Fast, Cheap Random Numbers

The keys needed to encrypt credit card transactions and other crucial information floating in cyberspace often rely on an infusion of random numbers. Generating true random numbers is actually harder than it seems since the generation process generally follows some deterministic algorithm, permitting the possible reappearance of unwanted predictability. James Gleeson, a physicist at Kent State University (330-672-9592, gleeson@physics.kent.edu) has come up with a cheap, fast solution. He shoots laser light into a sample of liquid crystals. But because the sample is subject to a turbulent flow, causing haphazard fluctuations in the orientation of the liquid crystals, the digitized transmitted light coming from the sample represents a stream of random numbers. Gleeson believes that because his device depends on standard liquid-crystal-display technology, his compact device can be used for many processes requiring random-number generation. (Gleeson, Applied Physics Letters, 9 September 2002.)