Here is SCO's Memorandum Regarding Discovery as text. Our thanks to Steve Martin for transcribing and coding.
Brent O. Hatch (5715)
HATCH, JAMES & DODGE
Robert Silver, Esq. (admitted pro hac vice)
BOIES, SCHILLER & FLEXNER LLP
Stephen N. Zack (admitted pro hac vice)
Mark J. Heise (admitted pro hac vice)
BOIES, SCHILLER & FLEXNER LLP
Attorneys for Plaintiff
IN THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF UTAH
THE SCO GROUP, INC.
Case No. 2:03-CV-0294 DAK
Honorable Dale A. Kimball
Magistrate Judge Brooke C. Wells
On March 3, 2004, this Court directed International Business Machines Corp. ("IBM") to
produce certain AIX and Dynix files. March 3, 2004 Order ("March Order") at 4, ¶ 1. It also
directed The SCO Group, Inc. ("SCO") to explain whether and how these files supported its
position in this case, and further stated that it would consider ordering IBM to produce additional
code upon SCO's identification of additional files and the reasons for such requests (id.). SCO
states as follows in response to the March order.
SCO needs the discovery requested in this Memorandum for at least three reasons:
To respond further to IBM's discovery requests. IBM claims discovery
misconduct and delay by SCO in providing discovery responses, while IBM has the information
SCO needs to respond.
To gather evidence relevant to IBM's theory of SCO's contract claim, under
which IBM appears to assert that to establish IBM's breach of the Agreement at issue, SCO
requires specific evidence of derivation from UNIX System V lines of code through the versions
of AIX and Dynix code to Linux.
To gather relevant evidence in support of SCO's theories of its contract case
against IBM, including without limitation: (a) that every transfer from AIX and Dynix, as whole
programs, into Linux constitutes a breach of the Agreement, and/or (b) that AIX and Dynix are
derivative works of UNIX System V, and consequently, IBM's contributions to Linux from AIX
and Dynix are breaches of the Agreement.
A. SCO's Breach of Contract Claims.
SCO claims that IBM breached its software agreement (AT&T Technologies, Inc.
Software Agreement, executed on Feb. 1, 1985 ("Agreement") [Exh. 1]) by, among other
actions, distributing parts of the AIX and Dynix operating system software programs to Linux.
Because the Agreement does not permit IBM to lease or transfer any parts of the AIX and Dynix
programs, IBM's contribution of AIX and Dynix code to Linux violates that Agreement.
contributions to a public forum also violate the requirement to maintain the "resulting materials,"
AIX and Dynix, as confidential. Id. at § 7.06.
SCO also claims that IBM's contribution of parts of AIX and Dynix to Linux breached
the Agreement because both the AIX and Dynix programs, as whole programs, are modifications
or derivative works of UNIX System V, so no parts of those programs may be sold, leased or
otherwise transferred. See note 3. Again, for example, because the Agreement prohibits the
transfer of software programs in whole or in part, IBM's contributions to Linux from AIX and
Dynix -- as derivative works of UNIX System V -- violate the Agreement. SCO is entitled, under
the Federal Rules, to evidence relevant to these claims.
B. IBM's Apparent Assertions as to Proof SCO Must Offer to Establish IBM's
Breach of Contract, SCO's Entitlement to Establish Liability under
Alternate Theories, and SCO's Entitlement to Relevant Discovery.
IBM is apparently taking the position that in order for SCO to succeed on its contract
claims, SCO must prove copyright infringement. This position -- if taken as a description of the
sole basis for IBM's liability for breach -- confuses contract with copyright and thereby
essentially eliminates the protections of the contract for SCO. Nevertheless, SCO recognizes
that the rules of discovery obligate it to respond, to the best of its ability, to IBM's discovery
requests. And in any event, SCO, like any plaintiff, is entitled to proceed under alternate theories
of liability, and to obtain the discovery needed to allow it to do so.
SCO must have access to the requested discovery to (1) show, in one of many possible
ways, that AIX and Dynix are modifications of and derivative works based on UNIX System V;
(2) obtain evidence that IBM donated AIX and Dynix into Linux, and to prove the origin and
extent of such donations; (3) have access to all relevant information regarding IBM's
contributions to Linux through AIX, ptx, Dynix/ptx, and Dynix(2); and (4) review all relevant
sources reasonably calculated to lead to the discovery of admissible evidence (Fed. R. Civ. P.
C. IBM's Discovery Demands and SCO's Entitlement to Obtain Evidence
Relevant to Responding to those Demands.
IBM's broad discovery requests relating to the derivation of Linux from AIX and Dynix
code require SCO to do a detailed code analysis. The means for conducting such an analysis and
presenting proof, however, is solely in IBM's hands.
