Sunday, May 2, 2010

Floating-point math

Most programmers don't get it right. They forget that floating-point numbers aren't precise, and that different numbers are irrational in base 2 than in base 10. End result: math errors and weird results.

So, on the web: a basic guide to floating point math.

And for the true math geeks: David Goldberg's paper on floating-point math, and the original ACM journal article if you have an ACM account with library access.

No comments: