- Don't Test for Equality; instead, do something like (fabs(x - y) < tolerance)
- Worry about Addition and Subtraction more than Multiplication and Division; "the relative errors in multiplication and division are always small"
- Floating Point Numbers have Finite Ranges; eg "float f = 16777216; f+1 == f"
- Use Logarithms to Avoid Overflow and Underflow
- Numeric operations can return INF and NAN -- http://www.codeproject.com/Articles/29637/Five-Tips-for-Floating-Point-Programming