Groklaw is, as you know, a group effort. So, while I was kicking up my heels at LinuxWorld, Dr Stupid kept slaving away in the Groklaw research dungeons. And he has found something so amazing, I think you will join me in thanking him for effectively blasting out one rung in SCO's ladder theory, the most important rung too, the one about methods and concepts, and their claim to contractually control what IBM can do with its own derivative works.
Both judges in the SCO v. IBM case have highlighted that issue, the contract issue, of whether AT&T retained control over modifications and derivative code, based on methods and concepts as opposed to straight copyright infringement, as being the one IBM needs most to disprove, at least by my understanding of their rulings. No doubt SCO's witnesses it offered declarations from, including Mitzi Bond, were a factor in the judges having to at least consider this issue and address it. It is with great pleasure that I present Dr Stupid's findings, therefore. As you will see,
it shows that Otis Wilson and David Frasure, in their declarations for IBM, are
recalling how things worked back in the day.
For context, you will recall that Mitzi Bond's declaration [PDF], which contradicts her bosses, Wilson and Frasure, (and AT&T attorney Geoffrey Green and all the other IBM witnesses), indicates that according to her current memory, methods and concepts were in fact covered by the AT&T license agreement terms, not just the code itself.
Specifically, on page 4, she claims that "the agreements protected all of the innovations embodied in the licensed UNIX software, including, among other things, the structures, sequences, patterns, methods, and concepts." She also says that licensees had to treat any work "prepared with the benefit of having been exposed to our product -- regardless of how the license further developed or changed that product -- as if it were part of our original licensed product."
Is that true? Let's see what a stroll down Google's memory lane reveals about AT&T policies regarding methods and concepts and also what Ms. Bond testified to in the earlier BSDi case back in the '80s. Ms. Bond quoted only a portion of that earlier testimony, and herein lies the tale. (Here [PDF] is the selection SCO attached to the Declarations of Jeremy Evans regarding SCO's Opposition to IBM's Motion for
Partial Summary Judgment on Breach of Contract Claims, and it's the section of her testimony that they feel is useful for their side.) .And let's also note an email from Ms. Bond herself on this very subject, dated July 24, 1987.
What I now believe really happened is this: AT&T, after divestiture in 1984, briefly tried to put methods and concepts in as a part of the definition of what the term "Software Product" in the contract meant. The licensees, however, wouldn't accept the change and insisted on that language being removed, which AT&T did, by means of letter agreements and side agreements, just as we have seen with respect to IBM's license agreements. So at first, methods and concepts were not covered; then AT&T tried to make them covered; objections made it impossible for them to maintain that change, and it was dropped, or in AT&T's language, "clarified" away. SCO, to my eyes, is taking that brief attempt to broaden AT&T's rights, which failed, and trying to claim enforcement of something not even AT&T was able to enforce originally. See if that is what you get from the historical documents Dr Stupid has found.
Your Word is your Bond?
~ by Dr Stupid
The Bond declaration references parts of Mitzi's 1992 deposition. This
in itself is interesting, since earlier SCO has denied
having any access to these depositions:
[IBM has asked for]“The pleadings, deposition transcripts
and deposition exhibits from UNIX System Laboratories, Inc. (“USL”)
suit against BSD.”
SCO does not have such documents in its custody, control or
While we don't have the full
deposition, we do have knowledge of other excerpts, thanks to the amicus
brief filed in the USL/BSDI litigation.
"[fn11] The SOFT-00089
license agreement was intended to 'clarify' the prior intent of the
parties.Wilson Depo., at 139:10-140:1. AT&T's Director of
Licensing, Otis Wilson, testified that the language used prior to that
contained in the SOFT-00089 license agreement was 'somewhat confusing
to some people in that they thought we were trying to assert ownership
to anything they created, even though it contained nothing of ours. So
this is to clarify that what's yours is yours and what's ours is
ours' (emphasis added). Wilson Depo., at 75:24-76:4.
"USL's Mitzi Bond admitted that she understood 2.01(b)ii) meant
that 'enhancements and modifications made by the licensee were to be
made available to anyone so long as they did not include any portion of
the software products licensed under the agreement' emphasis added).
