|
The SCO Unix timeline - Members Only |
|
Saturday, November 29 2003 @ 10:49 AM EST
|
Here is some more on our project to try to come up with a better UNIX timeline. Charles Puffer has actually made a start on a graphic. If you take a look at the SCO's page on Linux issues, the chart seems to show only two links between SCO's code and Linux - one from BSD 4.4, and another from Linux 0.01 which I believe was written by Linus himself. 2. On the bottom it states that "Original UNIX history chart created by Eric Levenez. Copyright © 1996-2003, Eric Levenez. January 2, 2003. Used with permission.". Eric's chart can be found here. They are not identical, however, despite the copyright notation which would imply that they should be. Has anyone asked Eric if he gave permission to change the chart? Does anyone know him? I wrote to him, but he never answered.
Note the excellent input in the comments section of the earlier members only discussion on this subject.
[ Update: Oct. 19, 2009 - Because SCO has now removed their version of the UNIX timeline, it seems appropriate to place it here for historians. So I am opening the article to the public.]
Update 2: In typical SCO fashion, SCO failed to recall that they'd also put this chart on Darl's site darlmcbride.com, so it's still there in its faux glory, just the way they posted it, with clickable features.
*********************************************
SCO's Unix Timeline
By Charles Puffer Copyright 2003
When I became aware of SCO's SCO Intellectual Property Pedigree I wondered if it would change as the court case continued. In correspondence, Pamela Jones noted that the SCO Intellectual Property Pedigree was based on Éric Lévénez's history and that it had already been changed. So I set out to document the changes as well as the present SCO version of the history. These are my comments based on review of both histories and a graphical comparison. You can view the histories yourself, Éric Lévénez's history is located here and SCO's history is here.
In the 1969 through 1982 sections, (and throughout the SCO version of the history), the only changes of note are the addition of colored lines to emphasize the SCO's view of Unix history and the change of the font/size of some of the labels mostly for emphasis and to move them away from the colored lines. The choice of colors does not seem to have meaning: Green for Linux referred to as (SCO Linux) Pedigree; Blue for the OpenServer Pedigree; Dotted lines of the same color are used for the History of Linux and Open Server. A Separate yellow line is used to emphasize the UnixWare Pedigree. The use of SCO Linux in this context suggests a level of proprietary on SCO's part.
In the 1983 section, the first interesting thing happens as the Linux pedigree is shown to be following the path of Sinix, when Sinix forks with one fork staying "pure" and the other combining with material from Xenix 3.0. SCO's green dotted line would like us to see GNU/Linux as derived from this Xenix-Sinix hybrid.
In the 1984 section, we see SCO's green dotted line follow down from the Xenix-Sinix hybrid to Minux. But wait--that's not what Levenez's chart shows. Levenez's history shows Minix coming from the pure Sinix line. This may be an attempt to more strongly connect Linux with Xenix, which SCO might like us to see as part of the OpenServer pedigree, or, it may be a mistake.
Between 1985 and 1991, we do not see any changes of note other then the coloration and font/size changes. There are not even any connections between any of the Unix and Minix (the sole carrier of the Linux pedigree) or vise-versa.
In the 1991 section, Linux comes into the picture. At this time we see more small differences in the histories, though I suspect many of these relate to different versions of Levenez's history and not to any actions on SCO's part.
In the 1994 section, there is a code addition from BSD Lite. This would and should be legitimate as long as the BSD license is observed.
In the 1996 section, MKLinux is created as a union of Mach4 and Linux.
In the 1998 section, IBM's Project Monterey begins.
In the 2000 section, we see connections from Linux (several types of Linux) to OpenServer. This creates a strong suggestion that there is more Linux in SCO than SCO in Linux. We also see arrows from Linux to Monterey followed by the Monterey dissolving into AIX.
In the 2002 section, we see three arrows from AIX to Linux. These arrows only exist in Levenez's history and may be a result of his continued updating while the SCO version has not been updated. We also know that IBM did make contributions to Linux and it may be the legitimacy of these contributions that SCO wishes to litigate.
I would not say there is a smoking gun or a clear view of SCO thought process in SCO's version of the history. Changes in the SCO version, (including the removal of Lévénez's copyright notice) seem to indicate a level of proprietary attitude. They also suggest a willingness to reach for small threads to hold up big weights and, if necessary, to find these threads where they do not exist. To the intent and ethical grounding of SCO's SCO Intellectual Property Pedigree, I think that is best left as an exercise to the reader.
|
|
Authored by: eamacnaghten on Saturday, November 29 2003 @ 11:25 AM EST |
If you go to Lévénez's web site, it says t the bottom regarding the chart...
