James Giles wrote:

> steve wrote in message <3498B5A6.C404C703@seasoned-software.com>...
> ...
> >ACTUALLY, Linux itself, and many programs, are SYSTEMS code!  Some
> >features can't or shouldn't be used, and it is ALWAYS more prone to
> >errors.  It surprises me how well Linux generally runs, and rarely
> >crashes!
>
> An interesting attitude.  The system is the one component of
> the programming environment that everyone *must* use every
> time they use the computer.  It *should* be the most robust,
> efiicient, and secure piece of code on the machine.

You're right(and I said always more PRONE), but there always seems to be
problems somewhere.  Especially if you are constantly upgrading it like
Linux.  Whenever *I* see an error, it almost always gets fixed in the next
release.

> The attitude
> that it needn't be seems to be a UNIXism.

Most of what people consider UNIX is in peripheral programs.  It isn't a
part of the kernel itself.One error in Linux that I have seen that I never
saw fixed(although others didn't seem to have the problem), is with
support of >2GB drives.  THAT is one place where tricks have to be
pulled.  I noticed that my disk repair utilities(though the latest from
the author) aren't doing such things, though Linux IS.

> To be sure, pre-UNIX
> systems had security an reliability problems (they were trying to
> make general purpose use of hardware we would now consider
> inadequate to run a toaster).

Well, TODAY all you have to do is diasable the fans, and VOILA!  TOAST!
8-)

> There's no excuse for this attitude
> today - except that the vendors of systems have noticed that
> UNIX never failed in popularity because of these weaknesses.
> Bill Gates owes a lot to UNIX, if nothing else: for the pervasiveness
> of the attitude that the system *needn't* be secure, robust, and
> efficient.

And windows with its DLLS is?  What of the security problems in IE4?  ...

> >How many C programmers learned ONLY C?  I doubt that it is very many.
>
> Well, you're welcome to your perceptions.  I'd guess that
> only about 1 in 10 C (C++, maybe Java, etc...) programmers
> I've ever met had any real substantial experience with any
> language outside that group.

OK, I have different experiences with C programmers.

> Unless you count AWK or SED
> as real languages!  Yet those with the least knowlwdge of other
> languages are always the quickest to dismiss languages
> with pat excuses ("Ada is too big", well it's smaller than C++
> and more cleanly designed - "Fortran makes you use GOTOs
> all the time", well it has all the "structured" flow control of any
> other language and actually restricts the use of GOTO more
> than C does - etc...).
>
> Now, among other things, I have in the past provided on-site
> maintenence for C compilers at a national laboratory.  I dealt
> with C programmers from around the world and at a time when
> they could *not* claim that the language never caused them any
> problems: when they had a bug they couldn't identify or fix.  So,
> while your perception may be different than mine, I'll still claim
> that mine is at least as valid.  I'll say one thing, C programmers
> are extremely loyal: they always accept the blame for their bugs.
> Even when it's a kind of bug I see often in C programs and almost
> never in any other language, they don't blame the language.

Well, most things common to C that you would call bugs ARE generally
called features.  It IS easier to make a mistake in C.  I never heard
anyone claim otherwise!  Although I HAVE seen COBOL people blame the
compiler for mistakes(and it often WAS responsible), THEY are certainly a
lowal group too!

> --
> J. Giles
> Ricercar Software