[Irl-dean] Color contrast and WCAG 2.2 priority?

[Paul Walsh, Segala]

Color? 

Barry, shouldn't your spell checker correct you on the spelling of colour
(wink).

*Personally* I really dislike the threshold for these particular points,
they are way too strict for my liking. Moreover, they are more strict than
HP's guidelines (no time to dig out references sorry!), which has a lot of
experience in this area.

I'm sure Sorcha (specifically) or Gez can contribute to this particular
point given that they're both on the WCAG task force. Naturally, the w3c
members would have discussed this in details. Furthermore, the specification
is public and therefore open for review.

I'm afraid I'll have to ask Sorcha to keep her comments to a minimum as she
spends a lot of time (every week) replying to comments on the specification
itself, as I'm sure Gez does.

Kind regards,
Paul

-----Original Message-----
From: irl-dean-admin at list.eeng.dcu.ie
[mailto:irl-dean-admin at list.eeng.dcu.ie] On Behalf Of Barry McMullin
Sent: 17 January 2007 14:30
To: irl-dean at list.eeng.dcu.ie
Subject: [Irl-dean] Color contrast and WCAG 2.2 priority?

Hi Folks -

Another slightly "picky" question.  It's motivated by further
reading of the Nomensa UN report, but it stands separately from
it.

WCAG 2.2 reads:

  Ensure that foreground and background color combinations
  provide sufficient contrast when viewed by someone having color
  deficits or when viewed on a black and white screen. [Priority
  2 for images, Priority 3 for text].

Now there are various practical difficulties with evaluating
against this checkpoint (e.g., how much contrast is "sufficient")
but I want to focus here on a slightly more subtle aspect.

In the simplest case, of a HTML page with text correctly marked
up semantically and styled using CSS, this checkpoint only
applies at priority 3.  Presumably this reflects the fact that,
providing the text *has* been correctly, semantically, marked up,
then any user with a serious deficit in color perception can opt
to override the author CSS (with a CSS stylesheet tuned to their
own specific requirements).  Yes, it does impose some extra
effort on such users, hence the checkpoint is still applicable,
but it is not a "show stopping" barrier, so the checkpoint
applies only at priority 3.

Now consider an image.  By definition, we want to consider an
image that is conveying some significant information or function.
And we suppose that there is indeed poor color contrast in the
image, making this information or function potentially difficult
to perceive. On the face of it, WCAG 2.2 immediately applies; and
it is at priority 2, presumably because now the user does not
have a (relatively?) easy way to override this poor contrast,
since one cannot change the internal contrast of an image just by
tweaking the CSS.

But wait.  If the image conveys significant information or
function, then checkpoint 1.1 already requires that that
information also be conveyed in a text alternative. So let us
suppose that there *is* an appropriate text alternative
provided (typically via alt and/or longdesc attributes; but
possibly also via redundant textual content of the page).

*Now* how do we interpret the priority of checkpoint 2.2?

To spell it out, even though the image itself has poor contrast,
the same content/function is conveyed by text; and exposure (and,
indeed, styling) of this text can indeed be controlled (via CSS)
by the user. So it seems to me that we would properly conclude
that checkpoint 2.2 is still only priority 3, even though the
"primary" content was an image with poor color contrast.

I *think* this is a sensible and logical interpretation of WCAG
2.2 in this situation.  Perhaps more importantly, I *think* is
sensibly represents the practical implications for access by
people affected by color deficit.  But ... I would certainly
appreciate other views, on either of these opinions.

(I did do a quick search of some of the "likely" sources on the
web to see if there was any prior discussion of just this point.
I didn't find anything, but it does "feel" like something that
would surely have been discussed before.  So again, if anyone has
any references, that would also be great!)

Thanks - Barry.

_______________________________________________
Irl-dean mailing list
Irl-dean at list.eeng.dcu.ie
http://list.eeng.dcu.ie/mailman/listinfo/irl-dean

-- 
No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.432 / Virus Database: 268.16.13/632 - Release Date: 16/01/2007
16:36