mpl and colorcet names about colorcet HOT 4 CLOSED

It’s even more confusing since both inferno are listed under https://bokeh.github.io/colorcet/#Testing-perceptual-uniformity. BTW, I should note here that I don’t agree with this sentence “Here the inferno map seems to have better discriminability than the viridis map, despite both being perceptually uniform.”. Maybe an issue with screen calibration? Or just a difference between peoples.

But I’m not sure how much an issue this is, and how it might be fixed: changing the name means breaking compatibility.

from colorcet.

ping @jbednar

from colorcet.

Strange, I don't seem to have gotten a notification for there being an issue on this repo! They don't come up often, so I haven't been checking explicitly. Probably got caught up in my filters for the overwhelming number of bokeh-related messages. :-)

Anyway, when used with mpl, the colorcet maps are all prefixed with "cet_" (i.e., "cet_inferno"), to avoid any ambiguity. But I'm happy to change the name "inferno" to the more literal "bmy" to avoid confusion. It's true that it breaks compatibility, but this is precisely the sort of issue that made me call the version 0.9 rather than 1.0 -- I wanted a chance to fix things. Anyone have any other thing they want to change before officially declaring it 1.0? E.g. the "blues" colormap could be confused with "Blues" from matplotlib?

As for the statement "the inferno map seems to have better discriminability than the viridis map, despite both being perceptually uniform", I just looked at that section on three different monitors, and asked a friendly nearby graphic designer to do the same, and in all three cases for both observers we found the "comb teeth" extending further down for "inferno" compared to "viridis". So yes, screen calibration or observer could be an issue, but the claim is true at least for these three very different looking monitors and two very different observers...

from colorcet.

I prefer the name bmy to avoid any ambiguity.

from colorcet.