Bond Depo., at 137:19-138:19. [Dr S: this SCO do not
reference for some strange reason]
"However, Ms. Bond also has espoused a mental 'contamination'
theory by which any university student exposed to
AT&T code would be beholden to AT&T for any software
product he/she might subsequently develop. Bond Depo,at 220:2-13,
237:24-238:8. [Dr S: this is the section SCO reference]
"USL's Mitzi Bond participated in the preparation of [$echo]. She
testified that the clarification to 2.01 'indicates that all we
did was add a sentence to the existing one so that licensees would
clearly understand that we were not claiming the ownership in code that
they developed that didn't contain ours' (emphasis added). Bond Depo.,
Unfortunately, the copy we have loses where the BSDI lawyers added the
emphasis they noted.
The first thing to note is that originally Ms Bond agreed with Mr
Wilson as to the intent of the agreement. How can we square this with
her espousal of the "mental contamination" theory? We can get a clue
filing in the same case, BSDi's Opposition
To USL's Motion For Preliminary Injunction. Note that in this filing, the actual
quotes are redacted:
"After divestiture, which took place in 1984, AT&T began to market
its UNIX source code actively in the United States and in many other
countries as a commercial product. As a result of this widespread
licensing and the access permitted to students all over the world, as
USL admits, the UNIX source code has become 'world-renowned,' and
many portions of it have become industry standards. Although USL
may have strict internal safeguards in place, it has so widely
licensed its UNIX source code that it cannot be considered a 'secret'
by any stretch of the imagination. Indeed, since the source code has
been licensed to universities all around the country, most people
who have studied computer science at universities over the past
fifteen years have had access to the source code at some time,
E to the Forte Aff.). Further, the strict safeguards USL claims to
observe have not been observed by its thousands of licensees.
consequently, USL's licensed code periodically appears on publicly
accessible computers. (Bond Dep. T. 71-24 to 74-16;[fn4] Adams
Aff., ~ 10).
"[fn4] All citations to the deposition transcript of Mitzi D. Bond are
attached as Exhibit F to the Forte Aff.
"USL has also permitted numerous books to be written on the UNIX
source code. These books describe nearly every aspect of the code,
including its design, architecture, sequence, organization, structure,
implementation and algorithms, and provide instruction on how to
write UNIX-compatible code. See infra. Among other things, such
publicly available reference materials disclose the specific features
USL's expert relies upon as evidence of copying and trade secret
misappropriation. (See generally McKusick and Bostic Decl.; Kashtan
Aff., ~l 6, 8). . . .
"2. The Header Files Which USL Claims Have Been Infringed Have Been
Included in Binary Distributions Made by USL Licensees With No
"The binary version of the UNIX operating system has been even more
widely distributed than the source code version. Although current
binary versions of the code contain copyright notices, earlier
versions did not. (See Bostic & McKusick Decl., 28.4).
"Moreover, these binary distributions included some source code
files, and in particular, included the header files USL accuses the
University of copying in Net2. (Bostic & McKusick Decl., ~ 28.4;
Kashtan Aff. ~ 8).
"USL itself has never distributed binary code, but has sublicensed
numerous other companies to distribute binary versions of its UNIX
operating system, placing no limitations on whom they could license
it to, other than those required by law. Thus, USL cannot possibly
argue that the distribution of the binary code (including the header
files in source form) was a limited publication. Not only is it
impossible for USL to meet the 'definitely selected group' test, but
simply offering a work 'for further distribution,' as USL did with its
binary code, in itself constitutes a general publication of the work.
See White, 193 F.2d at 746-47; Rexnord, 379 F. Supp. at 1197. Thus,
USL has no basis for claiming any copyright in the header files. . . .
"Although the language in these agreements was clarified over time,
the basic intent with respect to the University's right to use the
software and its confidentiality obligation was the same in each of
them. (See, Karels Aff., 1 24; see also Wilson Dep. T. 36-17 to 37-8).
