Suppose I wanted to check if a given floating point value is within a certain range. What would be Your recommended approach when considering simplicity, speed, portability, etc.?
I thought floating point comparisons was one of those things which should not be done due to the inexactness of representing certain numbers, such as 1/3.
That's only if you test for equality, i.e. == and !=. That is why checking a range (of acceptable error) is recommended over checking equality (Look at firedraco's suggestion and Stewbond's first example to see how to check a range).