On Wed, Mar 15, 2023 at 5:38 PM KenUnix <ken.unix.guy@gmail.com> wrote:
So, at this point what is the safest road to take?

Stick with v7?
I'm not a lawyer - this is not legal advice. This is how I personally analyze these Ancient UNIX license wording against the history of how different UNIX releases we made publically available. YMMV - get a legal opinion and form your own opinion and make a personal choice.

The Ancient UNIX license -- the document Warren has on the site (https://www.tuhs.org/ancient.html) says [please go read it yourself]:

1.9 SUCCESSOR OPERATING SYSTEM means a SCO software offering that is (i) specifically designed for a 16-Bit computer, or (ii) the 32V version, and (iii) specifically excludes UNIX System V and successor operating systems. 

My take ...
  • Any UNIX package based on the Research Editions 1-7 and 32V is allowed. This family includes the16-bit 1BSD, 2BSD, 2.9-11BSD and 32V based 3BSD, 4BSD, 4.1BSD, 4.2BSD, 4.3BSD and 4.4BSD.
  • PWB1.0 and 2.0 are 16-bit - although PWB 2.0 was not officially released outside of the Bell System (except possibly for the note suggested about Wollongong advertising they had it).  But under the 16-bit rule - I interpret both PWB 1.0 and 2.0 as being covered.
  • PWB 3.0 (a.k.a. System III) was released for both PDP-11 and Vaxen and was thus generally available to the Unix (source) licensees.   Under the 16-bit rule, I would personally interpret that PWB 3.0 is covered since it is not explicitly called out (as System V is called out).   
  • This also puts PWB 4.0 in an interesting place. Like PWB 2.0, it was never released outside of the Bell System; although Bell folks had 16-bit versions, they were starting to be depreciated in favor of Vaxen and WE32000/3B-based ISAs. Given the exclusion starts at System V and there was a 16-bit version for PWB 4.0 internally, again, I personally suspect it's okay, but get your own legal opinion, please.
  • Clearly, anything no matter the ISA any release is based on System V, SVR1, SVR2, SVR3, SVR4, and SVR5 has been excluded in that license, which means unless the current IP owners of System V-based UNIX make a new license, I personally interpret that as a no-no according to this license.

Some other random thoughts..

  • Some of the commercial UNIXs (as described by Charlie WRT to Dell), have encumberments beyond AT&Ts - say IP from MIPs whose compiler was often used and was not based on the AT&T IP and Transcript or PostScript, which came from Adobe.   For instance, besides Dell, DEC, HP, IBM's versions have these types of IP issues in Ultrix/HPUX/AIX.  I suspect many if not most commercial UNIX released would be in the same situation - particularly given Charlie example of Dell who was making a 'Wintel' release.
  • IBM, HP, and Sun all bought out their UNIX license from AT&T at some point and owned the right to do whatever they wanted with it. And as has been discussed here, a version of Solaris which had SVR4 code in it was released by Sun and later taken back in by Oracle.   Some questions for your lawyers would be:
    1. If Solaris was released, doesn't that make at least the bits from that release available forever?      Clearly, some people on this list have made that interpretation - I'm personally not willing to take that risk.
    2. Assume 1 seems to mean that the Solaris IP from that release is free to be examined and used since it was partly based on SVR4, does that make SVR4 available also? i.e. it does not matter what the owners of the SVR4 IP think, Sun legally released it with their license?     To me, this gets back to the USL vs. BSDi/UCB case ok what was what, and the question is how to show some portion of the code base was or want not released by Sun and what parts had been.   Again, I personally will not take that risk.