Note 1 : an arrow indicates an inheritance like a compatibility, it is not only
a matter of source code.
I think that puts the kaibosh on any possibility of this being used in evidence
for anything....[ Reply to This | # ]
|
|
Authored by: JanC on Saturday, November 29 2003 @ 12:02 PM EST |
In the 1984 section, we see SCO's green dotted line follow down
from the Xenix-Sinix hybrid to Minux. But wait--that's not what Levenez's chart
shows. Levenez's history shows Minix coming from the pure Sinix line. This may
be an attempt to more strongly connect Linux with Xenix, which SCO might like us
to see as part of the OpenServer petigree, or, it may be a
mistake.
IMHO the chart shows Minix being "derived" from the
old Unix TSS. (Look at where the arrow starts...)
[ Reply to This | # ]
|
- Correct - Authored by: Dan M on Saturday, November 29 2003 @ 12:33 PM EST
- Correct - Authored by: Waterman on Saturday, November 29 2003 @ 01:28 PM EST
- Correct - Authored by: PJ on Saturday, November 29 2003 @ 02:26 PM EST
- Correct - Authored by: D. on Sunday, November 30 2003 @ 11:57 PM EST
|
Authored by: eamacnaghten on Saturday, November 29 2003 @ 12:43 PM EST |
I find it fascinating that SCO are claiming "SCO-Linux Pedigree"
from Linux 0.01 that was written (released?) in Aug 01 1991.
Someone ought to tell Linus that he was working for SCO then, and any memories
he has that he was a student in Helsinki that worked on this project in his
spare time was just some kind of illusion or hallucination....[ Reply to This | # ]
|
|
Authored by: Thomas Frayne on Saturday, November 29 2003 @ 12:54 PM EST |
This chart is just a start. To take the next step, each arrow needs a paragraph
or so to explain the relationship of the arrow's ends, the major features
introduced at each point, who owned which features, and how the products were
licensed.
Examples of relationships:
a. Superset of source with minor modifications;
b. Superset of source with major modifications;
c. Similar technology developed independently;
d. Concepts used, not code;
e. Part of one product ported to the other;
The chart by itself displays too many relationships to keep track of. If the
explanations are added, we can zero in on the important nodes more easily, and
follow important paths through the history.
[ Reply to This | # ]
|
|
Authored by: rsmith on Saturday, November 29 2003 @ 01:19 PM EST |
The Levenez chart indeed states that an arrow between X and Y does not have
to stand for "Y's code derives from X". But it it's current form it is easily
misused by entities SCO.
It would be good to make a more visual
difference between UNIX versions actually borrowing code from each other and
versions that are compatible or borrow ideas (but not source) from each
other.
E.g. like the Berkely Packet Filter idea was implemented in Linux
from a published specification, not from BSD code.
--- Confucius say:
It is impossible to sling mud with clean hands. [ Reply to This | # ]
|
|
Authored by: rand on Saturday, November 29 2003 @ 01:19 PM EST |
When I readsomething like this it makes me think
that not all the important links are represented:
- SVR4 (1988),
mainstream of Unix implementations, merge of System V, BSD, and
SunOS.
- From SVR3: sysadmin, terminal I/F, printer (from
BSD?),
RFS, STREAMS, uucp
- From BSD: FFS, TCP/IP,
sockets, select(), csh
- From SunOS: NFS, OpenLook GUI, X11/NeWS,
virtual memory
subsystem with memory-mapped files, shared
libraries
(!= SVR3 ones?)
--- Dim
gstrIANAL As String
(Oh, Lord, get me off this project...) [ Reply to This | # ]
|
|
Authored by: cdru on Saturday, November 29 2003 @ 02:12 PM EST |
If you want the images arranged they way they are suppose to, look here. They
are suppose to be arranged horizontally into one long image, not vertially into
one tall image. [ Reply to This | # ]
|
|
Authored by: gleef on Saturday, November 29 2003 @ 03:33 PM EST |
I'm unclear on the distinction in this chart (and SCO's) between
"Pedigree" and "Heritage". Is the claim of
"Pedigree" that each work is based on code from those before, and
"Heritage" that each work gets some inspiration (but no code) from
those before? Or is it some attempt at sorting out whether or not A is legally
a derivative work of B (you can share code without being a derivative work, eg.
