[SystemSafety] Fault, Failure and Reliability Again (short)

[Martyn Thomas]

That's very helpful, Mike.

What evidence do they consider adequate to demonstrate 10^(-9) for
digital hardware?
And do they specify a confidence level for the probability?

Regards

Martyn

On 04/03/2015 15:58, Mike Ellims wrote:
> Some time back I had the opportunity to sit down with the heads of the FAA
> certification panel (we were doing something unusual) and we discussed this
> to a great extent. The outcome is that for the purposes of certification the
> failure rate of software is not considered as the FAA do not believe it
> possible to reliably calculate a failure rate for software.
>
> The figure of 10^(-9) applies ONLY to hardware; software is required to meet
> the process/quality requirements as given in DO-178 and verified means of
> audit (i.e. they come in and work you over ;-).