[css-color-hdr] Why do the new non-rgb/xyz colorspaces use the color() function?

[weinig]

In CSS Color HDR, the new colorspaces all use the color() function notation, even though Jzazbz, JzCzHz and ICtCp are not RGB or XYZ colorspaces.

Way back when, I filed #5887, arguing we should only use color() for additional colorspaces, but at the time it was rejected.

If these are being added to the color() function, does it stand to reason that lab()/lch()/oklab()/oklch() should also be added as well?

[svgeesus]

Fair point, and there is still time to change this if that is what we want.

[weinig]

@svgeesus, I can't tell from your response what you are suggesting is still possible to change.

Are you saying we should give Jzazbz, JzCzHz and ICtCp their own functions, jzazbz(...), jzczhz(...), and ictcp(...)?

Or, are you saying we should add color(oklab ...), color(oklch ...), color(lab ...) and color(lch ...)?

(My personal preference would be the latter, deprecating the existing forms, and also adding color(hsl ...) and color(hwb ...) to the party).

[svgeesus]

Sorry.

Are you saying we should give Jzazbz, JzCzHz and ICtCp their own functions, jzazbz(...), jzczhz(...), and ictcp(...)?

Yes. These are not implemented (well, color.js parses them and emits them, but no-one depends on that). Ant it makes CSS Color HDR align more with CSS Color 4 in terms of color functions.

Or, are you saying we should add color(oklab ...), color(oklch ...), color(lab ...) and color(lch ...)?

(My personal preference would be the latter, deprecating the existing forms, and also adding color(hsl ...) and color(hwb ...) to the party).

That seems like a lot of testing work for little actual benefit (especially if we then deprecate the other forms).

[Crissov]

If you are in fact going to add separate functions for J_za_zb_z, J_zC_zH_z and IC_TC_P, then please leave out the indices! jab(), jch() and icc() (or perhaps itp()) would work just fine.

In another world, we might have had unified lab() and lch() functions, e.g like lab(… in CIE), lab(… in OK) and lab(… in HDR).

[svgeesus]

If you are in fact going to add separate functions for J_za_zb_z, J_zC_zH_z and IC_TC_P, then please leave out the indices! jab(), jch() and icc() (or perhaps itp()) would work just fine.

Good point, J_za_zb_z in particular is super hard to say in conversation. And we left out the * in CIE L* a* b*.

On the other hand, there is a separate colorspace called ITP which might be confusing, and icc might be confused with ICC color profiles.

But on balance yes, simple and readable color function names.

[svgeesus]

CIECAM02, CIECAM02-UCS and CAM16-UCS also have Jab so unpronounceable Jazbee names it is, I think.

[svgeesus]

Another argument in favor of separating out the non-RGB spaces from color() is RCS. It would be odd to write

color(from var(--base) jzczhz calc(r/2) g calc(b+180)) to halve the lightness and get the complementary hue. These functions should have their own component keywords.

jzczhz(from var(--base) calc(j/2) c calc(h+180))

[weinig]

We do already have different constants when using the XYZ colorspaces in an RCS color() function.

[svgeesus]

@weinig so the best option seems to be

give Jzazbz, JzCzHz and ICtCp their own functions, jzazbz(...), jzczhz(...), and ictcp(...)?

Yes. These are not implemented (well, color.js parses them and emits them, but no-one depends on that). Ant it makes CSS Color HDR align more with CSS Color 4 in terms of color functions.

[svgeesus]

PROPOSED: In CSS Color HDR, move Jzazbz, JzCzhz and ICtCp from color() to their own functions, for consistency with CSS Color 4 Lab, LCh, Oklab and Oklch

[svgeesus]

@astearns suggest the CfC ends before the next HDR breakout in 2 weeks

[weinig]

Proposal seems good to me. +1.