JFS was written for OS/2, and shared with AIX and Linux, this doesn't make
Linux JFS derivative of AIX JFS or vice versa).
The solid "Pedigree" connection between Minix and Linux is
particlarly troubling to me. To my knowledge, Linux 0.01 used no Minux code,
and is not a derivative work of Minux. My understanding is that Minix provided
the inspiration and the file system specification (and possibly the boot
loader?) of the early Linux versions. The line doesn't even make sense, to
have a Pedigree of that version of Minix, but not the earlier versions of Minix
is nonsensical to me.[ Reply to This | # ]
|
|
Authored by: gleef on Saturday, November 29 2003 @ 03:57 PM EST |
In the "1999" section, on the Linux 2.3 line, There's a big label on this
chart of "2.2.14 August 25, 1999". This label is incorrect on many
counts.
- It is on the wrong line, the 2.2 Linux kernel is two lines
down.
- According to kernel.org, the 2.2.14
release of the Linux kernel is dated January 4, 2000, over four months
later.
- If you assume that it's a typo, and they meant 2.3.14, then according
to kernel.org, the
2.3.14 release was dated on August 18, 1999, more closely matching the small
"2.3.14 August 19, 1999" entry that this larger tag is written over (some
kernels were released the day after they were packaged, kernel.org shows the day
it was packaged).
Does anyone know what they're trying to say
here? It might be useful to do a map of Caldera/SCO kernel releases, for what
it's worth, I've got Caldera OpenLinux 2.2, and can break out the CDs and see
what kernel it shipped with, and what date was on it (the docs only say that it
has a "Linux 2.2 kernel"). [ Reply to This | # ]
|
|
Authored by: rakaz on Saturday, November 29 2003 @ 04:53 PM EST |
According to the timeline Minix is linked to UNIX TSS and Linux is linked to
Minix. This suggests a common heritage, but I do not think there is one, apart
from the obvious that Minix was created to work and look like Unix and that
Linux was created to work and look like Minix and Unix. As far as I know there
were no exchanges in source code or ‘methods’ between these three operating
systems when they were originally created.
Minix was created at a time that BSD still required an AT&T Unix license. So
there were no free Unix-like systems available. Andy Tanenbaum, teaching at the
Vrije Universiteit in the Netherlands, needed a Unix-like system for educational
purposes. So he created one from scratch, one that didn’t require a license from
AT&T. Although Minix was inspired by Unix it was quite a difference animal,
Unix has a monolithic kernel and Minix has a microkernel.
Minix was not free. The license for Minix was very liberal, but you still needed
to pay a small licensing fee. In the late 1990s Minix became open source.
Linux was born out of frustration. Linus Torvalds bought a PC and wanted to run
a Unix-like system, like he was used to working on at the University of
Helsinki. First he tried Minix, but he did not like it. So he build his own. He
did have access to the source code of Minix but didn’t use any of it in Linux.
The design of Linux is actually very different from Minix. Like I said before,
Minix uses a microkernel and Linux uses a monolithic kernel. In fact there is a
nice discussion archived about his between Andy Tanenbaum and Linus Torvalds:
http://www2.educ.umu.se/~bjorn/mhonarc-files/obsolete/
So, it is true, Minix is linked to Unix TSS and Linux is linked to Minix, but
certainly not in the way the SCO version of the timeline suggests. Minix is not
a descendent of Unix, but simply a look-alike and so is Linux.[ Reply to This | # ]
|
|
Authored by: PJP on Saturday, November 29 2003 @ 06:20 PM EST |
First comment on the diagram: getting this completely right would be a big
undertaking, but do we really need to? I would suggest that the first thing to
do is to prune all branches which have no relationship at any time to Linux.
This would make the relationships much clearer. Unless fo course, the intent is
simply to make clear just how convoluted and cmplex the history of Unix really
is.
So, decide on the purpose, then work on only the parts of the diagram
needed to fulfill that purpose.
Now on to MINIX.
