I think every person (me included) who knows shader languages (at least a bit) get this idea at some point: “I want swizzling in my C++ math library!”. At the end of the day tho, I don’t believe it’s that useful.
Swizzling exists in HLSL/GLSL out of necessity. One has to write efficient and readable code with optimal memory layout. Swizzling helps there a lot, I won’t argue.
In C++ tho? When one have good optimizer and much less restrictions (e.g. functional stack instead of registers)… I don’t believe it’s needed. Maybe it’s just my biased experience, but I know only one place that will benefit from it – converting (x, y, z) to (x, z). Does this syntax simplification really worth having a huge chunk of template magic in math code? Don’t think so. I’m asking myself: “how will it help a person to make a game”, and I fail to find an answer.