comp.lang.ada
 help / color / mirror / Atom feed
From: MFELDMAN@GWUVM.BITNET (Mike Feldman)
Subject: Ada compilers for IBM PC
Date: 6 Feb 88 15:45:00 GMT	[thread overview]
Message-ID: <8802061828.AA25233@ajpo.sei.cmu.edu> (raw)

I am aware of three full-language Ada compilers for MS-DOS machines.

AdaVantage - Meridian Software Systems - 800-221-2522

This compiler was validated under ACVC 1.8. It requires a 640k machine
with hard disk. Current release requires a coprocessor for floating
point arithmetic (not fixed point); a forthcoming upgrade will support
software-emulated floating point. I have used versions of this compiler
for about 1 year. It is sound and pretty fast at compile time (I don't
have specific benchmark numbers - don't believe in them).

My students have encountered a few bugs but they are not crippling, are
relatively subtle things, and are, I think, easily fixed in later
versions. I've been quite happy with it overall. Nice DOS libraries
are available for, I think, $65.00. Price is $795; a starter version
with some program-size limits (acceptable for learning and reasonable
term projects, etc.) can be had for $99.

Janus-Ada - RR Software - 800-PC-Ada-4U

Latest version passed ACVC 1.9 tests in December - they say they are
waiting for their validation certificate. I have the $99.00 starter
version, plan to order the full-scale one. The differences are mainly in
support for "real" programming - DOS libraries, etc. The $99. one has a
64K limit on executable image size. Generates .COM, not .EXE file.
Full-blown versions at $395. and $799. are available.

Janus-Ada has been around for years in different versions. They have gotten
a (largely undeserved) bad press. This release supports all of Ada, and
works, as far as I can tell. Any compiler will eventually show small and
subtle bugs, easily coded around, as fixed in subsequent versions. We all
know that validation does not guarantee perfection.
I believe that Janus-Ada is a good and useful product now that it handles
all of Ada. Both it and AdaVantage are new and both will benefit from
the usual maturation process. But money spent on them is far from wasted.

Many of my students have bought the $99. version of one or the other and
have been quite happy. Both compilers need better development environments;
both will get them in due time.

The third validated compiler is the Alsys system. I don't have it; it
comes with a 4 meg memory expansion. Has a good reputation in the industry;
good industrial-strength development system, if you have the $$ to pay
for the additional hardware they make you get.

When reading magazines like Byte and Computer Language, keep in mind that
they have a lag time of perhaps 6 months between acceptance of an article
and when it hits the newsstands. The articles were probably _written_ 6
months before that. The Ada business is changing _fast_. Info-Ada is,
I think, an excellent vehicle for keeping up with breaking news.
I'm sure that Ada compiler vendors must get frustrated with the lag time in
the magazines, as reviews are based on _really_ outdated versions!

Michael B. Feldman                     residence address for SNail:
Professor, Dept. of EE&CS              Michael B. Feldman
The George Washington University       6218 Wagner Lane
Washington, DC  20052  U.S.A.          Bethesda, MD  20816  U.S.A.
202-994-7593
MFELDMAN@GWUVM.BITNET

             reply	other threads:[~1988-02-06 15:45 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
1988-02-06 15:45 Mike Feldman [this message]
  -- strict thread matches above, loose matches on Subject: below --
1988-02-04 17:33 Ada Compilers for IBM PC "BACH"
1988-02-01 20:51 Ada compilers " Ric Messier
1988-02-04  0:13 ` Daniel F. Conway
1988-02-05 19:26 ` Daniel F. Conway
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox