[css-color] Merge Lab() and LCH() in CIE()?

[Crissov]

With CIE lightness becoming a <percentage> per #4477, it becomes possible to unify Lab and LCH in a single functional notation. Iʼm not sure this is a useful thing to do, but perhaps others do.

CIELAB a and b might not be percentages in disguise, but CIELUV u and v apparently are.

cie() = cie( 
 /* Lab */    <number>{3} 
 /* Lab */ | [<percentage> && <number>{2}] 
 /* Luv */ |  <percentage>{3} 
 /* Luv */ | [<number> && <percentage>{2}] 
 /* LCH */ | [<number>{2} && <angle>] 
 /* LCH */ | [<percentage> && <number> && <angle>] 
           [ / <alpha-value> ]? 
           ) 

By this reasoning, though, rgb() should consequently also support <angle> && <percentage>{2} for either HSL or HWB, or <angle> <percentage>{2} for one of them and <percentage>{2} <angle> for the other.

rgb() = rgb(
 /* RGB */    <number>{3} 
 /* RGB */ |  <percentage>{3}
 /* HSL */ | [<angle> <percentage>{2}] 
 /* HWB */ | [<percentage>{2} <angle>] 
           [ / <alpha-value> ]? 
           )

[tabatkins]

Ooh, I really don't think this is a good idea. The fact that the grammars don't technically clash doesn't mean merging them makes sense.

We're not trying to conserve function names here; that namespace is wide open. Having the color functions all be separate functions that name their arguments in the same fashion seems far friendlier to authors than asking them to memorize a number of argument-order patterns and divine which color space is being used.

[Crissov]

Iʼm absolutely fine with rejecting this idea, although Luv support might still be worthwhile. It originally was just a point in favor of enforcing explicit units wherever they apply.