I can't count the amount of times I see this text being posted. Yet it takes something like a degree in math/CS/... to read and understand it completely; it is really not suitable for everyone.

As such I bookmarked Essential facts about floating point calculations so next time a psychologist or biologists comes asking me if I already knew matlab has a bug when comparing 0 to 0, I can point him/her to this document. And some proper code to compare floating point numbers using a relative error or so.

For anyone noticing the “Oracle” in the domain name: The same document used to be at http://docs.sun.com/source/806-3568/ncg_goldberg.html