Slashdot Mirror
SCO Announces Final Termination of IBM's Licence
ickle_matt writes "SCO have announced the final termination of IBM's UNIX license, despite Novell telling them they can't. Interestingly enough there's a new set of "stolen code" figures in the release - 'approximately 148 files of direct Sequent UNIX code to the Linux 2.4 and 2.5 kernels, containing 168,276 lines of code. This Sequent code is critical NUMA and RCU multi-processor code previously lacking in Linux. Sequent-IBM has also contributed significant UNIX-based development methods to Linux in addition to the direct lines of code specified above.' "
To avoid the slowness of SCO's Cold Fusion, use this Yahoo! Finance mirror of the Press release:
http://biz.yahoo.com/prnews/030813/law050_1.html
2.5 isn't a from-scratch rewrite of the kernel so large portions of the kernel source code are common. Apparently the changes were introduced in the 2.4 kernel so the 2.2 kernel isn't a problem. The above isn't implying that I believe that there is any validity to SCO's fraudulent claims.
Chris Kuivenhoven is a thief, beware
Actually, SCO is simply asserting what it considers to be a legal right over the software via licensing. Their legal arguments would be weaker if they did not take steps to protect their IP. Please note I am NOT taking SCO's side here but these events are perfectly reasonable from a legal point of view as strategic posturing.
Courts are just a forum for dispute resolution - don't confuse this with justice.... an entirely different concept.
This is the license for some product called Dynix, quite distinct from IBM's AIX licensing.
Interestingly, this contract is with Sequent, not IBM. They're now alleging that Sequent gave IBM (it's parent) the code, breaching this contract. But seeing as IBM is a different company than Sequent, surely they were not under an obligation to keep that code to themselves; Sequent fudged up by giving it to IBM, so why should IBM's license to AIX be revoked? The Sequent contract read, according to this press release, that derivative could should be treated "as if" it were UNIX code, not that it actually became the property of SCOX.
It looks to me that the code was a trade secret at Sequent ("treaded as if..") but they retained copyright. Then it got divulged to IBM, who even obtained the copyrights, and could disperse of it as they pleased. Liability is then ristricted to Sequent's officers who breached the contract with SCOX, and perhaps some IBM officers if it can be prover they coerced Sequent. IBM's AIX license and copyrights (and thus ability to GPLize) stand, IFF this press release contains the actual facts!
SCO employee? Check out the bounty
There is an interesting article "Novell letters throw new light on SCO-IBM case" published in The Age that discuss letters sent to SCO by Novell asserting that Novell has "the right to compel SCO to waive or revoke any of its (SCO's) rights under the contract [involving SCO, IBM, and Novell]." In the letters Novell basically tells SCO that SCO cannot terminate IBM's license. These letters were apparently used as exhibits in IBM's countersuit of SCO.
-- Signature withheld by popular demand.
They'll be plenty of people with older (licensed) versions of SCO Unix cose against which SCO's copies could be compared. It may be under NDAs, but no court in the world would punish anyone who breaks that NDA to expose any barratry and fraud on behalf of SCO.
Besides, the court could probably subpoena such escrow code as part of the discovery process.
PS : I'm a lover, not a lawyer
Neither of these servers are actually in the sco netblock, however. www.sco.com is being hosted by Noorda Family Trust, at www.nft.com. ir.sco.com is being hosted, ironically enough, by Sequent Computer Systems.
I'm not talking about patents here, but about literal source code.
According the press release 'SCO's System V UNIX contract allowed Sequent to prepare derivative works and modifications of System V software "provided the resulting materials were treated as part of the Original [System V] Software." Restrictions on use of the Original System V Software include the requirement of confidentiality, a prohibition against transfer of ownership, and a restriction against use for the benefit of third parties.'
So IBM's patents aren't transferred or licensed to SCO, but if they write code for a System V derivative work (such as AIX) they can't use the code in Linux. The contract basically says 'All your AIX code are belong to SCO'
This is totally meaningless.
