Re: Towards a free GNU Ada

[Robert Dewar <dewar@GNAT.COM>]

<<I am forced to use GNAT because the other Ada vendors for Windows have not
completed the Information Systems Annex.  I still believe that one of the
weakest points of all of the Ada compilers that I have used is the error
messages.  I applaud Professor Dewar's suggestions about future
>>

A very odd comment, we have received almost no input from you commenting
on specific error messages. To me, GNAT's messages are among the best I
have seen from a compiler, and we put a huge amount of effort into them,
aided greatly by users who submit cases where messages could be improved
(special thanks to Mars Gralia here).

Don't just be a non-constructive complainer!

Send along examples where you think messages can be improved!


<<Robert Dewar has been critical of my suggestions as to how to increase
ACT's revenues.
I suspect that many of us, at present, who use GNAT would like to
compensate ACT. However, ACT's pricing structure is an example of what I
like to call the human nature boolean type, too much or too little.
>>

To run a company like ACT costs well over a million dollars a year, running
any small business is not an inexpensive operation. FOr the kind of support
we provide our customers, the prices we charge are highly competitive.

Yes, I know that people would like to think that they could get full GNAT
support for very modest sums (a while ago someone suggested that ACT should
provide a service where bugs would be fixed for $20 a piece -- enough to
pay for about 15 mins of time -- which of course is way off.

<<2. * Form a consortium of GNAT users, with annual dues which will be paid
to ACT to provide public support for GNAT.
>>

This is not an unreasonable proposal, but our guess is that there are not
enough users, and/or they are not willing to contribute enough, for this
to be worth our while (the conversation I referred to, the $20/bug one,
was enthusiastically followed up by lots of people saying, yes that would
be great!)

The other trouble is that we really work by guaranteeing our customers
a level of support that guarantees success. This is not inexpensive, but
it is worth it to those who need it, and the resulting costs are still
mcuh less than many competing technologies. Any kind of partial low level
support would not meet this goal.

There is also a real danger that serious users would try to get by on
this lower cost partial support, run into troubles, and end up frustrated
at GNAT in particular and Ada 95 in general. This is not a desirable outcome
for anyone.

Furthermore, these days, a great deal of our support activities are not
related to bug fixing, but rather answering peoples questions abo0ut how
to use GNAT and how to use Ada 95. In the context of the large applications
that people are developing, this can be a time consuming and costly operation.

Robert Dewar
Ada Core Technologies