[SystemSafety] Collected stopgap measures

[Paul Sherwood]

> If, on a list devoted to safety critical systems, your best is to
> suggest that we need neither documented requirements nor documented
> designs, then you may have to acquiesce in having some people think
> that your best is a long way short of good enough.

I don't believe I suggested that these things are not required **for 
critical systems**.

I've said that lots of successful **software** has been created without, 
including some popular software which has been deployed in critical 
systems, which is clearly true.

Without system-level requirements and architecture I don't see how there 
could be a basis for making any claim of safety. It's once we get into 
the practice of software that your argument breaks down, imo.

Clearly some software continues to be constructed based on pre-existing 
requirements and architecture, but most is not, these days. From my 
perspective the risks associated with starting from scratch on large 
projects may outweigh the risks of using 'cots', even if the existing 
code came into existence without ticking the boxes that became enshrined 
in lore some decades ago.

>>> ... and heaven forfend that anyone make the contrapositive inference.
>> 
>> I'm not going to bother attempting to parse that.
>> 
>> 
> .... engineers and language again ... and if you think this is
> trolling, let me point out that, just as there is a distinction
> between irony and sarcasm, so also is there a distinction between
> trolling and sublimated ennui.
> 
> I reserve the right to express irritation at being confronted by
> patent incompetence on a list like this.

And I reserve the right to be wrong, whereas you appear to believe that 
you are correct always.