[SystemSafety] Bursting the anti formal methods bubble

[Derek M Jones]

Michael,

> The 
> condescending attitude of the speaker, his ignorance of the reality of 
> commercial software development and assurance practices, his arrogant 
> assertions of superior knowledge without any supporting evidence, and 
> his thinly veiled claims of ethical failures in non formalists changed 
> the attitudes of quite a few attendees. But not in the way he 

You're right, formal methods people are their own worst enemy.

I think it boils down to a belief that anything done using mathematics
is (magically) correct.

Next time you talk to one of these people, ask them how they know the
mathematics is correct.  They will probably look completely baffled.
I have been told more than once that mathematics is more than two
thousand years and and mathematicians know what they are doing (being
two thousand years old did not keep the sun+planets orbiting the
earth).

Switching to a relative improvement argument, rather than absolute
arguments based on 'proof', might enable some real progress to be made.

> (presumably) hoped. Whereas many came to the conference thinking formal 
> approaches may be able to help them, nearly all left the room after the 
> talk thoroughly repulsed by the very idea. Someone may argue that the 
> folks shouldn't have been so sensitive. That they should've been able to 
> get past the bombastic rhetoric to the truth of the ideas presented. 
> Perhaps. But arguing that way ignores human nature, and presupposes that 
> evidence was presented about the truth of the ideas.

-- 
Derek M. Jones           Software analysis
tel: +44 (0)1252 520667  blog:shape-of-code.coding-guidelines.com