[TUHS] Tree of late BSD releases

[TUHS] Tree of late BSD releases

[Diomidis Spinellis]

For reconstructing Unix history on a single repository [1], I'd need to 
represent the branches, merges, and chronological sequence of the late 
BSD releases (after 4.3).  However, I've found on the internet some 
conflicting and simplistic information, so I'd welcome your input on how 
to straighten things up.

First, consider this widely reproduced BSD family tree [2].  It has 
4.4BSD-Encumbered derive from a line that includes Net/1, which was 
freely redistributable.  Wouldn't it be clearer to create two branches, 
one with distributions free of AT&T code (4.3 BSD Net/1, 4.3 BSD Net/2, 
4.4 BSD Lite1, 4.4 BSD Lite2) and one with full distributions (4.4 BSD, 
...)?  On which side would Tahoe and Reno stand?

Also, the same tree [2] shows 4.4 BSD having as its ancestor 4.3 BSD 
Net/2, whereas another tree depicted on Wikipedia [3] has shows 4.4 BSD 
and 4.3 BSD Net/2 having as their ancestor 4.3 BSD Reno.  What's the 
correct genealogy?

Finally, I have a conflict with release dates.  Wikipedia gives the 
following dates for Tahoe and Net/1 [4]:

4.3 BSD Tahoe June 1988
4.3 BSD Net/1 June 1989

However, looking at time-stamp of the newest files available under the 
corresponding directories in the CSRG CD-ROMs [5] I find the opposite order:

cd2/net.1/sendmail/src/util.c 1989-01-01 12:15:58
cd2/4.3tahoe/usr/src/sys/tahoevba/vx.c 1989-05-23 13:47:43

What's the actual time sequence, and what's the corresponding genealogy?

[1] https://github.com/dspinellis/unix-history-repo
[2] 
http://ftp.netbsd.org/pub/NetBSD/NetBSD-current/src/share/misc/bsd-family-tree
[3] https://en.wikipedia.org/wiki/File:Unix_history-simple.svg
[4] https://en.wikipedia.org/wiki/Berkeley_Software_Distribution
[5] https://www.mckusick.com/csrg/

Many thanks,

Diomidis Spinellis

PS Thank you all for the help you've provided so far.

[TUHS] Tree of late BSD releases

[Jeremy C. Reed]

On Sun, 15 Jun 2014, Diomidis Spinellis wrote:

> First, consider this widely reproduced BSD family tree [2].  It has 
> 4.4BSD-Encumbered derive from a line that includes Net/1, which was 
> freely redistributable.  Wouldn't it be clearer to create two 
> branches, one with distributions free of AT&T code (4.3 BSD Net/1, 4.3 
> BSD Net/2, 4.4 BSD Lite1, 4.4 BSD Lite2) and one with full 
> distributions (4.4 BSD, ...)?

I don't have any preference on that.

>  On which side would Tahoe and Reno stand?

Tahoe included both the proprietary code and the code which could "be 
freely redistributed". The same with Reno, which was also "within the 
usual licensing constraints" (as it still had the proprietary code).

> Also, the same tree [2] shows 4.4 BSD having as its ancestor 4.3 BSD 
> Net/2, whereas another tree depicted on Wikipedia [3] has shows 4.4 
> BSD and 4.3 BSD Net/2 having as their ancestor 4.3 BSD Reno.  What's 
> the correct genealogy?

Note that Net/2 was not a complete or ready-to-use system, so a better 
genealogy (than those two) may be like:

4.3BSD Reno
  |      \
  |       \
  |      4.3BSD NET/2
  |       /
  |      /
4.4BSD Alpha 

> Finally, I have a conflict with release dates.  Wikipedia gives the 
> following dates for Tahoe and Net/1 [4]:
> 
> 4.3 BSD Tahoe June 1988

Announced in June 1988 (see admin/postings/4BSD/INDEX) but even users 
outside of Berkeley reported bugs specifically against "Tahoe" source  
tree as early as at least May 1987 (see admin/bugs/4.3BSD-tahoe/).

> 4.3 BSD Net/1 June 1989

That date is in McKusick's open sources chapter, but the "Yesterday, 
Today and Tomorrow" article says Spring 1988. It was announced in 
November/December 1988 (see admin/postings/4BSD/V1/73 and the 
/admin/postings/4BSD/INDEX). (930108.oppose.txt also says was 
distributed in 1988.) (Two Salus books also mention that it was 
announced at the BSD Workshop in Nov. 1988 to be available.)

As for dates for the files, they were probably still updated even after 
initially released (which you can see in other releases too).