Call for testing: colorblindness vs side-by-side diff

[Matt Mackall]

On Tue, 2015-01-06 at 16:21 -0500, cowwoc wrote:
> On 05/01/2015 3:18 PM, Matt Mackall wrote:
> > (FYI, there are 7 distinct categories of colorblindness of varying
> > frequency and degree, so a sample size of one is not very useful.)
> >
> > The thing I tried to convey to you when you brought this up the first
> > time was that _we've already done this_. And we got significantly more
> > input than you have and arrived at the current result.
> >
> > Yes, some people are unhappy with the result. That's unfortunate, but
> > that is going to be the case with every design decision. If we accept
> > your scheme, a different set of people are going to be unhappy[1] with
> > the new result and lobby to switch and so on forever.
> >
> > At some point, we have to make a choice and stick with it. And as far as
> > I'm concerned, that point is already in the past. So it's really
> > doubtful this is going to go anywhere, especially since few people seem
> > to care this time around.
> >
> > [1] Starting with me, because ugh are you serious those colors? [2]
> > [2] https://en.wikipedia.org/wiki/De_gustibus_non_est_disputandum
> >
> Hi Matt,
> 
> I get the feeling that there isn't much anyone could present to you at 
> this point which would make you consider changing the color scheme. Am I 
> wrong?
> 
> While it is true that in the case of taste there can be no right and 
> wrong, it should be relatively trivial to put up a SurveyMonkey poll 
> asking people to vote for their favorite colorblind-approved scheme and 
> go with the majority vote.

Let's imagine we do this. Now the question becomes: what do we do the
NEXT time someone wants to change the colors? Because there will always
be users unhappy with the colors.

Should we repeat the whole process? How about the time after that? At
what point should we say "no, damnit, we've already done this damn
thing, why are we still wasting time on this?"

I submit to you that a trivial application of mathematical induction
reveals the optimal time to say that is right after the first time. And
that first time, as I've mentioned multiple times now, HAS ALREADY
PASSED.

> Alternatively, would you consider offering separate colorblind and 
> non-colorblind color schemes?

That's a really hard sell to a colorblind project leader, yo.

But here's something I'd consider: I hover over the color key and my
pointer changes. Huh. I click on it, and the color scheme cycles through
several different sets that nobody needs to spend any time arguing about
since they're not the default.

-- 
Mathematics is the supreme nostalgia of our time.