Under each of these agreements, the University was authorized to
make modifications, enhancements and derivative works using the
licensed software. If a derivative work contained the licensed
software or disclosed any proprietary methods or concepts used in
the licensed software, it could be distributed only to those who had
purchased a source code license from AT&T. As demonstrated below,
however, if an enhancement, modification or derivative work did not
contain or disclose any AT&T proprietary information, it belonged to
the University, and was not governed by the restrictions applicable
to the licensed software. See discussion infra.
"2. The License Agreements Permit Derivative Works Not Containing
USL Trade Secrets To Be Made Available To Non Licensees
"Section 2.01(b) of the E-SOFT-00089 Agreement expressly provides
that the University was to distribute to anyone, on an unrestricted
basis, modifications and enhancements that did not include USL's
proprietary information. Likewise, Section 2.01(a) of the E-SOFT-
00089 Agreement, provides that derivative works "that contain"
licensed software must be treated as licensed software i.e, must be
subject to the same restrictions on disclosure as the licensed
[Licensee's] right to use includes the right to modify such software
product and to prepare derivative works based on such software
product, *provided that any such modification or derivative work
that contains any part of a software product subject to this
Agreement is treated hereunder the same as such software product.*
(See Exhibit D-25 to Bond Dep.T. (emphasis added)).
"The negotiations leading up to the execution of this agreement, as
well as the prior course of dealing between the parties, make it
clear that the words "that contains any part of a software product
subject to this agreement" were intended to distinguish between
derivative works that actually contained licensed software and
derivative works that did not- while the former were subject to the
license restrictions, the latter were not.
"In the negotiations leading up to the execution of the E-SOFT-00089
Agreement, USL initially proposed language providing that any
derivative work 'based on' the licensed materials was to be treated
as licensed software. The University objected to that language,
however, and insisted that it be changed to cover only derivative
works 'that contain' licensed software. USL accepted this
modification, providing a new version of the E-SOFT-00089
Agreement in which the language 'based on' licensed materials had
been changed to 'that contains' licensed software. (Wilson Dep. T.,
Exh. D-67 and D-68 (attached as Exhibits Y and Z to the Forte Aff.).
"This same clarification i.e., that only derivative works containing
AT&T code had to be treated as licensed software had also been
made in discussions surrounding a prior license agreement, the 'UNIX
System V Agreement'. This agreement initially included within the
definition of 'LICENSED SOFTWARE' any derivative works prepared by
the licensee 'based on' the licensed materials. The University
objected to that provision, which had not appeared in any of its prior
agreements with AT&T, and was contrary to its understanding of its
obligations. (See Karels Aff., ~ 23-26). Following discussions
regarding the provision, USL wrote a letter agreement to the
University 'clarifying' the definition of 'Licensed Software' by
deleting from the definition 'derivative works prepared by the
licensees'. The letter further stated that AT&T never had intended
to claim any interest in such works. (See Exhibit B to Karels Aff.;
Exhibit D-34 to Wilson Dep. T. (Bates P000243) (Exhibit aa to the
Forte Aff.)). In place of the prior language, a sentence was added
stating 'LICENSEE agrees that any modification or derivative work
prepared by it *that contains any LICENSED SOFTWARE* shall be
treated as LICENSED SOFTWARE hereunder.' (Emphasis added).
"USL now attempts to read into the earlier 32V agreement (which it
mistakenly argues applies), the same language that was deleted
from both the System V Agreement and the E-SOFT-00089
Agreement even though no such language appears in either of those
agreements. Contrary to USL's argument, this same point also was
clarified with respect to the April, 1979 Agreement: In a letter
dated January 18, 1978, the University inquired whether it could
distribute UNIX-derived software.[fn18] In a written response,
Edwin G. Baldwin, who was AT&T's Director of Technology Licensing
at the time, stated:
We take no position with regard to your use or distribution of
software developed by you which does not contain any of our
proprietary information such as, without limitation, the computer
programs and documentation, or any portion thereof, related to the
UNIX operating system, including the methods and concepts utilized
With regard to the distribution of any software which does contain
some of our proprietary information, our primary concern is the
prevention of unauthorized disclosure, either intentional or
inadvertent, which might jeopardize our valuable proprietary rights
in such information.
