bharris at wikimedia
Jul 27, 2011, 4:19 PM
Post #12 of 13
On 7/27/11 3:49 PM, Maciej Jaros wrote:
Re: RFC: Modifying diff colors due to color blindness issues
[In reply to]
> I like the first red green (without the border), but if at it there
> should probably be more distinction between the grey, red and green
> colours. I was working on a picture on Commons some time ago after being
> poked by a colour blind person and found out that the problem was not
> the colour itself, but Value of colours (Value as in HSV model). So what
> I did is changed one colour a bit and made a greyscale image of the
> picture and I've then changed a bit more until I was able to
> differentiate the colours in greyscale mode.
> So in this case I would at least change the colours so that unchanged
> line (grey) would be different (in Value) from the changed line (red/green).
Picking the correct intensity for the colors is something that we can
deal with as needs be. The actual colors themselves are more important,
from a psychological point of view.
As I said in the bug itself, by using red and green we are
(accidentally) making a moral judgement about the revision. Humans are
hard-wired to see red as bad and green as good (blood vs. forest/life).
Revisions are, by their nature, amoral - so tagging one as "bad" and
another as "good" goes against our purpose.
Hence why I say they should be red/blue or green/blue. I originally
suggested red/blue because we want to know what was removed, but
thinking over it more I think it should be green/blue.
As far as the boxes go, I think they're essential. You should *never*
use color alone to indicate important information. Bolding the text and
increasing the saturation of the background color is a step, but it's
still possible for people with poor vision to mistake these things.
Macular degeneration can make distinguishing color variants difficult
(it sort of blurs the point of the eye's focus together), so including
express contrast elements (e.g., boxes or lines) helps to offset this.
While this thread is nominally about color blindness, I don't think
we'll go wrong by increasing scope to a larger set of visual problems.
(Note that we CANNOT and SHOULD NOT design for people with advanced
macular degeneration; this is beyond the scope of nearly any visual
design. But there are people who have similar problems [cataracts, for
instance] that lend to "lighter" versions of the same visual acuity
Protip: Photoshop has a way to show how things look to color blind people:
View -> Proof Setup -> Color Blindness (there are two options, you
should try both)
View -> Proof Colors
To return to normal, just unset View -> Proof Colors.
Fun facts: There are three major types of color blindness. Two are
red/green and one is blue/yellow. The blue/yellow (tritanopia) is rarer.
The *rarest* form of color blindness is called "monochromacy" - black
and white vision. It's *extremely* rare. . . unless you live on one
specific island in Micronesia.
(There's another kind of monochromacy but it's mostly found in people
with really, really bad eyes and its kind of the least of their problems).
(I know all this crap because I once worked for at a company where the
CTO was color blind. So it is fairly ingrained in me now.)
Wikitech-l mailing list
Wikitech-l [at] lists