Unix copyrights (was: Retypesetting Mathematics now on GitHub)

[G. Branden Robinson]

At 2023-06-02T21:54:21+1000, Damian McGuckin wrote:
> On Thu, 1 Jun 2023, G. Branden Robinson wrote:
> 
> > In keeping with the recent focus on eqn on this mailing list, I
> > thought I'd put my "Retypesetting Mathematics" undertaking of 11
> > months ago on more permanent footing.
> > 
> > https://github.com/g-branden-robinson/retypesetting-mathematics
> 
> Nice job. Very nice job.

Thanks!  :)

> Don't you need to have AT&T copyright in the files?

I don't honestly know.  They didn't come that way (I got them from TUHS
as noted in the first commit).  These documents are so old that they
might have come into existence still under a legal regime where if you
failed to include a copyright notice when creating the work, you
forfeited the right to sue for civil damages.  And since these were
inputs to a computer program, in the murky waters of U.S. copyright law
of the 1970s, they might not have been considered copyrightable at all.

As yet another factor, this document and its revisions straddle the
Copyright Act of 1976, before which a work had to be "published" to
enjoy copyright protection, and after which it merely needed to be
"fixed in a tangible medium".  Since CSTR #17 (the first edition of this
document) was published in CACM in about 1974, that condition would seem
to be fulfilled for its rendered, typeset form.  But the 1976 act did
not contemplate computer programs.

Further, I don't know if the documentary assets of the Unix system were
treated as part of the Unix source code throughout the complex history
of its the code's ownership.  They could well have been overlooked,
because Unix was successively acquired by firms looking to extract
license fees from deployments of the operating system.  If no one
considered that they could extract significant economic rents from the
papers in the Volume 2 of the Version 7 Unix Programmer's Manual--and I
think the case for such would have been weak even in the 1980s--then it
would not have been efficient for these firms to direct their attorneys
to expend effort ensuring that these assets remained part of the
package.

These things appear neatly bundled together to _us_, and to people like
TUHS archivist Warren Toomey because the engineers who prepared tape
archives of the Unix system were conscientious.  And in turn, we
research and consult these things for love, not money.  There is no
reason to assume the rights-holders in these works exercised similar
care, and every reason to believe they didn't.

If you want to simulate the decision processes of an executive, you must
adopt the attitude of a sociopathic mercenary.  Pick a number and make
it go up (or down).  Anything that does not move that number is
irrelevant.

Anyway, Diomidis Spinellis would appear to be a subject matter expert.
He keeps a copy of the Caldera's open letter applying, essentially, the
4-clause BSD license to historical Unices (restricted to "16-bit Unix"
and "32-bit 32V Unix").[1]

As I understand it, there is therefore a series of questions that must
be answered.

1.  Did copyright attach to these files at the time of their creation
    and revision?
2.  If the answer to the foregoing in "yes", has copyright survived all
    these years (not a fact to take for granted before 1976 in the
    U.S.)?
3.  If the answer to the foregoing is "yes", did chain of title _in this
    work_ pass to Caldera?
4.  If the answer to the foregoing is also "yes", then I would be
    comfortable assuming the work to be under Caldera's 4-clause
    BSD-style license.

Two big lawsuits were fought over Unix copyrights: USL v. BSDI and SCO
v. IBM.  In both cases it was revealed--although the in the former case
only after many years once the private settlement agreement was
disclosed--that the copyrights on Unix files were stewarded negligently,
and, as I understand it, even valid copyright notices placed by the
Berkeley CSRG were effaced by AT&T staff (possibly by USG or one of its
successor organizations, not necessarily Bell Labs CSRC staff).  These
days, that can be a criminal act.[2]  That's how BSDI got USL to
basically go away (with much resentment--nothing embitters a free market
capitalist more than failure to dispose of a rival through means other
than open competition at a profitable price point), and why 4.4BSD
happened at all.

Given that miserable record, I am reluctant to participate in
perpetuating the false notion that this history is anything but
ambiguous.

Therefore until I hear from someone who has a great command of the
factual and legal history--there do exist authorities I would trust--I'd
prefer to just leave the matter unaddressed.  If harassed by GitHub, I'm
not sure what I'll do.  I might take the repo down, or I might put up a
fight and try to make somebody's attorneys earn their fees.  (For
several years I worked professionally in software licensing compliance
for products including FLOSS[3] at Cisco.  Might be fun to run into
someone I know.)

I'll respond to the technical points of your message in a separate
thread.  :)

Regards,
Branden

[1] 
https://github.com/dspinellis/unix-history-repo/blob/Research-PDP7-Snapshot-Development/Caldera-license.pdf
[2] 17 U.S.C. 506(d) "Fraudulent Removal of Copyright Notice."
    https://www.copyright.gov/title17/92chap5.html
[3] Free/Libre/Open Source Software

signature.asc
Description: PGP signature