Re: What ada 83 compiler is *best*

[eachus@spectre.mitre.org (Robert I. Eachus)]

In article <to.reply-0812980710010001@129.197.97.40> to.reply@read.my.sig (Rick Thorne) writes:

  > Honestly, this is irrelevent.  Whether or not we as a handful of people in
  > the profession like Ada has little to do with the trends.

  Then let's look where the real trends are.  When someone asks "Is
XXX a dead language, they are not at all concerned with whether or not
there is a small cabal of programmers still using the language, they
are concerned with the answers to three questions:

  1) Are compilers available for popular and currently available hardware
and operating systems?

  2) Is the language itself maintained?

  3) Will compilers be available for future hardware and software
releases?

  Let's put a few languages through those filters:

      Ada95  Algol68  C      C++     Cobol    Fortran  Java    Pascal  PL/I
1)    Yes    Few      Yes    Most    Most     Some     Most    Some    Few
2)    Yes    No       Yes    Yes     Yes      Yes      Soon    No      No
3)    Yes    No       Yes    Most    Most     Few      Yes     Few     Few

   The only things I see as controversial in these answers are the
mosts under C++ targets, which reflect experience in the embedded
market, and the most and soon under Java.  If you are willing to
accept g++ as a viable C++ implementation, it is almost as widely
available as Ada.  But then I would have to extend the same "courtesy"
to Pascal and Fortran, where it is IMHO not true.

   But in any case, the reality is that in most cases if you want
widely portable code, you use Ada or C.  (If it needs to run on 8-bit
targets, you probably are forced to C, although there are some "8-bit"
chips with Ada compilers available.)  If you are targeting just data
processing environments, then COBOL is an alternative.  Similarly for
desktop computing and C++, or number crunching and Fortran.

   Pascal, PL/I and Algol all were nice languages once upon a time,
and there are still compilers for some popular platforms.  But as
choices for major projects they are as dead as IPL-V or Autocoder.

   Will C and Ada continue to be the only possible choices for
widespread applications in the foreseeable future?  With the
exceptions noted above, probably.  Some people think that C++ and Java
will be potential choices in the near future, but I just don't see it
in the embedded market.

   On the flip side, how long will C and Ada continue to be valid
choices?  I can't see a time when they won't be.  There is enough
momentum behind both that when new hardware is announced there will
almost always be working compilers for one or both languages
available.  If not, both will probably be available before beta
hardware.  (As it happens I do know of a couple chips without a good C
compiler, but the only new chips I can think in the past ten years
without an available Ada compiler are some members of the TI TMS320C
family.  Chips without a good C++ compiler?  Dozens.  Without a good
Java compiler?  Weeelll, I could say all of them, but there do seem to
be some good Java COMPILERS showing up. I just haven't tried any out.)


--

					Robert I. Eachus

with Standard_Disclaimer;
use  Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...