[SystemSafety] Bossavit's Leprechauns book

[Les Chambers]

PBL
Please explain. Is Bossavit denigrating these concepts by calling them memes, myths or legends? Is he saying that because we can’t reliably quantify notions such as “The rising cost of discovering defects later in the “Waterfall" chain.” we should not be informed by them when making decisions?
If so he is wrong.
Story theory intrudes.
Take Adam and Eve. Was there ever a garden of Eden? Did the tree of knowledge of good and evil exist? How about the snake that told Eve how yummy the fruit was? Do snakes talk? And most important of all was there a God to tell them not to partake of forbidden fruit. Is there a god?
We could knock ourselves out arguing the credibility of detail of but we’d be wasting our time. The value of this story lies not in it’s verifiable/quantifiable details but in the overall truth it conveys.
Example:
Faced with the decision to obey authority or seek knowledge at any cost A&E chose the latter. Good on em I say.

My point is that our culture is founded on myths and legends. They inform our actions from minute to minute. So rejecting the ‘rising cost’ meme because it cannot be statistically proven to 6 Sigma over all projects for all time is counterproductive .

Do you really want to be debugging your Mars lander a range of several million miles and a loss of billions if it crashes into the surface? No? Why don’t you spend a few million dollars on independent V&V then.

Granted some stories convey falsehoods. But there are a suite of legends that we should preserve and cherish. Bossavit listed several of them.

Some of these legends need to feed engineering fervour to the strength of religious belief. Spawning acts of faith if you like. Mature engineers often make decisions based on gut feel unsupported by numbers. This is where the feeling in the gut originates.

These notions should resonate with engineers building safety critical systems because we often have to make the argument, ‘spend these millions and your return will be : nothing bad will befall you.’ Holding your ground with this shakey argument in the face of the unwashed and powerful requires deeply held beliefs. You don’t get them from statistical analysis.
Bossavit sounds immature. I don’t think I’ll be reading his book.

Les

PS.
RE: quantifying the ‘rising cost’ meme.
It can be quantified for classes of system developed in specific environments. Same company, same app, same people, same processes. This is the basis of every book ever written on software metrics. I believe this to be true because I was there, working for several years in fixed-priced software development. You cannot survive in that business without collecting metrics and augmenting the legends with your local numbers.

Managing Director
Chambers & Assoc. Pty Ltd.    
 www.chambers.com.au
www.systemsengineeringblog.com