[SystemSafety] Software Safety Requirements according to IEC 61508

[Peter Bernard Ladkin]

On 2016-05-22 02:59 , Matthew Squair wrote:
> I think that the 61508 statement regarding validation simply reflects a prevalent world view.

I agree with you there. But it was shown, twice, nearly a quarter century ago in the literature that
this world view is inappropriate for systems required to be ultrareliable. So (an interesting but
secondary question) why has it persisted?

> From a Bayesian perspective, for this behaviour to be rational we would need to start with a very
> high prior probability (belief) that our system under test delivers the safety function and
> associated properties. 

Yes.

> So yes validating through testing is rational, as long as we have that
> extreme level of confidence. 

No it's not.

If you start out with a strong enough prior, all the testing it is feasible to do cannot improve
that prior ("improve" here means increase the prior reliability estimate, that is, decrease the
estimated failure rate).

Why do something that demonstrably brings you nothing, but costs resources?

PBL

Prof. Peter Bernard Ladkin, Faculty of Technology, University of Bielefeld, 33594 Bielefeld, Germany
Je suis Charlie
Tel+msg +49 (0)521 880 7319  www.rvs.uni-bielefeld.de





-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 455 bytes
Desc: OpenPGP digital signature
URL: <https://lists.techfak.uni-bielefeld.de/mailman/private/systemsafety/attachments/20160522/39c5bffa/attachment.pgp>