[SystemSafety] Statistical Assessment of SW ......

[David MENTRÉ]

Hello Mr. Ladkin,

Le 2015-01-23 07:42, Peter Bernard Ladkin a écrit :
> Suppose you have a piece of SW S which is deterministic.

Therefore you assume your software is free of uninitialized variables 
and similar internal sources of non-determinism. This can be proven but 
is not regularly applied as far as I know.

> And S reverts to an initial state with no
> memory of its previous behavior each time it produces its output.

Is it a valid assumption, except for the most basic software (e.g. 
"emergency button handler")? Most safety critical software I know of 
have modes, memory of current mode and so on.

> Suppose the distribution of inputs to S has a stochastic character.

As others have pointed out, I'm skeptical about that point. Besides 
hardware failure and other unexpected inputs to the software, you need 
also to consider an attacker that would systematically use the weakest 
point of your software (e.g. the special input sequence in your example).

Sincerely yours,
D. Mentré