[SystemSafety] RE : Qualifying SW as "proven in use"

[Peter Bishop]

I agree with John that caution is needed
While stochastic behaviour of the inputs means software failures are 
also stochastic, the big uncertainty is what happens in a new 
environment with a new stochastic behaviour.
Showing environment X is the "same" as environment Y is pretty tricky
and it is not clear if anything can be claimed if the environments are 
different.

SPRIGGS, John J wrote:
> When Peter's contribution arrived, I was drafting a reply to
> Matthew's posting that also justified the use of statistical
> techniques where failures are systematic.  I will not bother with
> that justification now, but I will add a bit of advice for Bertrand's
> postulated safety-engineering rôle-holder.
> 
> Once you have established a quantitative model, it is easy to
> continue working in terms of the numbers and disregard the, perhaps
> tenuous, correspondence with the real world.  This is similar to the
> situation with qualitative models where one can be lulled into
> working in terms of a 'feeling', and not acknowledging real-world
> complexity.
> 
> Do not present analysis results with many 'significant figures' and
> do not use them as the sole focus of assurance arguments.  Rather,
> use predicted failure rates and associated analyses to support your
> assurance arguments, be they for reliability or for safety.
> 
> Statistical methods are used for quantifying equipment and system
> reliability through the analysis of failure data.  Due to the high
> levels of uncertainty involved, these analyses do not offer the level
> of precision that the engineering user may expect.  Treating these
> highly-uncertain numbers in the same way as precise measurements is
> unsound and is likely to lead to false conclusions.
> 
> 
> John
> 
> Usual disclaimer about my opinions not being those of my employers,
> clients, et alia -----Original Message----- From:
> systemsafety-bounces at techfak.uni-bielefeld.de
> [mailto:systemsafety-bounces at techfak.uni-bielefeld.de] On Behalf Of
> Peter Bernard Ladkin Sent: 27 June 2013 12:35 To: Matthew Squair Cc:
> systemsafety at techfak.uni-bielefeld.de Subject: Re: [SystemSafety] RE
> : Qualifying SW as "proven in use"
> 
> <snip>
> 

-- 

Peter Bishop
Chief Scientist
Adelard LLP
Exmouth House, 3-11 Pine Street, London,EC1R 0JH
http://www.adelard.com
Recep:  +44-(0)20-7832 5850
Direct: +44-(0)20-7832 5855