[TUHS] Were any of the commercial unixes source-available?

[TUHS] Were any of the commercial unixes source-available?

[Joseph Holsten]

I’ve been playing around trying to link the OpenSolaris launch commit to various pieces in the Unix History Repository, and it’s making me wonder if we’ll ever have a chance to see the history of the systems.

I’m less concerned about HPUX, AIX and SCO’s offerings since I presume someone has copy inside these companies. But what about A/UX, Irix, Tru64? Did these ever get sold with licenses to source tapes? Are there copies we need to preserve in-camera so something can exist 120 years after creation or whenever copyright expires?

--
Joseph Holsten
http://josephholsten.com
mailto:joseph@josephholsten.com
tel:+1-360-927-7234

[TUHS] Re: Were any of the commercial unixes source-available?

[Ron Natalie]

Note that most UNIX was protected by Trade Secret rather than copyright 
so there’s no statutory expiration date.

We were contractors for IBM so we had XENIX and AIX sources in our 
facility under NDAs, so I don’t believe IBM dealt with it other than 
OCO.
I did have a license for the source code for Interactive Systems IS/1 
but that was a fairly straight forward System V kernel.

[TUHS] Re: Were any of the commercial unixes source-available?

[Warner Losh]

[-- Attachment #1: Type: text/plain, Size: 1242 bytes --]

On Sun, Sep 3, 2023, 4:42 AM Ron Natalie <ron@ronnatalie.com> wrote:

> Note that most UNIX was protected by Trade Secret rather than copyright
> so there’s no statutory expiration date.
>

But only so long as they remain secret... there are a few leaks of early
stuff in bitsavers... the leader broke the law, perhaps, but once the
secret is out there, its no longer secret. See the 32V preliminary rulings
for variations on that theme. System V and newer did have copyright
protection as well as trade secret contract stuff...

It isn't hard to find at least one copy of most of the major and many minor
players. It's fair use to copy for study and commentary. But building a
full system for commercial benefit is not.

We were contractors for IBM so we had XENIX and AIX sources in our
> facility under NDAs, so I don’t believe IBM dealt with it other than
> OCO.
> I did have a license for the source code for Interactive Systems IS/1
> but that was a fairly straight forward System V kernel.
>

Several variations on that are out there...

What isn't out there are the full source code control trees. At least Sun
distributed them. But so far only a symlink farm for that has been found.

Warner

>
>

[-- Attachment #2: Type: text/html, Size: 2116 bytes --]

[TUHS] Re: Were any of the commercial unixes source-available?

[Andrew Warkentin]

On 9/2/23, Joseph Holsten <joseph@josephholsten.com> wrote:
> I’ve been playing around trying to link the OpenSolaris launch commit to
> various pieces in the Unix History Repository, and it’s making me wonder if
> we’ll ever have a chance to see the history of the systems.
>
> I’m less concerned about HPUX, AIX and SCO’s offerings since I presume
> someone has copy inside these companies. But what about A/UX, Irix, Tru64?
> Did these ever get sold with licenses to source tapes? Are there copies we
> need to preserve in-camera so something can exist 120 years after creation
> or whenever copyright expires?
>
There are source leaks for quite a few commercial Unices floating
around in various places. These are the ones I'm aware of:

A/UX (0.7, which is complete, and 2.x, which is only the kernel)
AIX 4.1.3 (most of the kernel and some of user space, possibly
complete enough to build)
BSD/OS (various versions, probably complete)
DEC OSF/1 (1.0 and 2.0; these seem reasonably complete)
DYNIX 3.x (several versions, possibly complete enough to build)
DYNIX/PTX (4.x?; possibly complete enough to build)
IRIX 6.5.5 (missing quite a few major packages and nowhere near
complete enough to build)
MIPS RISC/os 4.52 (possibly complete)
SGI System V GL2-W3.7 (for the IRIS 3000 68K machines; probably complete)
SunOS 4.1.3 (seems to be the complete base system)
System V for the 3b2 (several 3.x versions, possibly complete)
System V for the UNIX PC (3.51, possibly complete)
System V/386 4.2 (possibly complete)
ULTRIX-11 (at least 3.1)
ULTRIX-32 (2.0, which has been confirmed to build by someone else, and
4.2, which is also fairly complete)

I haven't looked at any of these in depth, so I'm not completely sure
of the status of any of them except for A/UX and ULTRIX

[TUHS] Re: Were any of the commercial unixes source-available?

[Marc Donner]

[-- Attachment #1: Type: text/plain, Size: 2131 bytes --]

My team at Morgan Stanley had a source license to SunOS in the early ‘90s.
We tried to secure a license to AIX source but never succeeded.

On Sun, Sep 3, 2023 at 7:26 AM Andrew Warkentin <andreww591@gmail.com>
wrote:

