Excuse me, but I would not call the CACM article "excellent". The authors were simply trying to push for adoption of their particular favorite as a "standard". There are considerably better random number generators available on BRL UNIX systems if you know where to look.