Re: GNAT GPL 2005 Edition is now available

["Steve Whalen" <SteveWhalen001@hotmail.com>]

Jeffrey Carter wrote:
> Samuel Tardieu wrote:
> >
> > Do you think that the DoD (the funder of the early GNAT versions)
> > intent was to restrict the use of GNAT to programs using the GPL
> > license and to AdaCore (which didn't exist) customers?
>
> The DOD program that resulted in GNAT was created specifically to build
> a freely-available compiler that could be used to create proprietary SW.
> The issue of the run time, including parts of the run time that might be
> added to a program by instantiating a generic, are why NYU's lawyers
> came up with the GMGPL in the first place.

Thanks to you and Samuel for the posts.  I was more active in this
group from about 1994 to 1999 and have mostly been "lurking" since
then (not enough time to get involved in most discussions, though I
still care about Ada). I'd read all of this discussion and was having
a hard time figuring out why I was so mad at AdaCore, but your
reminder of some of the history of GNAT helped me figure it out.

I do believe AdaCore (or any company) has the right to choose what
they do with what they create.  But I think this move is really bad in
several ways. Bad for Ada. Bad for AdaCore. Bad for free software
(including GNU/FSF because of the licensing confusion this will cause:
only programmers will understand the significance of a run time
library and it's license, but Microsoft and others will have another
opportunity to spread confusing FUD about the GPL).

You reminded me that DOD and NYU and Robert Dewar and the others were
in many ways doing a very "good thing" by creating a free Ada compiler
to help offset the removal of the Ada "mandate" from within the DOD.

What bothers me about this move by AdaCore is that Robert and AdaCore
seem to have forgotten a lot of what they said back at the beginning
of GNAT.

10+ years ago Robert and the other AdaCore founders were making a lot
of very sensible statements of practice, principle and philosophy back
when GNAT and AdaCore were being "born", including:

1) GNAT would be good for the industry, and good for Ada. A free
compiler would encourage teaching and use of Ada in places it would
otherwise not reach.

2) There would always be demand for "validated" compilers for DOD /
safety work and for high quality support for the toolset so AdaCore
(or whatever it was called at the start) would be able to make enough
money to survive.

3) GNAT would be good advertising and not cost AdaCore any serious
money beyond the work necessary to "package" the non-validated
releases.

4) There was a lot more but that's enough for now (Anyone who wasn't
hanging around comp.lang.ada back in the mid 1990's I invite to go
back in google groups and read what Robert and the other AdaCore
founders were saying back then).

It all made sense then, and nothing has changed to invalidate it.

I've realized that I'm mad because in a way, I'm feeling betrayed.
Robert Dewar and the AdaCore people are going back on what they said
they were going to do, and why they were going to do it. (I repeat, I
agree AdaCore has the absolute right to do what they are doing, but I
think it's a really big mistake on their part and certainly NOT in
keeping with the pronouncements they made repeatedly during the first
five years of GNAT's existence ... this seems more like something I'd
expect from a Microsoft type company than from GNAT's creators).

It makes me wonder who's running AdaCore.  Have the founders been
pushed aside? Are they having so much trouble meeting their payroll
that they're getting desperate? Have they forgotten all of what they
said when GNAT was just getting started? I was glad when they founded
AdaCore (or whatever it was called at the beginning) because it seemed
appropriate that those who'd put so much work into getting GNAT off
the ground should get something back from it.  Suddenly it seems
they've forgotten all they said they believed in ten years ago.

Worse, I'm afraid it's a sign AdaCore no longer has confidence in it's
ability to attract customers to GNAT by their high quality of service
and depth of knowledge and support. That does not bode well for the
long run, for either AdaCore or Ada.

Steve