comp.lang.ada
 help / color / mirror / Atom feed
From: dewar@gnat.com
Subject: Re: SGI GNAT Question? (Long)
Date: 1999/03/05
Date: 1999-03-05T00:00:00+00:00	[thread overview]
Message-ID: <7bos1q$ogq$1@nnrp1.dejanews.com> (raw)
In-Reply-To: 36DFA6FB.D3A2AD84@spam.com

In article <36DFA6FB.D3A2AD84@spam.com>,
  spamwithchipsplease@spam.com wrote:
> dewar@gnat.com wrote:

> From later comments in your post re the interpretation of
> the letter versioning, you could go further and state
> that 3.11b2 "IS" 3.11p in source tree and binary build
> terms.

No, we would not go further and make any such statement.
We make no statements at all about the 3.11p version, since
we have no control over it once it is out there. We only
make guarantees with respect to our supported commercial
technologies.

Clearly the source tree is NOT identical, it is at the
least different with respect to the version number embedded
into the source at several points. The most we will say
is that 3.11p is based on 3.11b2.

> Specifically, you have clarified that the commercial GNAT
> is GPL code, ACT irrespective of their commerical
> commitment to open source, are compelled by the terms of
> GPL to release any and ALL future versions of GNAT under
> the GPL

This is quite wrong in two important respects.

  The GPL never forces anyone to release anything. If ACT
  decided that all future versions of GNAT will be private
  to ACT and used only within ACT, that would be completely
  consistent with the GPL.

  Similarly, if ACT decided that future versions of the ACT
  copyrighted components of GNAT were to be released in
  fully proprietary form, that would be perfectly
  consistent. Giving someone a license to your copyrighted
  creation does not place limitations on YOU, the author!

These are indeed common misconceptions. In particular, I
have talked with several companies recently who were quite
surprised to find that they could do the following:

   1. Issue an open source version of software X under GPL

   2. Issue a deriviative work that was fully proprietary

FOr example, you could have a crippleware product that was
under the GPL, and the fancy version with bells and
whistles as fully proprietary.

Cygnus does something a little like this with Cygwin. The
public version is under the strict GPL, which means that
it CANNOT be incorporated into proprietary programs. If
you want this kind of incorporation you have to buy their
proprietary version, which, to a first degree of
approximation, is identical except for the license. This
is perfectly legitimate.

In fact ACT is committed to making future versions of our
technology publicly released under the GPL or GNAT-modified
GPL (GGPL) as appropriate, but this is a result of
corporate policy, there is nothing in the GPL that requires
this.

Here are two useful things to remember:

  The GPL NEVER requires anyone to distribute a program
  under ANY circumstances at ANY time. It does place
  restrictions on you if YOU choose to distribute.

  The GPL is just a license. Open source software under
  the GPL is like any other software on the market. It
  is copyrighted software to which you are granted a
  limited license. Granting a license to people in no
  way restricts the copyright holder's ability to do
  anything they like with their own work.

> Strange that you inflict your "Beta" versions exclusively
> on your paying customers. I think the success of open
> source has been based on public releases feeding back to
> the developers bug reports.

Not so strange. Beta versions of this type are given only
to customers who request them, and who have full support
should they run into any problems. That makes a big
difference. What may be a small installation glitch in
a version given to a customer, where the problem can be
solved instantly under their support contract, may be a big
problem for a general public release. The criteria for a
public release, to be used by lots of people with no
support are different.

> Having  run a GNU/linux system for 3 years, I like many
> others are use to "feature-rich" pre-releases.

More than you think, development of GNU and Linux software
happens on private trees that do not begin to be public. In
fact they are private enough that often people are not
aware of them.

I often meet people who think that EGCS is the Cygnus
development tree. In fact of course it is not, there are
major developments going on internally in Cygnus which
they do not tell the outside world about.

The difference with ACT is that we tell people what is
going on, and share our future plans much more openly.
But no major open source development that I know of is
much different. There is a public version that represents
technology that is one step behind the current development
technology. In this respect open source software is really
very little different from normal proprietary software.

