Re: Bug-free software (was: Re rant about viruses on VMS...)

[der Mouse <mouse () Rodents Montreal QC CA>]

["David Crocker" <[EMAIL PROTECTED]>, replying to me]
> > Any nontrivial piece of software has bugs.
> No, there are some nontrivial pieces of software in which bugs have
> never been found and there is no reason to believe any exist.

"Never been found" doesn't mean they don't exist.

And whether there is no reason to believe any exist is a matter of
opinion, not fact; I certainly have reason to believe at last one bug
exists, reason which you do not see as being such a reason.

> I'm thinking here of safety-critical software developed using formal
> specification, verified refinement and automatic code generation,

If the english for the program is a "formal specification", then that
is the appropriate measure of size and that is where I would expect
bugs to exist.

> "The 115 000 lines specification

Even if the translator that compiles this specification to machine code
is perfect - which itself is something I question - I do not believe
humans can generate over a hundred thousand lines of anything,
including formal specifications, without bugs.

You choose to believe otherwise.  For the sake of everyone whose safety
depends on the result, I hope you're right - but I can't make myself
believe it.

/~\ The ASCII                           der Mouse
\ / Ribbon Campaign
 X  Against HTML               [EMAIL PROTECTED]
/ \ Email!           7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B