$ dig ir.sco.com
; <<>> DiG 9.2.0 <<>> ir.sco.com
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 64936
;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 2, ADDITIONAL: 0
;; QUESTION SECTION:
;ir.sco.com. IN A
;; ANSWER SECTION:
ir.sco.com. 14393 IN CNAME cald.client.shareholder.com.
cald.client.shareho lder.com. 1793 IN CNAME client.shareholder.com.
client.shareholder. com. 1793 IN A 170.224.5.43
;; AUTHORITY SECTION:
shareholder.com. 14286 IN NS ns1.shareholder.com.
shareholder.com. 14286 IN NS ns2.shareholder.com.
;; Query time: 18 msec
;; SERVER:
;; WHEN: Wed Aug 13 11:08:36 2003
;; MSG SIZE rcvd: 132
This investor relations page is being hosted by someone else.
rooooar
This letter specifically gave back to IBM the rights of any code they created to enhance or extend the AIX/SYS5R4 OS.
This straight away rules out such claims unless SCaldera believe they can convince a Judge and Jury that the superceeding agreement is invalid in someway.
Something I rather doubt :)
Looking though the 2.4.21 sources it looks like RCU isn't even merged into the mainline kernel yet. As for NUMA, it effects all of about 20 files. So can somebody tell me where the hell SCO is finding 148 files of copied code!?! Hell, the core code in ipc, mm, and kernel is only 58 files.
mm/numa.c:
/*
* Written by Kanoj Sarcar, SGI, Aug 1999
*/
arch/x86_64/mm/numa.c:
/*
* Generic VM initialization for x86-64 NUMA setups.
* Copyright 2002 Andi Kleen, SuSE Labs.
* $Id: numa.c,v 1.6 2003/04/03 12:28:08 ak Exp $
*/
/*
* linux/arch/alpha/mm/numa.c
*
* DISCONTIGMEM NUMA alpha support.
*
* Copyright (C) 2001 Andrea Arcangeli <andrea@suse.de> SuSE
*/
I've just had a look at the original RCU patches contributed by IBM for 2.4.1, (RCU web site, http://lse.sourceforge.net/locking/rcupdate.html)
One of the copyright notices from a file:
Support for deferred freeing of memory using Read-Copy Update
mechanism.
Copyright (c) International Business Machines Corp., 2001
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
Author: Dipankar Sarma [dipankar@sequent.com]
(Based on a Dynix/ptx implementation by
Paul Mckenney [paul.mckenney@us.ibm.com])
IBM aquired Sequent in 1999, the copyright notice on this file from 2001. The important thing is that this code is based on an implementation and not the actual original code as such this is not Dynix/pty code.
SCO could argue that since it is based on the code then it is actually a copy, however from what I have seen of the contract clauses (mainly from groklaw) IBM are free to use any knowledge/processes/etc gained from adding to AIX.
Probably the two pieces of code look similar and I can believe even the comments since the implementor had access to original code. In addition to this it could be argued that it would be impossible to implement the code without there being similarities or even being nearly identical in places. For example try implementing a bouble-sort/linked list/etc without the code looking similar.
The only thing left are the RCU patents and IBM own these.
Having looked at this now SCO's case is pretty thin, they could win of course nothing is impossible (though unlikely), however it would just be a case of copyright violation and someone outside of IBM could do a re-implementation of RCU. In addition since RCU is n't part of the mainstream Linux kernel so I cant see how they could claim any substantial damages let alone $3 billion.
Actually the GPL is not anywhere near as viral as this license - if (big if) it means what SCO say it does.
The GPL doesn't take away any rights _you_ have to _your_ code, including adding it to other products under other licences. What it says is that if you combine your code with other GPL code you have to release the result under the GPL.
SCO seem to be claiming that if you add your code to their unix code, it isn't just the resulting unix variant that is a derivative work, but your code (in isolation) is also.
