[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: PD status v.s. DARPA?


Rob passed your query on to me.  He has taken over a good chunk of the
technical load for CMU CL, but I'm still running the project.

DARPA has paid for our whole project, except for a bit of tangential
support from IBM in the form of RT workstations.

CMU Common Lisp has been public domain since the earliest releases of Spice
Lisp on the Perq -- maybe 1985 or so.  We thought for a while about trying
to do some sort of license, either a "copyleft" or perhaps even trying to
make money from commercial users like DEC, H-P, and Xerox who at various
times have product-level Lisps of their own on top of the base we provided.
But CMU was unwilling at the time to procide us with the necessary legal
advice, so we just decided to put it all into the public domain and keep
the damned lawyers out of the loop.  Of course, we've been careful not to
inadvertently include any code copyrighted by others.  (We do intentionally
include and acknowledge a couple of modules that are freely usable but that
are copyrighted by others: PCL from Xerox and Dick Waters's pretty

There's no issue with CMU.  Our intellectual property policy explicitly
allows authors to put their work into the public domain, absent any
explicit agreements to the contrary.  Now some of the CMU administrators
see this and wonder if they should have tried to make money from it, but
they still don't have a decent organization on campus advising people how
to protect and exploit their work (with CMU getting a cut).  People keep
recommending that they set up something like this, but the university
administration figures it might cost money to do this, and they don't think
about the long-term possibilities.

As far as DARPA is concerned, I don't think they've ever explicitly told us
it was OK to go public domain, but they've known about our plans at every
step of the way and have never objected.  ("They" in this case are
Ohlander, Squires, Scherlis, Boehm, and others more recently.)  Remember
that when we started down this road, the big question at DARPA was whether
the Lisp community could ever settle on a single "common" dialect, and I
always pointed out that the existence of a professional quality,
public-domain Common Lisp would go a long way toward solving that problem.
They always seemed pretty enthusiastic about this vision.

We have had some hassles about shipping our software outside the U.S.  At
various times, I've heard third-hand about requests from DARPA that we not
send any DARPA-funded software outside the country.  But when this issue
first came up, copies of our sources had already reached England, Germany,
and Japan, either via FTP, visitors taking back a tape, or U.S. companies
passing the code along to foreign associates.  I must have tried ten times
over the years to get a clear, official answer from someone at DARPA about
how they wanted us to handle this.  All such queries have gone unanswered.
So we took the position (and told DARPA, in case they wanted to object)
that we weren't going to mail tapes outside the U.S., but that the code was
available via anonymous FTP and if foreign companies got hold of it, we
would be happy to answer their questions.  Again, no complaints from DARPA.
As it happens, most of the beta testers and one of the developers for the
SunOS version are in Europe.

So that's the story.  I would be interested in hearing about the
experiences you have had with DARPA.

-- Scott