[SystemSafety] What do we know about software reliability?

[Brendan Mahony]

> On 16 Sep 2020, at 12:46 am, Peter Bernard Ladkin <ladkin at causalis.com <mailto:ladkin at causalis.com>> wrote:
> 
> I have no idea why anyone might complain generally about such definitions. They are ubiquitous in
> electrical engineering and make plain sense (although one can quibble about the precise phrasing -
> and does!).

Prediction is very difficult, especially if it's about the future.

Most of the activities associated with software reliability estimation would seem to be worthy, but do they really result in any ability to predict future behaviour (in the absence of mathematical analysis of algorithmic features)?

For example, is Bayesian prior style reasoning applicable to (very complex) software controlled systems?

Does the Central Limit Theorem even apply to a sequence of test events involving a software controlled design?

Even if applicable, will not the process of averaging hide potentially crucial features of system behaviour?

The existence of persistent memory, Y2K, memory overflow, quick release of security patches, etc would tend to cast considerable doubt on this.

Hardware reliability processes leverage off centuries of experience confirming the normal distribution of manufacturing errors, wear and tear etc on very simple systems (components).

Even so, the advent of pressurised jet aircraft led to some tragically inaccurate reliability predictions.

What do we have to counter the claim that ever single software controlled design compromises a completely new class of system for the purposes of reliability assessment?

Indeed why is this not true of any sufficiently complex system design, such as a bridge or skyscraper?

Why is it not the case that reliability estimates should be synthesised from the component reliabilities through mathematical analysis of the proposed design?

You don’t predict the reliability of a bridge design by building thousands (billions?) of copies and testing them!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.techfak.uni-bielefeld.de/pipermail/systemsafety/attachments/20200924/770b3286/attachment.html>