ATT Unix license

[HEDRICK at RUTGERS.ARPA]

From:  Charles Hedrick <HEDRICK at RUTGERS.ARPA>

How do other Universities interpret the provisions in the Unix license
referring to availability of the results of research?  We are just
getting into Unix.  One of our major research projects is doing work in
collaboration with an industrial lab.  While our research itself is
public, they will no doubt be hacking on software from this lab, and it
is not public.  The researchers are worried that they are going to be
placed in an unresolvable conflict between provisions of licensing
agreements with this lab and with ATT.  Here are the alternatives I can
see:

 - every piece of software on the system must be in the public domain.
	This is obviously crazy.  It wouldn't allow us to buy a compiler
	from a 3rd party.  Clearly we must be able to run proprietary
	software.

 - piece of software that we work on must be in the public domain.  This
	makes a bit more sense, but would mean that we couldn't fix bugs
	in proprietary software.  I hope this isn't what is meant.

 - the substance of our research must be public.  This would allow us to
	work on proprietary software when it is needed for our work.

 - the term "results" as used in the ATT license refer to the actual
	scientific findings.  These are normally embodied in
	publications in journals, and in working papers. Those are of
	course public. Any software that we develop is merely a tool for
	conducting the research, and is not the actual results.  This is
	the normal interpretation that government agencies place on
	research, and is the reason why software developed under
	government support is no longer required to be in the public
	domain.  The only thing that this would rule out is a project
	whose primary goal was software development.  As long as the
	software could be represented as being secondary to a scientific
	goal, it would not be considered the results of the research.

My inclination is to adopt the last alternative.  As I understand the
law, ambiguous terms are interpreted in the light of current practice.
I believe that current practice in research contracts considers 
research to be non-commercial if the scientific results are published,
but does not require that software developed as a side-effect be
freely distributed.
-------