MINIX really doesn't
derive from Unix V7 in any way other than as described by its author, Andrew
Tannenbaum as follows:
... I have written a new operating system, MINIX
from scratch. MINIX has the same system calls as version 7 UNIX (except for the
omission of a small number of unimportant ones). I have also supplied a shell
that is functionally identical to the UNIX shell along with more than 60 other
programs that are similar to their UNIX counterparts (e.g. cat, cc cp, grep ls
and make). In short, to the user, MINIX looks very much like
UNIX.
The licensing has no dependencies on AT&T or BSD. The
notice with the source reads as follows:
Copyright (C) 1987 by
Prentice-Hall, Inc. Permission is hereby granted to private individuals and
educational institutions to modify and redistribute the binary and source
versions of this system to other private individuals and educational
institutions for educational and research purposes. For corporate or commercial
use, permission from Prentice-Hall is required. In general, such permission will
be granted subject to a few conditions.
It looks (to me) as though
the inclusion of MINIX in these discussions is a bit of a red herring. It is not
derrived from unix other than in spirit and the adoption of the Unix system call
API.
For the purpose of tracing the relationship of Linix to SCO owned code,
I believe MINIX is irrelevant (Unless Darl is going to claim that he owns MINIX
to, since it used the Unix V7 sys. call API, and provides similarly functioning
utilities). [ Reply to This | # ]
|
|
Authored by: Khym Chanur on Saturday, November 29 2003 @ 07:24 PM EST |
Good God! If the UNIX family of operating systems was a family of humans,
they'd be so inbred that they'd have negative IQs!! [ Reply to This | # ]
|
|
Authored by: Sunny Penguin on Saturday, November 29 2003 @ 10:25 PM EST |
Caldera derived Unix from Linux profits. Without profits from Linux sales SCO
could not have purchased the IP they are so upset about.
---
Norman Madden
Chuluota, FL[ Reply to This | # ]
|
|
Authored by: D. on Saturday, November 29 2003 @ 10:31 PM EST |
These charts have some information that can be quite misleading.
For example the "inheretance" v7 -> MINIX -> Linux. (see
comments in another thread.)
The various AT&T distibutions are not clearly delimianted. For example the
research verisions (v1, v2, v3 ...) were seen by Thompson & Ritchie, et al
as a continum, with version numbers changing when it came time to update the
manual. In-house production versions like the programmers work bench (PWB) and
the documenters work bench (DWB) were separate projects that had some cross
fertilazation with Murray Hill.
Then there are the commercial unices. SysIII and SysV (there was no SysI or II).
The charts show a SystemIV. While
there may have been plans for one, no SysIV was released to the Public.
The charts also show a 1BSD, there ain't no such beast. The first tape that
Bill Joy made was called Berkeley Software Distribution, and only contained the
Berkely Pascal Compiler and the text editor "ex" and maybe the vi
editor.
The first full OS released by UCB was either on the second or third tape.
D.
[ Reply to This | # ]
|
|
Authored by: jmc on Sunday, November 30 2003 @ 09:31 AM EST |
It seems to me that a whole lot of extra FUD is created by treating Unix or
Linux kernels as a great amorphous blob both in this timeline and elsewhere.
Think of scheduling and process management, syscall handling, device drivers for
all sorts, networking, configuration, boot and also support for the developing
chipsets with ISA, EISA, PnP, PCI and so forth.
Not only do you end up with "manageable pieces" to compare and talk
about the history of but you don't get the feeling that a bit of scheduling
code isn't being compared with a network card driver and SCO are saying that
lines are copied because they have similar trivial things on them like
"return 0;" or close brace.
But I forget, that wouldn't help SCO. It would reduce by an order of magnitude
the stuff they had to argue about.[ Reply to This | # ]
|
|
Authored by: markus on Monday, December 01 2003 @ 10:29 AM EST |
There is a very nice explanation of the early Unix history on the Bell Labs
website:
http://www.bell-labs.com/history/un
ix
--- Markus Baertschi, Switzerland [ Reply to This | # ]
|
|
Authored by: pjcm on Saturday, January 17 2004 @ 11:01 AM EST |
I have tried to produce an accurate timeline for a large number of OSes that do
not come under the existing Unix timeline.
The current early stage diagram is available at :-
http://firedrake.org/paddy
I would be very grateful if you could have a look at it and comment on any
omissions or point me towards any missing information. I am especially
interested in the early years, prior to 1975, as these are currently vague.
Thank you for your time and patience.
Paddy[ Reply to This | # ]
|
|
|
|
|