GU: Control of distribution

[adler@pulsar.cs.wku.edu (Allen Adler)]

In reply to my critical comments about the control which Mike Schmidt
exercises over gamess distribution, I have received a lot of email
from people defending the policy. In no case are the reasons the
one which Mike himself gives, namely that he wants to keep it
out of the hands of governments who don't respect human rights.

The main argument I have received is that it is impossible to maintain
quality control without having distribution control. Here, I would
like to point out that the Free Software Foundation manages to have
free distribution of its GNU software but nevertheless manages to
have quality control. For details, see the GNU Public License
Agreement, which is distributed with all GNU software.

A secondary argument I have received is that there is no need for
freeer distribution since no one is inconvenienced by the present
controlled system of distribution. I have already described how
stifling the present system of distribution has been for my own
efforts. So let us agree that some people are indeed inconvenienced by
it.

I don't pretend that my own projects are of such importance that
Mike or anyone else ought to drop everything and start thinking
about them. Nor do I exaggerate my own ability to contribute
to gamess; in fact, I am well aware of my limitations. But I
maintain that interested persons should be free to experiment with
the gamess code and with the gamess documentation and to distribute
freely the results of their experiments provided that:
(1) they do not represent their results as the official version, and
(2) they DO include the complete original version along with their own
    creation, clearly designated as such.

One of the advantages of this is that it would be possible to
distribute improvements (such as ports) of obsolete versions,
even though those versions are no longer distributed by Mike.
As long as the versions are clearly indicated, that ought to be
perfectly legitimate. I'm sure plenty of people on this list are
doing their research with obsolete versions of gamess.

If Mike were to go over to an ftp mode of distribution, he
could have a subdirectory for user contributed software
and literature, distributed with no guarantees on his part
(and that caveat clearly explained in a README file). He
could also keep the last few official versions available
in some other directory, so that if for some reason someone
wants to work with an older version, they can do so.

That way, nothing goes into the official distribution unless
Mike wants it there, but the user community is not hamstrung
by excessive control over distribution.

It works for GNU software and I think it would work with gamess too.

Allan Adler
adler@pulsar.cs.wku.edu