[fn18] Exhibit D-11 to Baldwin Dep. T. (attached as Exhibit bb to
the Forte Aff.).
(See Exhibit-12 to Baldwin Dep- T- (attached as Exhibit cc to the
Forte Aff.) (emphasis added)). Mr- Baldwin testified at his
(Baldwin Dep. T. 33-15 to 34-16). Mr. Baldwin further testified that
"Another USL employee, Mitzi Bond,
similarly testified to her understanding that
(Bond Dep. T. 138-14 to 138-19, 137-13 to 138-13).
"After consulting with Ms. Bond during a break at the deposition,
USL's attorney cross-examined her, at which point she changed her
testimony, now stating that it was her understanding [REDACTED]
215-24, 216-20 to 216-25, 217-6 to 217-15, 219-3 to 219-15,
224-18 to 224-25).
(Frasure Dep. T. 172-11 to 172-20).
36-18 to 39-10, 49-20 to 50-2).
"Given that USL twice attempted to include a provision that any
derivative work "based on" the licensed software would have to be
treated as licensed software (in the 1983 System V Agreement and
the 198S E-SOFT-00089 Agreement) and both times the University
insisted that the language be narrowed to cover only derivative
works that "contain" licensed software -- the argument that the
intent was to cover any work "based on" the licensed software, or
any work benefiting from knowledge of the licensed software, is
completely unsupportable, much less the argument that anyone who
has ever seen UNIX source code is "mentally contaminated" and may
never write another piece of code that is not subject to USL's
The first page range regarding the Mitzi Bond testimony corresponds to the "agreeing with Wilson" section of her testimony in the BSDi case, and
the latter page range corresponds to the section quoted in the Bond declaration in the SCO case. In
other words, it appears the "mental contamination" aspect was something Bond came
up with not as a direct
result of her training from Wilson, et al., but after what might be
seen as "coaching" from USL's attorney during the break. This is information that is lacking from her current declaration.
I must hasten to add that one must not rush to conclusions about Ms
Bond from this observation. There are many possible reasons why she
might change her story -- for example, not having been involved in the
devising and drafting of the licenses and USL policy in general, she
may simply have been open to having her understanding "corrected" by
some prompting, albeit to a interpretation of the AT&T license that
not even SCO have put forward. But the key fact to take away is that
she was originally of one mind with Wilson -- her "contradiction" of him
is a later construct.
These documents date from 1992, at a time when the people involved
(Bond and Wilson, to name but two) were having to recall past policies.
Interestingly, a USENET discussion that broke out during the lawsuit
unearthed some documents that are more contemporaneous.
post is the first salient one from the thread (emphasis is mine):
From: email@example.com (Diomidis
Subject: Re: What happened with AT&T's copyright clearing
firstname.lastname@example.org (Diomidis Spinellis) writes:
>I seem to remember, that some public domain source code that
>rounds of the net a number of years ago (five?) had a
statement from an
>AT&T person attached, stating that the code did not
contain and was not
>based on any AT&T intelectual property. It also
gave a contact for
>further enquires. The idea I got was, that if you had
some code that
>could have been AT&T's intellectual property you mailed
it to someone
>at AT&T who told you whether it was or not. Does
anyone know whether
>that procedure still exists, and if yes, why BSDI did not use
it, if no
>why it was dropped by AT&T?
Thanks to Alan Cox's memory, I managed to locate the file I was
about. It is part of the uuslave distribution (an early
written by John Gilmore. The file s named CERTIFIC.ATT and
Jun 8 1987; I enclose a copy at the end of this
point is, that John Gilmore had
almost definitely been exposed to Unix
source code at that time.
