[SystemSafety] IEC 61508 Committee Draft of Edition 3 published

[Derek M Jones]

Peter,

>> The view is not widely shared because it is not in
>> the economic interests of those involved to express
>> that they are doing involves an element of smoke and
>> mirrors
...
> In 1981 a very large and involved book was published by Barry Boehm, SW head of TRW, about TRW's experience with costs 

The title was "Software Engineering Economics".  It's a good book, but
I would recommend skipping the 100 pages of over fitting 63 rows of data
to the COCOMO model. Details here
https://shape-of-code.com/2016/05/19/cocomo-how-not-to-fit-a-model-to-data/

> of SW, including plenty of statistical evaluation, because that is how you derive costs in such a company. And how TRW 
> had been evaluating their costs for the previous quarter century of bidding on large SW contracts, often for USG.

The 1978 book “Software Reliability” by Thayer, Lipow and Nelson,
using TRW data and doing a much better job than Boehm, is will worth
a read.  It's available online, with photos of poorly scanned
pages here
http://shape-of-code.com/2016/05/20/software-reliability-lots-of-detailed-data-and-thoughtful-analysis/

A book from 1978 still a top read over 40 years later.
It shows that software reliability has not moved forward
by much.

> Edward Adams published an evaluation of IBM's most-used SW program (a facility configuration program for clients) and 
> its reliability, and the phenomenology of its failures, suitably redacted for non-IBMers, in 1984.

Adams' 1984 paper remains one of the few public software reliability
datasets.  The fact that such a tiny dataset is still useful 40 years
later shows how little progress there has been.

> Watts Humphrey published a lot of studies for SEI in the 1980's through the 2000's, collected together by SEI in 2009.

This is a great dataset.  You can find a detailed analysis here
https://arxiv.org/abs/2106.03679
It is criminally underused.

> Saying everybody who evaluates SW statistically is an economically-motivated smoke-and-mirrors faker is not only daft. 

How would you mark a student whose answer completely
misrepresents what the question under discussion?

This book has several 5-star reviews on Amazon (but one
is somewhat biased)
http://knosof.co.uk/ESEUR/

-- 
Derek M. Jones           Evidence-based software engineering
blog:https://shape-of-code.com