Re: Bye-bye Ada ?

["Warren W. Gay VE3WWG" <ve3wwg@cogeco.ca>]

Hyman Rosen wrote:
> Chad R. Meiners wrote:
> 
>> the tools were not appropriate for the job at hand
> 
> Fine, but the OP seemed to be saying that Perl, Fortran,
> and assembly are *never* appropriate. That's not correct
> even now, and is certainly anachronistic.

In fairness "never" is always a "strong" statement, and I
didn't really say that. ;-)

However, I'll confess that I don't believe that there are many
appropriate "production level" applications that should _today_
be written in Perl, FORTRAN or assembly language. Sure, device
drivers are still appropriate, and for small segments of an
operating system where squeezing the last byte out of the code
is important (like a boot sector program segment).

Perl is OK for a quick and dirty "hack something together"
to perform a one time job. Perl is absolutely the wrong
language to use if someone at a later point in time, and
especially a _different_ someone, is stuck with maintaining
it (it is IMHO, a write-once, "hope you don't have to read
it again" language).  I've also witnessed endless problems with
module version/compatibilities problems, on hosts where some
people felt Perl was appropriate for production.

FORTRAN IMO, is OK for legacy stuff that is already in that
language (tested and trusted), with the usual caveat that it
depends on the application (it _may_ never be good enough for
space shuttles, and other critical stuff).

This is a long winded way of simply saying, I don't believe
that there are many applications for those inferior tools
today. I also believe that better tools exist (Ada95), and
that too many people jump on the "efficiency bandwagon"
instead (C/C++). Efficiency becomes less important every
year that new CPUs advance in speed.

OTOH, _reliability_ becomes increasingly important for general
purpose computing as we try to build upon a foundation. Its
hard to build on a shakey one. ;-)

-- 
Warren W. Gay VE3WWG
http://home.cogeco.ca/~ve3wwg