🚀 go-pugleaf

In article <1991Jul24.211356.29432@news.nd.edu> wessels@thor.cc.nd.edu (Andy Wessels) writes:
>...  Third, by placing the blame on
>the hardware, you are admitting that there is a flaw in the software that 
>should have been corrected. 

>The common example of life critical software has been medical equipment.  There
>are other areas that are currently  being researched...

>Writing software is a science.
	So you believe it is feasible to generate *proofs* of program
correctness?  If you aren't arguing this, then your criticism is banal;
one of the best protections available is multiple mechanisms (e.g.
hardware which acts as a watchdog on inappropriate actions, multiple
checks in software (e.g.
ApplyXrays(act) {
	if (IsUnsafe(act)) abort("Danger!!!");
	DoXrays(act); 
}), etc.)  This is how NASA deals with the issue in space flight, and it
makes damn good sense.  Until you can *prove* a dangerous condition
can't arise, you should have multiple checks that it doesn't occur.
As I understand it, we can't really do proofs of this kind for large
systems (and to be safe, one would also need to prove things about
the operating system and even the hardware.)
		Ron