Re: GNAT Support Costs

[Robert Dewar <robert_dewar@my-deja.com>]

In article
<867lgtm6gz.fsf@ppp-162-235.villette.club-internet.fr>,
  Laurent Guerby <guerby@acm.org> wrote:
> "Rob Kirkbride" <rob@rk-comp.demon.co.uk> writes:
> > [...] For the project I am on at the moment all that support
is just
> > not required.  I would like just to buy a compiler, if I
find bugs I
> > would like to be able to report them. If a fix exists I
would like
> > to take that fix. I also don't need a support infrastructure
to
> > support 10 different people. [...]
>
> Hi Rob, I must admit I still fail to see why the public
version of
> GNAT doesn't meet your needs here. You pay $0 (or $x if you
get it
> from a cheap byte CD), you have an email to report bugs too,
you'll
> get fixes with the next public version (that you can buy too),
and if
> you're on Linux you have a dedicated team that is able to
produce
> Linux related fixes (also with an email). If you're stuck with
a
> problem, it's probable some experimented GNAT user (or
customer) will
> propose a workaround on mailing lists/newgroups.
>
> If it's the $0 tag or having to pay to some company that
doesn't
> improves GNAT directly that is a moral problem, you can always
> write a check to the FSF or to ACT (I don't know if they have
a
> special fund for this).

Please do NOT send contributions to ACT, we are not in the
contribution business. Send contributions to FSF instead,
which is indeed a non-profit organization partly supported
by donations.

If you are comfortable using unsupported software with no
support or guarantees from anyone, then indeed the public
version of GNAT may be just what you need. I would caution
that you can't necessarily rely on comp.lang.ada or the
chat list for help. Many questions go unanswered, and many
of the answers from volunteers are wrong or misleading, which
is not at all surprising, some of the issues are quite complex
(a good example of this is the thread on elaboration that
is current, the response is straigntforward here, especially
if you have read the GNAT documentation, but neither response
is really the helpful one so far :-) Furthermore, especially
with Ada 95, it is often hard for people to be sure whether
something is a bug report, or just a misunderstanding of the
language or the system. The majority of bug reports turn out
to be the latter!

ACT deals with clients who need fully supported software, often
they need formal validation (the public version of course is not
and cannot be validated unless you do it yourself), and they
find the high level of support that we can provide valuable (as
we see from the very high rate of support renewal from our
customer base). Part of the value of this support is precisely
that you can ask questions without worrying about whether they
are Ada questions, GNAT questions, or real bugs.

Yes, we quite understand that some folks would like a GNAT based
product with a lower level of support. That's quite a reasonable
idea, it is just not the business that ACT happens to be in.
Nothing is stopping anyone else from providing this service
(whether you can make a viable business doing it is another
concern, but that's a separate issue). Labtek did in fact try
to make a go of it in the NT area doing precisely this, and we
were happy to cooperate with that effort, but they abandoned
it precisely because it was not a viable business.

Robert Dewar
Ada Core Technologies



Sent via Deja.com http://www.deja.com/
Before you buy.