board
<EMAIL REMOVED>