May 27, 2009 17:48
In my compilers class, we are discussing optimization. Floating point numbers don't behave like other normal numbers. Our example of how you can't necessarily optimize floating point operations was to add a big number to a small number, subtract the big number, and compare what you expect to what actually occurs.
"So we add the very small number to a big number, say the national debt."
To which a student replies, "the national debt isn't that big."
And I had to post this because saying $13 trillion isn't a big number is quite possibly the most retarded thing I've ever heard.