SCO must have access to IBM's revision control systems, to interim versions of AIX and
Dynix which IBM has declined to produce, to programmer notes and design documents related
to modifications and revisions to the programs, and to other materials which will allow SCO to
provide complete answers to IBM's interrogatories asking for the lines of code from UNIX
System V from which IBM's contributions from AIX and Dynix/ptx were derived; to defend
against IBM's Cross-Motion for Partial Summary Judgment on its Claim for Declaratory
Judgment of Non-Infringement; and to supplement SCO's response to the Court's March Order
(see March Order at 2, ¶¶ 2-3).(3)
Further, IBM has argued that "SCO does not need any AIX and Dynix source code for its
analysis" -- and that "IBM has met all of its discovery obligations".(4) To the contrary, SCO needs
revision control system information(5) and versions of AIX and Dynix to further respond to IBM's
discovery demands, and to counter IBM's argument that SCO has "delayed in providing full and
complete responses to IBM's discovery requests". IBM Opp. at 4. Without the listed items (see
Section II infra), to which only IBM has access, SCO will not have the information to which it is
entitled under the Federal Rules, nor will it be able to further respond to IBM's interrogatories.
D. IBM's Previously Produced Files
The files previously produced by IBM pursuant to this Court's March Order show that
IBM improperly contributed code to Linux, but these materials are only partial responses to
SCO's requests for information and materials. IBM has produced selected pieces of AIX and
Dynix, consistently refusing to produce all requested versions of AIX and Dynix,(7) or the
revision control systems for either program. Because Dynix code contains some historical commentary,
SCO has been able to trace some sources of contributions to Dynix code. This research has
revealed that Dynix contains material and significant UNIX System V code.
AIX almost certainly has a similar history, but because AIX's code does not contain any
historical comments, or at least the AIX code provided did not, SCO has had difficulty
determining all the portions of AIX that were taken from UNIX System V. To trace the
foundations of AIX source code, and to have a complete understanding of Dynix's source code
origins, SCO must have access to all interim and final versions of AIX, Dynix and ptx; to
programmer notes and design documents that reveal the work behind the revisions to the
programs; and most importantly, to all revision control systems that track changes to AIX and
Dynix, thus exposing the sources of IBM's current AIX and Dynix code and revealing what
portions of UNIX System V made their way from AIX and Dynix into Linux.
E. SCO's Efforts to Respond to IBM's Allegations Without Complete
Production from IBM.
IBM has selected snapshots of AIX and declined to provide either the change-log
information or the revision control information showing the changes between the various
snapshots. Consequently, it is difficult if not impossible to find the original sources of the
improper contributions of SCO's contractually protected code. Without the listed items, SCO
has spent countless hours, and sometimes fruitless effort, trying to track the improper use of
UNIX System V code in Linux through AIX and Dynix.
In the absence of the code, programs and other materials requested, the following tasks
and efforts have been undertaken by SCO:
Identification of all publicly available code patches contributed by IBM/Sequent
engineers to Linux, to the extent identified on IBM's web site and in Linux forums.
Tracing of code in available code patches to versions of AIX and Dynix.
Identification of all publicly available categories in Linux that are or have been
worked on by IBM/Sequent engineers, to the extent identified on IBM's web site or
obvious from Linux forums.
Identification of IBM's role in Open Source Development Laboratory (OSDL), and
IBM's role in Carrier Grade Linux project and Data Center Linux project, as
described in the OSDL website.
Review of all publicly available sources to identify the modules of UNIX System V
code improperly misappropriated into Linux by non-IBM sources, to the extent possible.
Identification of IBM's public statements regarding its roadmap for building Linux
and its technology contributions to Linux.
Identification and tracing of certain UNIX System V file system code, including
structures, sequences, and interfaces in UNIX System V code, that became modified
or derivative file system (JFS1) code in AIX; identification and tracing of the same
file system code (JFS1) in AIX to modified, derivative file system (JFS2) code also in
AIX; and identification and tracing of the same file system code (JFS2) to modified,
derivative file system code in Linux, to the extent possible based on the limited
production made available by IBM.
Identification and tracing of certain multiprocessor lock-avoidance code (RCU) from
Dynix/ptx to become modified, derivative multiprocessor lock-avoidance code (RCU)
in Linux that appears to be based on proprietary UNIX-based code, methods or
Identification and tracing of certain multiprocessor memory access code (NUMA)
from Dynix/ptx to a code dump made available to Linux that appears to be based on
proprietary UNIX-based code, methods or concepts.
Identification of substantial activity by former Sequent engineers and IBM engineers
in SMP design, locking and performance enhancement in Linux that appears to be
based on proprietary UNIX-based code, methods or concepts.
Identification tracing of substantial activity by former Sequent engineers and IBM
engineers in debugging, testing and logging information for use by other Linux
engineers in improving Linux performance in other areas that appears to be based on
proprietary UNIX-based code, methods or concepts.
Identification and review of all design documents related to multiprocessor
technology created by USL engineers, or as work for hire by Sequent engineers
for USL relating to multiprocessor technology ("ES/MP design documents").
