News for hackers. Stuff that matters.
Re:Colorbox by Lyft Design by jacobolus on Friday May 03, 2019 @03:35AM (#19816316)
Cute, but people need to stop using HSL and similar color models for this kind of thing. It is horrifically bad for it, designed by software researchers in the 1970s who had very slow computers and as far as I can tell little background in color vision, color reproduction, visual art, or graphic design, and then passed down through the generations with very little thought along the way.

A human being who is not already an expert should essentially never be looking at numerical values of “R”, “G”, “B” or these trivially derived quantities of “H”, “S”, “L”, “V”, etc. and trying to assign meaning to those numbers. It just results in confusion.

A visual color picker interface (with no numbers involved) is pretty much the only justifiable context where HSL or HSV ever makes sense. And arguably not there either.

https://en.wikipedia.org/wiki/HSL_and_HSV#Disadvantages

Re:Colorbox by Lyft Design by moontear on Friday May 03, 2019 @03:56AM (#19816425)
people need to stop using HSL and similar color models for this kind of thing

What is your suggested alternative?

Re:Colorbox by Lyft Design by jacobolus on Friday May 03, 2019 @04:03AM (#19816448)
Munsell, NCS, OSA-UCS, CIELAB, CIECAM02, CAM16, ICtCp, Jzazbz, ....

Pretty much anything designed by color experts from the past 100 years would be better.

Here’s an implementation of CAM16 I was making today, https://observablehq.com/@jrus/cam16