Feature request: Bit manipulation

[Winterhuman]

A bit of a niche request (pun not intended), however, I think it'd be a useful feature for visualising how bit-based approaches differ to their "unoptimised" counterparts (e.g. The fast inverse square root algorithm).

There's a lot of further questions which greatly complicate this feature request:

1. How many bits are integers assumed to use?
   • Should there be a function for this, like bits(x, 4) for translating a given value of x into 4 bits, or a hardcoded value such as 32 bits? Or maybe a setting?
2. Are only integers allowed, with everything else being floored?
   • If not, what kind of floating point number scheme would be used? Similarly, should "longs" be considered?
3. Thinking back to the Quake optimisation linked above, should the 0x... syntax be included?

I doubt there's a reasonable way to cover everything mentioned here, since this issue is essentially asking for an entirely new syntax system alongside the existing one, but let me know what you think.