> On 9/2/23, Joseph Holsten <joseph@josephholsten.com> wrote:
> > I’ve been playing around trying to link the OpenSolaris launch commit to
> > various pieces in the Unix History Repository, and it’s making me wonder
> if
> > we’ll ever have a chance to see the history of the systems.
> >
> > I’m less concerned about HPUX, AIX and SCO’s offerings since I presume
> > someone has copy inside these companies. But what about A/UX, Irix,
> Tru64?
> > Did these ever get sold with licenses to source tapes? Are there copies
> we
> > need to preserve in-camera so something can exist 120 years after
> creation
> > or whenever copyright expires?
> >
> There are source leaks for quite a few commercial Unices floating
> around in various places. These are the ones I'm aware of:
>
> A/UX (0.7, which is complete, and 2.x, which is only the kernel)
> AIX 4.1.3 (most of the kernel and some of user space, possibly
> complete enough to build)
> BSD/OS (various versions, probably complete)
> DEC OSF/1 (1.0 and 2.0; these seem reasonably complete)
> DYNIX 3.x (several versions, possibly complete enough to build)
> DYNIX/PTX (4.x?; possibly complete enough to build)
> IRIX 6.5.5 (missing quite a few major packages and nowhere near
> complete enough to build)
> MIPS RISC/os 4.52 (possibly complete)
> SGI System V GL2-W3.7 (for the IRIS 3000 68K machines; probably complete)
> SunOS 4.1.3 (seems to be the complete base system)
> System V for the 3b2 (several 3.x versions, possibly complete)
> System V for the UNIX PC (3.51, possibly complete)
> System V/386 4.2 (possibly complete)
> ULTRIX-11 (at least 3.1)
> ULTRIX-32 (2.0, which has been confirmed to build by someone else, and
> 4.2, which is also fairly complete)
>
> I haven't looked at any of these in depth, so I'm not completely sure
> of the status of any of them except for A/UX and ULTRIX
>

[-- Attachment #2: Type: text/html, Size: 2628 bytes --]

[TUHS] Re: Were any of the commercial unixes source-available?

[Andy Wallis]

On 2023-09-03 08:40, Marc Donner wrote:
> My team at Morgan Stanley had a source license to SunOS in the early
> ‘90s.  We tried to secure a license to AIX source but never
> succeeded.

We had access to the AIX source at my work place. The AIX 3.2.5 source 
delivery included a build machine because we had requirements to be able 
to support and redistribute AIX 3.2.5 for up to 20 years. As I recall, 
it required a lot of high level negotiations between us and IBM to get 
that.

Access was limited to select people on an air-gapped machine. The source 
code itself was locked away on 8mm tape and required signatures to check 
it out.

IBM did have a public announcement letter for AIX 3.2.5 source access.
http://bio.gsi.de/DOCS/AIX/ENUSZP93-0158.printable.html

We also had access to the various AIX versions that we had to support 
our government customers. I know we had access to AIX 4.3.3, 5.1,5,2, 
and 5.3 source code. One of our developers would fix bugs that he found 
and file PMRs to tell the AIX developers what exactly to fix. A strip 
command change to support very large executables took 18 months for them 
to fix because they didn't understand why failing to strip very large 
binaries was a problem.
-Andy Wallis

[TUHS] Re: Were any of the commercial unixes source-available?

[Warner Losh]

[-- Attachment #1: Type: text/plain, Size: 2860 bytes --]

On Sun, Sep 3, 2023 at 5:26 AM Andrew Warkentin <andreww591@gmail.com>
wrote:

> On 9/2/23, Joseph Holsten <joseph@josephholsten.com> wrote:
> > I’ve been playing around trying to link the OpenSolaris launch commit to
> > various pieces in the Unix History Repository, and it’s making me wonder
> if
> > we’ll ever have a chance to see the history of the systems.
> >
> > I’m less concerned about HPUX, AIX and SCO’s offerings since I presume
> > someone has copy inside these companies. But what about A/UX, Irix,
> Tru64?
> > Did these ever get sold with licenses to source tapes? Are there copies
> we
> > need to preserve in-camera so something can exist 120 years after
> creation
> > or whenever copyright expires?
> >
> There are source leaks for quite a few commercial Unices floating
> around in various places. These are the ones I'm aware of:
>
> A/UX (0.7, which is complete, and 2.x, which is only the kernel)
> AIX 4.1.3 (most of the kernel and some of user space, possibly
> complete enough to build)
> BSD/OS (various versions, probably complete)
> DEC OSF/1 (1.0 and 2.0; these seem reasonably complete)
> DYNIX 3.x (several versions, possibly complete enough to build)
> DYNIX/PTX (4.x?; possibly complete enough to build)
> IRIX 6.5.5 (missing quite a few major packages and nowhere near
> complete enough to build)
> MIPS RISC/os 4.52 (possibly complete)
> SGI System V GL2-W3.7 (for the IRIS 3000 68K machines; probably complete)
> SunOS 4.1.3 (seems to be the complete base system)
>

There is (or was) a github repo that has Solaris 2.5, 2.6, 8 and 11 sources.

System III sources are out there, lots of copies, but they all appear
to come from the same root source and have just been repackaged.


> System V for the 3b2 (several 3.x versions, possibly complete)
> System V for the UNIX PC (3.51, possibly complete)
> System V/386 4.2 (possibly complete)
>

Bitsavers also has sources for the IS/1 or similar to a number of different
68k machines, but it's only the kernel. It's V7 based, but with bits of
system III
and system V tossed in.

Also, it's not hard to find SystemVr1 (for VAX and PDP-11), r2, r3 (several)
and r4 online.


> ULTRIX-11 (at least 3.1)
>

This is buildable, and legit with permission from DEC. At least parts of it
are buildable, I've not tried to do a full system gen from sources.


> ULTRIX-32 (2.0, which has been confirmed to build by someone else, and
> 4.2, which is also fairly complete)
>
> I haven't looked at any of these in depth, so I'm not completely sure
> of the status of any of them except for A/UX and ULTRIX
>

hpux has a lot of docs out there. And a lot of binaries, but little source.
It's
one of the few I've not found sources for when I was doing research on how
sync behaved, for example.

Warner

[-- Attachment #2: Type: text/html, Size: 4004 bytes --]