[SystemSafety] C for OSs

[Peter Bernard Ladkin]

On 2019-10-14 10:36 , Grazebrook, Alvery AN wrote:
> .... The agile manifesto has its place. 

Well, yes, in the following sense. Your description describes the advantages of incremental
development and advancemanship, two key processes described by Barry Boehm in SEE in 1981, which
came to be called the Spiral Model.

The original advantage of the "Agile" approach was the four-eyes approach to coding. The idea was
that there were fewer chances to make mistakes (this was also known from the High-Reliability
Organisations work of the sociologist of technology Todd LaPorte and colleagues). That is how
Alexander Reinefeld and I wrote our interval-constraint-solving code in 1991-3, which produced the
practical algorithms which lasted for a decade and a half (at least), but we didn't know it had a
name. It worked very well for us.

Then management theorists turned it into a whole-organisation body sock. Programmers who weren't
born when it started apparently didn't realise most of it was lifted from existing techniques such
as Spiral (most programmers do not seem to be familiar with any history of software engineering. It
is astonishing how many supposed experts in dependable computing aren't familiar with Ken Thompson's
perfect Trojan, for example).

The problem with both it and Spiral is that it violates the condition that allowed Waterfall to be
so helpful. Waterfall is a way of organising the documentation rationally, as pointed out by Parnas,
and indeed Royce in the original paper. What you need to substantiate a critical system is a
documentation stack that shows (a) the derivation of functional requirements, and (b) the system
development by formal refinement from functional requirements. That doesn't arise from anywhere in
the "agile" process descriptions. It has to be engineered from scratch. And if you are going to have
to engineer that from scratch in any event, you might as well use it to help guide your forward
development in the first place. Which is what happens in the better SW shops, as far as I can tell.

PBL

Prof. Peter Bernard Ladkin, Bielefeld, Germany
MoreInCommon
Je suis Charlie
Tel+msg +49 (0)521 880 7319  www.rvs-bi.de





-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://lists.techfak.uni-bielefeld.de/mailman/private/systemsafety/attachments/20191014/e3f8dc8b/attachment.sig>