If SCO are correct, if you took code from another of your products and added it to Unix you would lose the right to ship that other product. That wouldn't happen if you added that code to a GPL product.
From the Securities Exchange Act 1934 (as amended), section 32:
The SEC has a good page sumarizing the relevant US Federal laws. (Incidentally, according to Netcraft, www.sec.gov is running Linux. Do you suppose they've received their menacing letter and invoice from SCO yet? :-)
If you check out the Nasdaq stock report for sco the following shareholders with >5% is shown:
Ralph Yarro
Darcy Mott
Canopy Group Inc
John Wall
If you then check out The Canopy Group website you will see the following as board members
Ralph Yarro - CEO & Presedent
Darcy Motto - Vice President, Treasurer and Chief Financial Officer
This gives the Canopy group at least a 15% stake in SCO, possibly more.
If you look at SCO's board and then have a look at their share trading, you see the following:-
Charles BROUGHTON (VP world ops) sold about 120,000 USD worth of shares in the last two months
Robert BENCH (CFO) sold about 120K USD in last few months
Jeff HUNSAKER (s. VP)sold about 120K USD in last few months
Other than McBride (CEO) most of the board have ofloaded shares in the last month or two. To give you an idea of the "peak" of share selling. According to the Nasdaq the number of "insider" trades (i.e. board members) in the last 12 months was 15, and 12 (all of which were sales) of those were in the last 3 months (or as its know just after the Linux thing).
Yarro and Mott both also sit on the board of SCO.
Does that stink or is it just me?
Jaj
I don't thing a court would issue an injunction against IBM, as that would effectively nullify IBM's contract. I can't see that happening. And IBM doesn't have to worry about SCO's termination their license. SCO can't terminate the license, period. It's written in contracts.
From SCO's Exhibit D
Note the terms irrevocable and perpetual. I don't think it can be made any plainer.
If you people had really been following the whole idea of derivative works you would have realized that there are *many* things that cannot be considered derivative works. One of those things are algorithms that can only be expressed a certain way (ie: software patents). Another one is code that is hardware dependent (ie: hardware patents).
NUMA and RCU are almost certainly considered to be one or both of those.
BRRRRKKKK! Sorry, SCO. The time is up. Not only have you still been releasing all of the kernel code under the GPL, you are trying to encumber another companies software and hardware patents.
Don't pass GO. Go striaght to jail.
Exhibit F
Exhibit G
I agree with the meat of your analysis about Sequent. I don't know enough contract law to comment on SCO's thesis about Sequent.
The Opensource organisation have an evolution chart. This is a simplifed version from the UNIX timeline history. Bang up to date as well (July 31st is the most recent update). It's in a similar style to those civilisation charts that nearly every History teacher has pinned on the wall. Looks like one civilisation is about to become extinct.
The Linux RCU code is not from AIX nor Dynix, please read this post: http://slashdot.org/comments.pl?sid=74586&cid=6686 573
- 24-Copyr ights_Story01.html
Infact SCO admit they down own the copyrights to RCU, NUMA or JFS as per this story:
http://mozillaquest.com/Linux03/ScoSource
Cutting through the noise SCO are claiming that IBM can not add technology added to AIX to Linux it's more to do with contracts than copyrights.
>the class (contract) lower down in the hierarchy (chronology); i.e., the more immediate parent node, takes precedence.
In the OOP languages that I'm familiar with, that's exactly what happens. It could be argued that this is the whole point of OOP, or at least a major one. The interesting case is when a language supports true multiple inheritance and a subclass directly inherits from two classes that have the same functions. (I mean that both classes are "parents", rather than one parent and one grandparent.) My C++ is rusty, but if I remember right it handles this case by saying that the parent class that was listed first in the declaration wins. Java (and C# I think) solves this by disallowing multiple inheritance.
