On software quality and engineering

[Antti.Roppola at brs.gov.au]

Tom wrote:

> In my opinion, the problem is that we are dealing with a discipline,
> that has really had about 30 years of vigorous development in terms
> of us humans building up undertsanding of information and how to process
> it. Other forms of engineering have had many more years of this, I woudl
> dare say somewhere about 200 years or so, and in case of what we call
> Civil (as opposed to Military) or Structural engineering - even longer.

I was reading about this quite some time ago, and the observation that
we are still only one generation removed from Turing, and that many of
the pioneers are still very much alive. Can you imagine being an engineer
around the time of the industrial revolution? From wooden sailing ships
to iron steamships within a similarly short period of time. Because it
was all new, there were some amazing mistakes. But as you say, we figured
it out.

> If you look at history of engineering as a whole, you will see many a
> painful lessons there. Luckily as a body, engineersa were able to learn
> form the mistakes, and take advanatage of the scientific discoveries
> around them. Unfortunately for us, when society as a whole hears the
> term "engineering" then automatically, they come to expect the same
> standards of reliablility and rebostness as we see in most mechanical devices.

> Answers to the problem: I really do not have many apart from education
> and perhaps standards.

This is one of the *major* contributions of OS software. It'll be things
like Debian policy and the coding standards required for Ninja Hoards software
projects that will form the foundation of a mature science. Like how we now
take TCP/IP for granted.

Antti