Re: SGI GNAT Question? (Long)

[dewar@gnat.com]

In article <36E2531E.37AEECF1@chocolatesaltyballs.com>,
  root <white@chocolatesaltyballs.com> wrote:

> someone who works for a company who have bought 3.11b2,
> should do the honourable, get permission from their
> company and post it to alt.binaries.source.

In fact 3.11p and 3.11b2 are indeed virtually identical
(differing only in stuff related to the version name.
When we make public releases they are not somehow crippled
or incomplete, they are the full code at the time of
release.

As to whether it is viable for serious mission critical
projects to use unsupported software, that is something
that the project has to decide for itself. Our viewpoint
that the public version is more suited to student and
research use is an issue of support not bits!

It is true that some of our supported customers are using
more advanced versions of GNAT technology that is in alpha
or beta testing status. We generally ask that such test
versions NOT be distributed, since it definitely is NOT
helpful to have dozens of slightly different versions of
GNAT around, which have not been fully tested.

It would not be illegal for a company to distribute a test
version in this situation, but it would be unhelpful to
the community, and I doubt "honorable" would be the right
term for this :-)

In fact this is an old practice in the open source
community, and is used everywhere. For example, some
of Cygnus' customers have access to Cygnus internal
developments which are beyond the EGCS development, but
they are in a stage where it is not helpful to generally
distribute them, since work is needed to integrate them
into the current EGCS sources. Cygnus will do this work
in due course, so this technology will become available.

Richard Stallman has been quite comfortable with this kind
of request and understanding, and the FSF has often
operated this way in the past, and it has worked out well.

Hackers and hobbyists are often in a mode where they want
the latest and greatest whether or not it is fully tested
and working, and that is understandable.

However, one of the great fears that the "real world" has
of free/open source software is that it is out of control,
and that it is a free for all where no one knows what
version they have of anything or what's in it.

In actual fact, the important developments in open source
software (EMACS, GCC, GDB, LINUX, GNAT etc) are in fact
managed in a much more responsible way, and there are
stable well tested periodic releases, in much the same
manner as for proprietary software (well I think we do
a little better than Microsoft in getting out new versions
of GNAT :-)

P.S. GNAT 3.12 technology is on the way. We have most of
the work done for this new release, which has many new
interesting capabilities, and we are working towards an
early release, which will include a public version 3.12p.
We have no definite schedule yet, but we can say that the
wait for 3.12p will be much shorter than the 3.10p to 3.11p
gap (which was too long and too big a jump!)

Robert Dewar
Ada Core Technologies

P.S. It's interesting, we have variations on this thread
every now and then, typically sparked by a new GNAT
enthusiast who was not around for the old threads. Never
mind, it is a good opportunity to sort out confusions with
the GPL.

One of the important accomplishments of the last few years
is that now almost all major companies are MUCH more
comfortable using open source software than they used
to be!


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