My current favourite is this one.  It's new, but it's a good question with some great answers that teach you a bit about how processors, or rather programmers, handle operations that don't fit inside a registers width.

> [How can a computer calculate exponential maths without overflow error?](http://superuser.com/questions/452498/how-can-a-computer-calculate-exponential-maths-without-overflow-error#452640)

I find the "fit inside a register" to be an important part because many people assume that the processor cannot deal with a number that is wider than its register, but there are ways around that kind of limitation that humans use every day with their own brains forgetting that computers can use the same tricks.

Forgot to mention, I'd be interested to write about this myself but itay take me a while due to other commitments.