You're viewing a single thread.

LoL I just tried 10/0 and got "Not a number".

AFAIK that should give you +infinity, not NaN

Almost. 1/x approaches infinity from the positive direction, but it approaches negative infinity from the negative direction. Since they approach different values, you can't even say the limit of 1/x is infinity. It's just undefined.

it is possible to rigorously say that 1/0 = ∞. this is commonly occurs in complex analysis when you look at things as being defined on the Riemann sphere instead of the complex plane. thinking of things as taking place on a sphere also helps to avoid the "positive"/"negative" problem: as |x| shrinks, 1 / |x| increases, so you eventually reach the top of the sphere, which is the point at infinity.

https://en.wikipedia.org/wiki/Division_by_zero#Floating-point_arithmetic

In IEEE arithmetic, division of 0/0 or ∞/∞ results in NaN, but otherwise division always produces a well-defined result.

**Dividing any non-zero number by positive zero (+0) results in an infinity of the same sign as the dividend.**Dividing any non-zero number by negative zero (−0) results in an infinity of the opposite sign as the dividend. This definition preserves the sign of the result in case of arithmetic underflow.10/0 ≠ lim x->0+ 10/x

Or in other words, the thing you keep quoting does not apply in this case. Any number divided by zero is undefined, not positive infinity (or negative infinity for that matter).

It's undefined in math, but not floating point arithmetic

It's an error, since no amounts of zeros, even infinite, would make it equal 10.

https://en.wikipedia.org/wiki/Division_by_zero#Floating-point_arithmetic

In IEEE arithmetic, division of 0/0 or ∞/∞ results in NaN, but otherwise division always produces a well-defined result.

**Dividing any non-zero number by positive zero (+0) results in an infinity of the same sign as the dividend.**Dividing any non-zero number by negative zero (−0) results in an infinity of the opposite sign as the dividend. This definition preserves the sign of the result in case of arithmetic underflow.