For contracts, my understanding is that if a later contract between two parties explicitly contradicts an earlier contract, the later one wins. By explicitly, I mean that the later contract says that it's amending the earlier one, e.g. "This provision replaces section 3, paragraph 7 of the contract between X and Y dated Sept 2, 1997, hereinunder(1) attached as amendment A," or equivalent legalese. On the other hand, if the contracts are just plain contradictory (or can be interpreted that way), then the later one will usually win, but only after the lawyers and judges argue about it for a long time.
No, IANAL, but I've had to work with them a bit.
(1) I think that's the first time I've ever used the word hereinunder. Someone save me before it's too late!
People still don't seem to get it...
:->
IBM != Sequent (for year values less than 1992)
IBM had a contract/license variation with AT&T. Sequent did not. Anything IBM developed, IBM owned. Anything Sequent developed (before being bought and becoming part of IBM), based on the original AT&T license, belongs to the owner of SYS V, so the (rights to the) developements did not belong to Sequent and could not be sold to IBM in '92, so IBM does not own them, never did own them, and can not give them away - they belonged to Novell, and now to SCO.
Could this possibly be a (perhaps THE) legal leg for SCO's lawsuit?
This reply completely does not address your admittedly very good point about the patents, as I don't know how the license issue would afftect the patent issue. I.e., can I patent an idea that legally belongs to someone else, even thought it was my idea and I submitted it myself? I would guess 'yes', but it would depend greatly on the exact wording of any contracts/licenses - and we are right back to the license issue. Also, would you be able to charge me license fees on the idea you own, but I patent? Could I charge you for using my patent if you legally own the idea?
My head is starting to hurt...
Acts of massive stupidity are almost never covered by warranty. --me.
The company doesn't sell all its stock during an IPO. That would be stupid. Instead some of it remains with the company. They can do a split, sell stock, give it out as options. It is bargaining power and companies like bargaining power.
Maybe people should write letters to the following companys that are sponsoring SCOForum 2003, telling them you will no longer do business with them, as long as they support SCO.
List of Sponsors:
Hewlett-Packard
CRN
Microlite Corporation
Century Software, Inc.
Lone Star Software
DTR Business Systems
Digi International
TeleVideo
Multi-Tech Systems
InoStor
Open Systems
Rasmussen Software, Inc.
Tarantella
Equinox Systems
Maxspeed Corporation
TelSoft Solutions
BASIS International
Vultus, Inc.
Synergetic Data Systems, Inc. (SDSI)
fP Technologies, Inc.
TAKGroup
I found some similar code from System III in Linux
.h files
/2420/arch/ia64/sn/io/ate_utils.c - /sys3/usr/src/uts/pdp11/os/malloc.c; /sys3/usr/src/uts/vax/os/malloc.c /2420/drivers/macintosh/macserial.h - /sys3/usr/include/sys/reg.h; /sys3/usr/src/head/sys/reg.h /2420/include/asm-mips64/termbits.h - /sys3/usr/include/sys/tty.h; /sys3/usr/include/termio.h; /sys3/usr/src/head/sys/tty.h; /sys3/usr/src/head/termio.h /2420/include/asm-ppc64/termbits.h - /sys3/usr/include/sys/tty.h; /sys3/usr/include/termio.h; /sys3/usr/src/head/sys/tty.h; /sys3/usr/src/head/termio.h /2420/include/asm-ppc64/termios.h - /sys3/usr/include/sys/ttold.h; /sys3/usr/src/head/sys/ttold.h /2420/include/asm-s390x/errno.h - /sys3/usr/include/errno.h; /sys3/usr/src/head/errno.h /2420/include/asm-s390x/termbits.h - /sys3/usr/include/sys/tty.h; /sys3/usr/include/termio.h; /sys3/usr/src/head/sys/tty.h; /sys3/usr/src/head/termio.h
This code can not be found in BSD 4.4...
