All,

Michael Feldman <[log in to unmask]> writes:

> I don't think I'm all that cynical; I just haven't seen any evidence
> of public outrage about computer stuff, except a desire to punish
> obvious crime (like kids who write viruses, etc.). I think it's
> as much a problem of ignorance as of apathy.

If anything, it is a matter of low expectations. People are by now so
conditioned to having computers (well, software) fail on them that the
universal remedy (just reboot the d*** thing) has become deeply
engrained. I'm not sure who said that, but paraphrasing it: ``lowering
expectations of software reliability was perhaps the biggest crime
Microsoft committed''.

Most people -- including computer science graduates -- have now grown up
and were educated in an environment where highly complex and flaky systems
are the norm.

That theme was also brought up elsewhere in this thread: asserting that
yes, you can design and build systems that perform their mission with a
high level of accuracy and assurance is increasingly met with incredulity,
even within the profession. And in many ways that attitude is justified
given that the building blocks people are given to work with are
ridiculously flawed and overly complex.

> Well, I don't think it's realistic to expect ACM, a huge (~100,000-member)
> and diverse organization, to be fanning the flames of anything. Such
> an organization has to serve too many constituencies to "take sides".

Not all hope is lost. The ACM position statement on the DMCA and its clear
(if not literal) impetus of having research on security and reliability
suppressed was something I would not have hoped for.

But unless the theme is picked up by policy makers (and, I'm afraid,
horribly mangled in the process) that's not going to go very far.

So far the obscurantists and apologists for bad software have won the day.

> > I'll also have to admit that, if anything, "we" (ACM) presently are doing
> > just the opposite - aiding and abetting irresponsible practitioners.

Certification always has and always will carry the danger of being outdated
quickly and becoming a `cargo cult' process (again, too lazy to look up the
reference).

Unless there exists some mechanism to force the creators and consumers of
software to adhere to certain standards, that might just accelerate the
move of development to software sweatshops.

> Some people on the IEEE-CS side of things are pressing strongly for
> individual licensure of software engineers. After years of following
> the debate over this, I still wonder whether it would make much
> difference.

Getting students (and practicioners) to acquire skills and appreciation for
the tools available is a (weak) precondition for the development of high
assurance systems. I'm not sure how efficient I am in proselytizing, but
that does not stop me from trying to get people to raise expectations and
developers to use proper tools and techniques, including Ada and formal
methods for specification and validation. That probably puts me on Team
Jeremiah as well as on Team Ada, but so what.

--
        later,
        Stephen

Fraunhofer-IGD                 | mailto:
Stephen Wolthusen              | [log in to unmask]
Fraunhoferstr. 5               | [log in to unmask]
64283 Darmstadt                | [log in to unmask]
GERMANY                        |
                               |
Tel +49 (0) 6151 155 539       | Fax: +49 (0) 6151 155 499