17 Dec
2003
17 Dec
'03
11:17 p.m.
I would look at monitoring some large scale system that is not easily modelled so it can't be predicted faster than it actualy occurs. This way when the appropriate condition occurs its digitized state could be used as a key.