Here's an excellent article on floating point numbers, how they're stored and what kind of errors they can cause. http://www.regdeveloper.co.uk/2006/0...approximation/ .

Cheers,
cgkanchi