The big three -are- all based on 4.4BSD-Lite. OpenBSD forked off NetBSD around 1995. FreeBSD and NetBSD were originally (circa 1993) patched versions of 386BSD, which was based on the Net/2 release, itself intended as an unencumbered version of 4.3BSD. FreeBSD and NetBSD had to re-base their source trees to 4.4BSD-Lite when - as a result of the settlement of the lawsuit in early 1994 - Net/2 was considered tainted. As I recall FreeBSD 2.0 was the first version based on 4.4BSD. Of course the source code of the different projects has changed (and diverted) a lot since then.

As far as I know, BSD was not derived from System V but from V7 (via 32V). Early System V releases were based on V7 combined with a few internal Bell Labs systems.

You can get the 4.4BSD-Lite releases as tar.gz file from all over the Internet (use Google to find one) or you could order the BSD archive CD-ROM set from Kirk McKusick (www.mckusick.com). I have not seen a tape distribution offered recently.

For Sony NEWS, a fairly short lived product. The workstations were initially based on Motorola 68K, later on MIPS R3000 and up. The OS was 4.3 BSD in the beginning, later also SVR4. The X ports were quite straightforward, mostly adaption to hardware and fighting the compiler (slightly quick-and-dirty PCC based). And of course the bugs we found in X itself.

It is a while ago (more than 20 years!), but I am certain that we fixed most of the issues described in CVE-2014-8092 and CVE-2014-8095 when porting X11R4 and we did report those bugs to the X Consortium (actually a precursor of Xorg.)

I am also quite sure that the large companies (Sun, SGI maybe IBM) also fixed and possibly reported those bugs. I can only conclude that the software engineering practices of the X developers were a bit sub-standard, even for those days and it also makes me quite suspicious about the way-of-working of the present OpenDesktop and Xorg developers.

I have a strong impression that they are more focused on development of Wayland anyhow. And that is not a good thing...

I'm still "using" Minix (currently 1.6.25) both on my ancient Atari 1040 ST and on an Atari ST simulator. Part out of nostalgia of course, but also to remind myself what you could do using a CPU that is about 10000 times slower than current CPU's, running an OS that you could actually understand completely by reading the complete source code. And I am using a cross compiler these days, based on GCC 4.x running on FreeBSD (see www.beastielabs.net/prerel.html) and it is amazing to see what you still can get to run on the 1993 vintage Minix, although there is still no networking.

I remember having read the first edition of OSDI (including the source code) from cover to cover several times in the late 80's (the book split in two volumes as a consequence 8-) ) and a really learned a lot from it, although I was already heavily involved in Unix kernel hacking by then (both BSD and SVR4 based kernel ports). I still value Andy's OSDI in its several editions as a most educational experience.

Actually Sun co-created SVR4 8-). Quite early, I think in 1987 Sun and AT&T cut a deal (As I remember Sun stock was also involved) to jointly develop and market a unified version of Unix, containing features from SunOS, SVR3.2, 4.3BSD and Xenix (that originally was based on V7 and marketed by Microsoft).

Such a unified version was not unwelcome after a decade of incompatible Unix versions from many different vendors. When SVR4 was released in 1988 it also promised compliance with the then new IEEE Posix standard. And additionally, the Berkeley CSRG had made clear on several occasions that they would not and could not offer commercial-grade support on their BSD system: they were an academic research group after all. And since the BSD source code was not yet publicly available, these were quite convincing arguments for both the Unix system vendors and their customers to move away from BSD toward SVR4.

Of course the Sun-AT&T deal by itself caused quite a few new schisms. And Sun seriously botched the first releases of there SVR4 version (especially Solaris 2.2 and 2.3), which made Sun customers stick with SunOS 4 for a very long time (some even "downgraded" from Solaris to SunOS in response to continuing issues). The last release of SunOS (4.1.4) was in 1994, and I guess nobody expected it to last that long (especially not Sun).

I don't think hardware was much a driver for Solaris. SunOS was also available for the Sparc and the M68K was quickly forgotten after the introduction of the Sparc. And though there actually was an MP version of SVR4 available at a certain point in time, it really smelled badly 8-) The reputation of Solaris for SMP stems mostly from its later releases (from Solaris 6 onwards) that considerably diverged from pure SVR4.

So there probably were not many hardcore technical reasons to move to Solaris. But there was a class of users for whom the technical problems with Solaris (e.g. with its weird socket interface) motivated a move away from it during the first years of the Internet boom toward Linux and (of course) FreeBSD (both at least SunOS-like) and to Windows NT.