Re: 0.3*3 == 0.9?

[Jordi Gutiérrez Hermoso]

2010/8/25 Michael Creel <address@hidden>:
> I thought that comparison of floats returned true if the difference was less
> than machine precision.

Machine epsilon isn't always the best comparison size, not even with
relative error. Hardware floats don't carry information around them
saying, "this many of my digits are correct".

I do recommend you read the long-winded article I linked to earlier today.