Or how calculus can give a feel of why approximation errors can be great or small with floats.

Some reasons why Python and C behave differently on this matter. It's a source of mistakes.

Probably the definitive resource on how floating-point arithmetic works.

Interesting tool. Hopefully will help us manipulate floating point expressions better.

Floating-point numbers are really a complicated species. This is an interesting deep dive in some of their representation.

Good explanations of why you might want to revert the Z axis on your projection matrices. Of course it boils down to how float precision works.

Now this is a really neat way to explain how floats work and how you loose precision. Definitely a good trick I should keep in mind when I have to talk about them, it's always been a chore to explain them.