Originally: gpuweb/WSL#351

Decimal notation for floating point numbers is fine for common cases, but confusingly lossy. E.g. 1.5 vs. 1.3
Sometimes you really want to control exactly which floating point number is used, e.g. for mathematical constants like approximations to Pi, or for testing things like NaN-handling.

C99 hex float syntax is a worked-out absolutely unambiguous syntax for getting exactly the floating point numbers that you want. WGSL should support this additional syntax.