2 ms·
It seems like you can only prove something in computer software up to the point of some random failure in hardware. If there's a chance that a cosmic ray will r
by baddox 15y ago
It seems like you can only prove something in computer software up to the point of some random failure in hardware. If there's a chance that a cosmic ray will randomly flip a bit in memory every quintillion clock cycles, then surely you can only prove that an algorithm is correct with the explicit concession that it might be wrong every decade or so.