Thank ethan I would really like to see your colors used in our environment. Now I think that we have to work on the mapping to constructs. BTW we had fun on (not so related) microprints (was just an experiment: http://scg.unibe.ch/archive/papers/Robb05b-microprintsESUG.pdf Stef On Apr 2, 2011, at 5:34 AM, Ethan Schoonover wrote:
Tudor Girba <tudor.girba@...> writes:
the red in @param in java does not work so it is strange
The red has been revised for beta2 (it's the only color I had an alternate for -- two alternates actually, one of which was further tested and will be the final color). The old red was L value 45 and was the only accent color that dark. Red is very challenging in general (particularly given the contraints I set in order to toggle light and dark) and I have truly spent *days* testing reds for this scheme. No kidding. I dreamt Lab and hex values. The new red is L 50 which translates to sRGB space as a much less saturated value.
I tested it with Sim Daltonism: http://michelf.com/projects/sim-daltonism/ It does pretty well (except for Protanopia).
I'll take a look at that tool as well. For what it's worth, I did do some testing (loaded the scheme into colorschemer and simulated color blindness) as well during the development of Solarized. What I finally came back to was ensuring that L*a*b lightness was accurate and consistent.
My assumption was that L lightness would translate to an equivalent L lightness (regardless of *a*b perceived value) in most color blindness, but this may not be an accurate assumption. This would at least retain readability. I try to limit colors to "value added" accents only.
Feedback from those interested in this issue specifically is welcome. I'm sure I could do some more work in this area.
Best, Ethan Schoonover es@ethanschoonover.com