The other point here is that the public releases of GNAT
are quite conciously aimed at the large mass of Ada users,
NOT at hobbyists and enthusiasts who want to fiddle around.
That creates a rather different market place. By far the
most common use of GNAT is by beginning students, and the
requirements here are quite different from those of many
CLA readers :-)

> As while I wouldn't go so far as to say ACTs position as
> both commerical company and GPL code developer is unique
> in the GPL world, they are certainly in a minority of GPL
> developers.

If you think that, it is probably because you do not know
what is really going on!

It would certainly be nice to see more activity in the
publicly released tree. ACT itself does not have the
resources to support the kind of active integration of
changes that we see with EGCS (Cygnus budgets a substantial
amount of resources for this purpose). Marcus Kuhn and
others are trying to setup a similar environment for GNAT
on a volunteer basis, and in addition, the GNAT front end
will be integrated on some basis into EGCS (we are still
discussing with the EGCS folk how to do this effectively).

But not all GPL products have such an active public tree
by any means. For example, GDB does not, and this is quite
a problem, because, unlike the case with GNAT, there are
several major companies doing major work on GDB, and there
is no effective way to coordinate that work at the moment.

Robert Dewar
Ada Core Technologies

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    




  reply	other threads:[~1999-03-05  0:00 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
1999-03-02  0:00 SGI GNAT Question? (Long) Paul Colvert
1999-03-02  0:00 ` David C. Hoos, Sr.
1999-03-02  0:00   ` GNAT discussions should be here as well kvisko
1999-03-02  0:00     ` robert_dewar
1999-03-02  0:00     ` dennison
1999-03-02  0:00     ` Samuel Mize
1999-03-02  0:00     ` Mike Silva
1999-03-02  0:00     ` Larry Kilgallen
1999-03-02  0:00 ` SGI GNAT Question? (Long) dewar
1999-03-03  0:00   ` Paul Colvert
1999-03-03  0:00     ` robert_dewar
1999-03-04  0:00       ` SpamSpamSpam
1999-03-04  0:00         ` dewar
1999-03-05  0:00           ` SpamSpamSpam
1999-03-05  0:00             ` dewar [this message]
1999-03-05  0:00               ` dennison
1999-03-05  0:00                 ` robert_dewar
1999-03-07  0:00                   ` root
1999-03-07  0:00                     ` dewar
1999-03-08  0:00                       ` root
1999-03-09  0:00                         ` Some GNAT history (was Re: SGI GNAT Question? (Long)) dewar
1999-03-09  0:00                           ` Tom Moran
1999-03-09  0:00                           ` dennison
1999-03-09  0:00                             ` robert_dewar
1999-03-11  0:00                           ` Arthur Evans Jr
1999-03-11  0:00                             ` dennison
1999-03-09  0:00                         ` SGI GNAT Question? (Long) dewar
1999-03-10  0:00                           ` SpamSpamSpam
1999-03-10  0:00                             ` Chris Morgan
1999-03-10  0:00                               ` dewar
1999-03-10  0:00                                 ` Chris Morgan
1999-03-10  0:00                                   ` dewar
1999-03-10  0:00                             ` robert_dewar
1999-03-07  0:00                     ` David Botton
1999-03-07  0:00                       ` robert_dewar
1999-03-05  0:00             ` bourguet
1999-03-05  0:00               ` dennison
1999-03-05  0:00                 ` dewar
1999-03-05  0:00             ` dennison
1999-03-05  0:00               ` dewar
1999-03-07  0:00               ` root
1999-03-07  0:00                 ` dewar
1999-03-08  0:00               ` Marin David Condic
1999-03-05  0:00             ` GNAT Field Test scope (was SGI GNAT Question) Larry Kilgallen
1999-03-04  0:00         ` SGI GNAT Question? (Long) dennison
1999-03-02  0:00 ` Gautier
replies disabled

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