I think ate_utils.c is known but there are some
similar
I have the full report but is long enough to fit here
just use diff
If you do not have System III source, it can be found
on the Net as sys3.tar.gz
The issue is that Sequent developed the RCU code for Dynix/ptx (their SVR4/BSD personality hybrid) after publishing papers and submitting for patents on the related algorithms. The RCU code in Dynix/ptx is only an implementation of the IP, not a transfer of the IP itself.
All SCO has acquired is the right to use and distribute the RCU implementation Sequent deployed as part of Dynix. It does not transfer ownership of the RCU algorithms and patents to SCO, though it probably implies SCO has a perpetual right-to-use the RCU algorithms as implemented in Dynix.
SCO needs to prove that IBM used the Dynix implementation verbatim for AIX or Linux, rather than using code developed from the research papers and patents they acquired via Sequent. Further, they'll need to show that the Dynix code was not actually ported from a generic multi-OS code base (i.e. a generic OS sample implementation, such as you find in some textbooks.)
As multi-processor AIX was scaling quite nicely before IBM acquired Sequent, I don't think they're going to have any difficulty proving that AIX is not encumbered by Dynix code. Further, as Dynix was only running on x86 hardware while AIX runs on Power RISC cores, I really don't see how the core RCU code could even have made it's way from Dynix into AIX -- wouldn't the core RCU routines be hand-tweaked assembly rather than C code?
I do not fail; I succeed at finding out what does not work.
I just modded in this topic, so I had to logout to post.
I'm far far away from vegas, but SCO is having a show there in three days. If any slashdotters are in the area I suggest pulling the permits and protesting infront of the SCO show. As someone who has done work on SCO stuff in the recent past, I will now move every customer I can off their product and onto another operating system.
We need to show those that use SCO that the people who do support work on the SCO operating systems will no longer do any further work on SCO based systems (in the past couple of years I did some work on a Xenix system, so it's plural).
http://www.sco.com/2003forum/
Dynix/ptx started with a SVR3 source code tree. They became SVR4 by writing their own code to meet the API specifications in order to save some money over buying an SVR4 code base.
When a company is formed, a maximum number of shares is authorized. This number is usually quite large (many times what the company will issue for some time). Even then, this maximum can be changed at a later date. The number of authorized shares is pretty much meaningless. The idea that a company "holds back" shares or "doesn't issue all the shares" in an IPO is meaningless.
What is important, is the number of shares outstanding. That is the number of shares issued - shares bought back by the company (shares bought back are called Treasury Stock). Of course, if the company sells or gives away treasury stock, those shares are once again considered outstanding. The number of shares you own, divided by the number of shares outstanding is the percentage of the company you own.
The stock sold in an IPO is newly issued stock (in all but a incredibly small number of cases). Nobody has ever owned this stock before. The proceeds of this sale go directly into the coffers of the company (less the commission to the investment bank). One of the thing that the red herring & final prospectus describe how this money will be used.
The people and institutions that owned stock prior to the IPO still own the same amount of stock they did before. However, their shares are now diluted which means that they represent a smaller percentage of the company than they did before. Their shares are also Restricted Stock and cannot be traded on the open market. Their is a pain-in-the-ass process by which it can be converted in limited quantities, which is how insiders get to sell their stock eventually.
At a later point, the company may issue new shares to raise more capital. This is called a secondary offering. Once again, these are NEW, never before issued shares.
When a company goes through a stock split, no new shares are issued. It doesn't "use up" shares that are sitting around. The only thing that happens is that the old shares are replaced by a larger number (or in the case of a reverse split, smaller) of new shares.
The real limit on issuing new stock? Dilution lowers the price of the stock. The exchanges set minimum price levels and will boot a stock off the exchange if it trades below that price.
Sequent's licenses are available at SCO's web site as Exhibits G & F.
However, they are identical to the first two IBM licenses, Exhibits A & B at SCO's website, so if you've already seen those, then you've pretty much seen the Sequent licenses.