But math "says so" because it's a consequence of the very few axioms from which all of mathematics is deduced. What's your point?

To put a CS perspective on this: in floating point arithmetic, two numbers x and y are considered equal if x - y < z, where z is the precision defined by your architecture (1e-12ish?).

Was actually going to bring up something very similar in math later...

iff |x-y| < ε ∀ ε (where ε is defined as an arbitrarily small positive value), then x = y