On Jun 20, 7:48*pm, Bob Noel
wrote:
In article ,
*Le Chaud Lapin wrote:

And what analysis techniques would be applied to prove that the resulting
software intensive system is adequately safe?

The same techniques that employed, in general, by experts to test
software.

You don't have much training or experience with safety-critical software, right?

No.

I don't care how many "fastidious" people look at an architecture or the
as-built system, if they don't know what they are looking for and how to
find it, the odds of proving *anything useful are pretty small.

Well, assuming they are experts, each in their respective areas, they
would indeed know what to look for. *Also, peer-review (by other
experts) is a very good way to check structural integrity of software
(or any system).

The state-of-the-art for establishing/proving the safety of software-intensive
systems isn't particularly mature.

No argument here.

Every few years I meet someone who is doing research of proving
integrity of softare in general. In every case, the intellectual
effort involved in configuring the proving tool for the specific
application context is on par with the intellectual effort that would
have been employed to make the system correct in the first place.
When I point out this fact, the answer is generally, "Yes, we know,
but the idea is to eventually reach a point where the software can do
everything by itself."

Ahem.

-Le Chaud Lapin-