Trying to release old versions of Unix
This is the text of a letter I received from Dennis Ritchie after he saw this web site and read about the process I went through to get VisiCalc posted. Reproduced here with permission.
The first real attempt to release Unix for less-controlled
use was with our early research versions,
(say up through Seventh Edition and 32V). I began the
attempt a while after USL had been spun off,
I think during the USL vs BSDI and then UC Berkeley
lawsuit. I had the Intellectual Property agreement
between AT&T and USL, and to me it looked fairly clear
that what (amongst other things) was being turned over
was rights to code, trademarks, etc. in System III and
successors. So I thought I might pull a fast one,
and get a release from AT&T for the code for the
research systems, since (it seemed to me) that we
owned it. This looked like fun, and might short-circuit
the bad press AT&T was getting about the suit, even
though USL was mostly independent by then.
There was a several-year gap between the research
systems I was interested in, and that named in
the agreement, and it seemed that what I asked
for was separate from the USL product.
The lawyers basically agreed that my reading was
probably correct and defensible, but one phrase,
which was approximately "and test versions...",
seemed to give an anchor for USL to argue
that the early research systems were in fact covered.
Particularly since there was a lawsuit in progress,
the AT&T lawyers were probably well-advised not
to get involved in a sudden claim that the starting point
of the BSD system was now somehow released.
Once USL was sold to Novell, I tried again directly
with Novell, making discreet inquiries. The suit
had been settled by then. The folks I talked to
unanimously thought that they actually couldn't
see what harm it would do (and generally thought
it a nice idea), but the Novell lawyers had some qualms that
were never really articulated; no one wanted
to be the one to sign. (Your account to Winer
is reminiscent here).
The next thing to happen was that John Lions,
the author of the commentary on the Sixth Edition
system, was becoming ill with a progressive neurological disease.
Prompted by Peter Salus and Berny Goodhart,
I asked about a republication of Lions's "suppressed"
work from about 1975. The answer from Novell
was about the same (good idea, but Something Bad
Might Conceivably Happen).
Finally, USL was sold by Novell to SCO. I wrote to
Mike Tilson, an old acquaintance from Usenix
and elsewhere, and a VP of the company. He was interested,
though when I first wrote, the sale had not been
consummated and the company couldn't think
about such things then. He did introduce the
idea to Doug Michels, closer to the center of
the Unix acquisition. We kept working on this;
in fact Lion's book was republished by Peer-to-Peer
in 1977, barely in time for John to appreciate
that this had occurred. (He died last year.)
More recently, SCO has been issuing licenses for
the early Unix versions, including source (not
for free and not without a bit of hassle, though).
But better than nothing.
The release of the C compilers was accomplished
much more easily--I just did it. The Unix kernel
code in particular has had at least one lawsuit
associated with it, and there are traceable components
that extend even until today, but this compiler isn't
used any more, and it is less charged with IP issues.
I figured there was no one outside who could possibly
be worried, and that I could deal with any inside issues.
Side observations--the timing of your and my releases
seems coincidental. It was interesting to see the SJMerc
article and the others. (I got a relatively huge hit count
on the home page in the days after it and the /. article).