The standard calls for two infinities (positive and negative) as well as two zeroes (also positive and negative), making it similar to a 'discretized' version of the extended real numbers The standard is structured in such a way that positive floating point numbers are easily compared (greater than, less than, equal) by interpretation as integers.
Several values can be represented in more than one way, including zero (+0 and –0) and the special value not a number.
- Jean-Michel Muller, Nicolas Brisebarre, et al, Handbook of Floating-Point Arithmetic. New York: Springer (2009)
- IEEE Xplore Digital Library, 754-2008 IEEE Standard for Floating-Point Arithmetic (requires individual or institutional subscription)
- IEEE 754-2008 — From Wikipedia, the free encyclopedia.