Re: State of opinion of GNAT

Re: State of opinion of GNAT

[Chris Morgan]

In article <ouybrk38qa.fsf@i486.mcneil.com>, sean@mcneil.com (Sean McNeil)
wrote:

> When I joined this newsgroup a few years ago, there was a lot of talk
> about GNAT and how it is just a toy compiler for the academic environment
> and couldn't be taken seriously.  After all, it doesn't even have a
> library manager!
> 
> I have been working with GNAT for a while now and actually prefer it's
> environment.  Also, it is apparant that common opinion of GNAT's
> usefulness has completely turned around since those early days.  I
> was just wondering when (and how) these opinions were changed?

  I can't speak for others. My interest was sparked initially because of
the price and because it was always an Ada95 compiler. These two facts
caused me to investigate the possibility of downloading it and installing
it. To do that I had to get Internet access, a Unix account and learn
basic Unix (since don't have a PC and it doesn't run on VAXes). Once I had
done all that, the compiler and documentation speaks for itself i.e. was
well worth all that effort. The GNATDOC1.PS file is especially interesting
as it demonstrates how the target and host independence is achieved. Once
one has read that, one can see that irrespective of the current status of
the compiler (quite incomplete when I first got hold of it), the
architecture is uniquely powerful due to the GCC underpinnings.

  I think being a real 'fan' of GNAT probably also depends on
understanding and accepting the GNU ideal. Some people object to it, and
that's fine. If you are an Ada fanatic (which I am I suppose) and accept
the GNU movement's aims, then one cannot but be a fan of GNAT as well,
especially now it's virtually complete.  I think proprietary software has
an entirely rightful place in the market, but gradually the free software
movement can fight back against over-dominance of the market by certain
vendors and I am pleased by that. I work for a big company that buys
expensive tools, but I have the feeling that my Ada skills are more
portable entirely due to te existence of GNAT. It's obviously nice to see
products like ActivAda Personal edition for $99, and I expect GNAT had a
hand in allowing Dave Wood at Thomson to justify that price to the
accountants :-) (just a guess)

  Incidentally, last week I wrote up a document proposing limited use for
GNAT for real work on my project (limited in the sense of not yet
producing customer deliverables). When I went through it, it was actually
very easy to make a case for using the Solaris version as its so nearly
complete. I also got less resistance than I expected because Ada95 has now
loomed onto the horizon of senior managers for bids and future projects,
whereas back in 1994 it was just too far away to be an issue.

I now prefer the environment as well.

Chris

-------------------------------------
-- Chris Morgan, BAeSEMA Limited   
--   chris.morgan@baesema.co.uk    
-------------------------------------
--           Team Ada              
-------------------------------------

Re: State of opinion of GNAT

[Robert Dewar]

Chris says

"  I think being a real 'fan' of GNAT probably also depends on
understanding and accepting the GNU ideal. Some people object to it, and
that's fine. If you are an Ada fanatic (which I am I suppose) and accept
the GNU movement's aims, then one cannot but be a fan of GNAT as well,
"

I don't think this is true at all. GNAT is simply a tool, you should use
it if and only if it meets your technical needs, and the price is right,
taking into account whatever support you might or might not need (presumably
if you don't need support, the price *is* right :-)

Yes, there may be some people who will use GNAT because it is free software
and the "GNU ideal" appeals, and there may be some people who will not use
it because this ideal does not appeal.

But for the great majority of people it makes more sense to simply regard
GNAT as a tool which may or may not meet your needs, using or not using GNAT
really has nothing to do with whether you understand and accept the GNU
ideology.

Yes, it is *because* of this idiology that the technology exists in 
its current freely available form, but from a users point of view you
can simply take advantage of this and that's that!

In practice, people will find that some combination of GNAT with proprietary
tools, packages, applications, interfaces, bindings etc and with free
such tools etc. will be what they need, and such a mixture is perfectly
reasonable and possible!

In some cases it will even make sense to combine the use of proprietary
Ada compilers with GNAT. FOr example, in the DEC VMS world, the Ada 95
compiler will be GNAT-based, but some DEC Alpha users may prefer to
stick with Unix for development purposes, even though eventual delivery
of the product is on VMS. Such users have two choices, they can either
use GNAT on DEC UNix, or they could use the Rational APex environment
on DEC Unix, and then use the export facility of Apex to export the
code for recompilation by GNAT.

The Ada 95 environment will hopefully get richer as time goes on with
the addition of both free and proprietary tools, bindings and compilers.
Ada 95 users will then have a choice of approaches, which seems attractive
from two points of view. First there is nothing like competitoin to sharpen
up everyone's techology, second if you have a choice of tools, then you are
more likely to find exactly what you need.

Re: State of opinion of GNAT

[Larry Keeler]

Sean McNeil <sean@MCNEIL.COM> wrote:

   "When I joined this newsgroup a few years ago, there was a lot of talk
    about GNAT and how it is just a toy compiler for the academic environment
    and couldn't be taken seriously.  After all, it doesn't even have a
    library manager!

    I have been working with GNAT for a while now and actually prefer it's
    environment.  Also, it is apparant that common opinion of GNAT's
    usefulness has completely turned around since those early days.  I
    was just wondering when (and how) these opinions were changed?"

I suspect that the when was about release GNAT 1.7 and the how was the large
volume of traffic on comp-lang-Ada about it by those willing to try the early
versions and the willingness and large amount of information about it which was
provided by Robert Dewer and others on the GNAT team at NYU.  (Of course the
fact that they made GNAT a really serious compiler helped a lot :-)
In addition, I think that the efforts by others outside of NYU also played an
important part, especially those of Michael Feldman, for his work with the DOS
version (and whoever it was that did the DJGPP part to make a DOS version even
possible), and to other members of Team Ada like David Weller who, like Mike
and Robert have contributed to making comp.lang.ada lively and informative. (A
certain cleric will be omitted. ;-)

Of course Tucker Taft and all of those responsible for Ada95 also had a part to.
(I cannot imagine a swell of respect developing for a BASIC compiler.)  But I
ramble.  (Although I don't suspect that you expected a concise answer.)

Larry
Team Ada