>There is some hypothesis, offered by somebody I cannnot recall now, that says that no line of code can ever be proven to be bug free because it cannot be run in every conceivable scenario.

I recall being taught that the longer a program is in production the less likelihood of remaining bugs, so though you can't every say a system is totally bug free, when a system has run (and been debugged) for a long time you can get "close enough."