Re: No call for Ada (was Re: Announcing new scripting/prototyping language)

["Alexander E. Kopilovich" <aek@VB1162.spb.edu>]

Dmitry A. Kazakov wrote:

> >> Requirements and terms of
> >>liability are pretty easy to set and check.
> > 
> > Are you a lawyer? I think that a person who is not a lawyer should not
> > make statements of this kind anyway.
>
> I was not talking about technical details of a particular license agreement.
> I was talking about a common, long established practice of protecting
> customers from frauds.

The problem is that in software it often isn't easy to differentiate a fraud
or liable negligence from a reasonable risk, to which the user was knowingly
agreed. Almost all computer professionals will agree that given the current
state-of-art, there is a risk-cost tradeoff for common software products. And
with the risk of eventual software error being pushed too down the cost will
skyrocket. So users have a good reason to agree with some risks - because they
need (or want) the functionality, provided by the product, but only for a
certain price, and not higher. This is a user's choice - either to take a
proposed risk, or not to buy the product. The problem is that with huge
diversity of both software products and user's background, attitudes and
education, it is impossible to represent clearly that trade-off as it is set
in a particular product. And even to represent it vaguely, but more or less
enough for anticipated courts, the professional lawyers are needed.

> So, are you a lawyer to claim that this practice is
> inapplicaple to software?

Well, I'm not a lawyer, although once I attended law school for 2 months -
just for getting some understanding of basic principles of law and of logic
of lawyers.

> > By the way, I remind you that Robert Dewar (who, although not being a
> > lawyer himself, nevertheless have some real experience with these matters
> > regarding software) warned many times (both here and in GCC mailing list)
> > against making law-related statements about software without consulting a
> > professional lawyer.
>
> These discussions concerned concrete details of various GPLs.

Yes (as far as I remember), but the warnings often were general - for
law-related statements about software without consulting a professional lawyer.

> I didn't asked about purpose of a text processor. I did about the software
> quality.

Being accesible and usable for large number of diverse people is one of the
most essensial qualities for some kinds of software.

> >> > when you appeal to the government for such purposes, you just try to
> >> > exploit another market - a narrow one, for privileged parties only.
> >>
> >> and more competent ones.
>
> ...
>
> Should I quote Churchill?

If you know a piece of Churchill's speech or writing containing his opinion
about the government's competence in technological issues then please, quote
it.

> >> Again, there are issues essential to the existence of our civilization.
> > 
> > You are so focusing on our civilization... don't you think that
> > civilization is quite a complex thing, and it is not easy to determine
> > what is essential for it? But at least you don't refer to the God's will,
> > and this is good.
> > 
> >> Software development becomes one of them. Ada is just an indicator of how
> >> things are going on.
> > 
> > Don't you see that these two your sentences contradict each other?
> > Ada essentially defies the general notion of "software", Ada expects you
> > to analyze each your problem deeply, with all its specifics, not relying
> > upon some general "software" fashions and solutions. It does not mean that
> > you should forget all your previous experience and it doesn not preclude
> > any general computer science - you can use all that if it helps you, but
> > you can't rely upon that, you can't get excuses from that.
>
> Where is any contradiction?

You said that software development - general process for general notion/issue
- becomes critically important. Then you said that the things are going very
badly in this respect. Then you hinted that the negligence to Ada is a symptom
as the latter. But Ada essentially defies general notion of "software", so you
logically should perceive the negative attitude towards Ada as a good symptom,
that is, a symptom of widened understanding of importance of general notion
of "software".

> > Well, maybe you don't know, but in dark years of the Cold War, a
> > substantial part of the hope was that there are enough errors in missile
> > systems (that is, missiles themselves, navigation etc.) on both sides; so
> > that if they will be really launched then most of them either will not
> > takeoff or will not explode or will explode somewhere in ocean. And some
> > of us also hoped that the goverments are also somehow aware of that
> > possibility of those massive errors, and neither side can be sure whether
> > it has much less errors or much more errors than the opponent.
>
> Pretty silly, because it is actually no matter where it would explode.

If it would explode at all, and explode by nuclear way (that is, warhead) and
not by chemical way (that is, missile's body - engine, fuel tanks).

> > Having an atomic tonnage of America or Russia,

Crucially important was the bootstrap behaviour. The hope was that the
bootstrap will not succeed, but will abort - on both sides.

> >> >> As an example, I know a
> >> >> car vendor, which for years is unable to find a bug in its *one*
> >> >> engine controller, running *one* task, causing sporadic stop.
> >> > 
> >> > Poor vendor, he can't reach me -:)  Did he try to create a simulator?
> >>
> >> Of what? Motor simulator? Roller dynamometer? Climatic chamber?
> >> Autopilot?
> > 
> > You said - "a bug in its engine controller", so I mean a simulator of that
> > engine controller.
>
> Controller is just a processor, usually there is no need to simulate it.

I see (with surprise) that an explanation is needed: it is input signal
stream(s) that should be simulated. Whether the hardware involved should be
simulated also, or the actual hardware may be used instead - this depends on
particular circumstances and equipment.

> Anyway the cost of various testing, simulating etc, hardware/software is
> measured in tens of millions of dollars. Their maintenance..., who knows.

A particular simulator of/for a particular controller for investigating a
particular bug - tens of millions of dollars? Perhaps you mean some ideal
error-preventing system, which will cost more than than accumulated cost of
all errors, which it prevents over current error-prevention/debugging
technique,

> So a manager might think that this is more than enough to find a bug.

What can he wish else, if a bug is already present and perceived? 

> But it is quite strange to hear similar statements from a software man.

I don't know about software men, I never meet such people, at least in person...
or didn't recognize them as such.

> Debugging is the least productive way for preventing software faults in
> embedded/real-time world. In many cases it does not work at all.

But what to do when a bug was not prevented, and shows itself in released
product? You are proposing to drop away this product and rebuild it using
some (very costly) ideal error-free methodology, from scratch?



Alexander Kopilovich                      aek@vib.usr.pu.ru
Saint-Petersburg
Russia