[SystemSafety] Difference between software reliability and astrology

[Prof. Dr. Peter Bernard Ladkin]

On 2024-08-13 14:57 , Prof. Dr. Peter Bernard Ladkin wrote:
> On 2024-08-13 13:17 , Derek M Jones wrote:
>>
>> It does not work this was for software reliability models because
>> of the lack of evidence of what works or does not work.

While I am at it, I might as well note your misuse of the term "evidence".

You write as though something is "evidence" or it is not, as if "evidence were a property. It's not. 
It is (at least) a binary relation: something E is evidence for assertion A.

So when you write that there is a "lack of evidence", I (and everyone else who understands this) 
will be wondering: lack of evidence *for what*? But you don't ever say.

When Uber ran over Elaine Herzberg in Tempe, after (if I remember rightly) half a million miles of 
accident-free autonomous driving, it changed the estimates of reliability and confidence levels in 
specific ways detailed here by Peter Bishop. One event, which was evidence for/against the 
accident-freeness of Uber's autonomous-driving system. It wasn't evidence for Waymo's equivalent 
hopes/claims. Nor was it evidence for the reliability of Airbus FBW systems.

So I think it would help comprehension, although at the same time rendering your favorite tropes 
meaningless, if, when you say "lack of evidence", you would also specific "for/against what assertion".

PBL

Prof. Dr. Peter Bernard Ladkin
Causalis Limited/Causalis IngenieurGmbH, Bielefeld, Germany
Tel: +49 (0)521 3 29 31 00