Date: Mon, 23 Mar 87 23:49:43 PST
Listed below is the source code for uuslave that you
mailed to me on February 20, 1987. Our product management
have reviewed this code and have determined that it was not derived
from source code from
versions of our UNIX(r) operating system. If
you have any further questions, please feel free to call me at
David W. Frasure
AT&T UNIX Software Licensing
Diomidis Spinellis Internet:
Department of Computing, Imperial College, London
SW7 #include "/dev/tty"
John Gilmore had been exposed to AT&T's UNIX code, and was
naturally steeped in its methods and concepts. The program, uuslave,
was intended specifically for UNIX (the two u's refer to
"UNIX-to-UNIX", as they do in UUCP.) and in fact was an unabashed
"clone" of an existing UNIX program. Yet Frasure is entirely
unconcerned by this, since the code was not a derived work of the UNIX source code - viz., did not
infringe the UNIX copyrights.
post closes our circle - again, emphasis mine:
From: email@example.com (John
Subject: Re: What happened with AT&T's copyright clearing
Thank you, Mr. Spinellis (who posted this).
I and some friends proposed a "sift" project at Jan '87
idea was that some impartial Unix wizards would sift through
Unix and decide which parts were Bell, and which parts were
and then get Berkeley and AT&T to verify that decision.
I initially got involved with AT&T in checking the license
of software in a somewhat roundabout fashion. As a
correspondence from me describes it:
Actually they first contacted me; I was working on uuslave and I
some anonymous warnings (relayed through friends) that said
it was theirs and was going to "crack down" on whoever was
it. I posted a challenge to the net, saying if they thought
theirs they'd better come public or lose whatever rights they
had in it. Someone in
AT&T forwarded it to the right people, who turned
out to be at the licensing dept at
828-UNIX, I sent them a copy, and
they gave it a clean bill of
health. They end up sending it to Summit
to be read over by the experts, which is why it's so slow -- I can
imagine your reaction to having to read a bunch of this stuff and
whether it's part of Unix or not, rather than doing more
But there's way too much stuff that's not theirs which is hiding
shadows of "they might sue us because we don't really know who
and if AT&T really does want to control Unix, they'll have to
bullet and say what Unix really is and isn't.
(end of that excerpt. Here is another
with further history.)
The next two things I sent in were mdbm and rcs. I sent
mdbm on 25
March 1987, but it got lost inside AT&T because David
Frasure, who was
handling it, retired. I checked on May 29th and discovered
had lost it; talked with Chuck Green who took over for
Dave. I also
mentioned RCS and they said "send it too, we will do them both".
I sent both to Mitzi Bond,
who works for Chuck, on 1 June 1987,
in 9 shar files (one mdbm, 8 RCS). I checked to make sure
received it, and got email back from Mitzi on 13 June saying that
got it all and were sending it to the developers. I checked
got mail on 2 July saying that Charlie Thiel in Summit was still
reviewing the code but expected an answer by the end of next week.
On 13 July, Mitzi was on vacation for a week. On 24 July I
final answer by email:
Date: Fri, 24 Jul 87 22:47:47 PDT
To: John Gilmore
From: Mitzi Bond
Charlie Thiel at our UNIX software development group in Summit
called me back
to say that his developers has examined the code that you
sent. Their results
were basically indecisive. There was nothing in either set
of code to
determine that it was solely AT&T's. There was a lot of similarity but not
enough for AT&T to build a case to
say that this was strictly AT&T code
and its use in both instances is a
violation of the copyright of our
Charlie also says that it is the duty of the University of
Maryland to protect
the software and their lies the responsibility in that particular
I apologize for taking so long to respond. I talked with
Charlie on Monday
when I returned, and again today for the final results, which
don't seem to
be as conclusive as we would have hoped.
Thank you for your patience.
(end of 2nd excerpt of old correspondence.)
I think (personal opinion) that AT&T decided it was better to
losing the big court case over this (refusal to help potential
infringers check if they were infringing) than to have to
by piece, that indeed major chunks of what the world thinks of as
did not actually infringe AT&T's rights.
"It isn't given to us to know those rare moments when people
are wide open and the lightest touch can wither or heal."
So there you have it. If you sent some software you'd written to
AT&T to get clearance, did they look for "methods and concepts"?
Did they wonder
if you had "benefitted from" UNIX in writing it?
Mitzi Bond herself, in effect, tells the writer that the criterion is
whether the work is a derived work in
the copyright sense. This was back in 1987, when there were
arguably more secrets to be protected in UNIX; and it also chimes with
David Frasure's words regarding uuslave, as well as his (and Wilson's)