Identification of the areas in Unix System V ES/MP code base that contain
multiprocessor technologies based on the proprietary ES/MP design documents.
Identification of, to the extent possible from IBM's production, the former Sequent
engineers who had access to System V and to Dynix, and have made contributions to
Identification of, to the extent possible from IBM's production, the AIX
engineers and contractors who had access to System V and to AIX, and have made contributions to
Identification of, to the extent possible from IBM's April 19, 2004 document
production, the general areas of Unix SVR4 code, methods and concepts adapted for
use by IBM in AIX 5.1, AIX 5.2 and AIX5L. (Detailed code review has not yet
These efforts have not, however, yielded much of the information required for SCO to
further respond to IBM's discovery demands. SCO has attempted to follow IBM's scattered path
through the winding history of countless alterations, derivations, and revisions, but the task is
nearly impossible without a map, a map so easily accessible to IBM, so clearly relevant to this
case, and so absolutely essential to SCO that IBM's withholding of it and subsequent filing of a
summary judgment motion is unconscionable.
II. ADDITIONAL ITEMS REQUESTED(8)
SCO requests that this Court order IBM to produce the following:
all revision control system information (including documents, data, logs, files,
and so forth) for AIX, Dynix/ptx, ptx, and Dynix from 1984 to the present
source code and log information for all interim and released versions of AIX,
Dynix, ptx, and Dynix/ptx from 1984 to the present
In addition, SCO requests that this Court order IBM to produce all design documents,
whitepapers and programming notes, created from 1984 to the present, related to the following:
As to AIX:
Journaled File Systems (including PFS, JFS, and J2) and all other file systems
used or developed by IBM in AIX (including System V file system)
Extended Volume Management System (EVMS)
Async Input/Output (I/O)
symmetric multiprocessing (SMP) and other multiprocessing (MP)
documents related to performance measurement and management
SVR4 features in AIX
debugging and statistics gathering for both kernel and user-space functionality
As to Dynix, ptx, and Dynix/ptx:
read-copy update (RCU)
non-uniform memory access (NUMA)
Async I/O, MPIO and Standard I/O
SMP and all other changes related to MP functionality
performance enhancement tools and methods
SVR4 features in Dynix and/or PTX
debugger and statistics gathering for both kernel and user-space functionality
virtual file system
MP enhancements to network file system
System V file system
Each of these items falls within previous SCO requests to IBM, but IBM has resisted
production without SCO's filing of this Memorandum. See Letter from Christoper Kao to Mark
Heise, dated April 14, 2004 [Exh. 2].
A. The Listed Materials are Relevant and Necessary
Under the Federal Rules, parties "may obtain discovery regarding any matter, not
privileged, which is relevant to the subject matter involved in the pending action" or "reasonably
calculated to lead to the discovery of admissible evidence." Rodgers v. Hyatt, 91 F.R.D 399,
402 (D. Col. 1980), citing Fed. R. Civ. P. 26(b). This rule is to be liberally interpreted. Id.
"'[O]nce a party has requested discovery, the burden is on the party objecting to show
that the discovery is not relevant....'" Smith v. MCI Telecommunications Corp., 137 F.R.D
25, 27 (D. Kan. 1991) (citations omitted). This is a difficult burden to satisfy, especially in the
discovery stage of litigation. Id., citing Miller v. Doctor's General Hospital, 76 F.R.D. 136, 138-
39 (W.D. Okla. 1977) ("Relevancy is broadly construed at the discovery stage of the litigation
and a request for discovery should be considered relevant if there is any possibility the
information sought may be relevant to the subject matter of the action."); Gohler v. Wood, 162
F.R.D. 691, 695 (D. Utah 1995) ("[A]t the discovery stage, the concept of relevance should be
construed very broadly.").
Not only are the revision materials highly relevant to this litigation, but they are also
necessary for SCO to further answer IBM's interrogatory requests (IBM's Second Set of
Interrogatories to SCO, Interr. Nos. 12-13; Fourth Set, Nos. 15-16), which require SCO to list
lines of code which were derived from UNIX System V. Also, this Court has ordered SCO to
make these identifications. March Order at 2, ¶ 3. Unless and until IBM produces the requested
materials, SCO cannot complete these discovery tasks.
SCO's need for the materials is especially evident because IBM's discovery requires
SCO to compare one thing to another. In Weahkee v. Norton, 621 F.2d 1080 (10th Cir. 1980),
Weahkee had sued the EEOC for discrimination, and requested the personnel files of EEOC
employees whom he contended were hired or promoted in discriminatory preference over him.
The Court refused to compel their production. Finding that this refusal constituted an abuse of
discretion, the Tenth Circuit honed in on the plaintiff's need for access to those files so that they
could be compared to his: "[t]he qualifications and job performance of these employees in
comparison with the plaintiff's qualifications and performance is at the heart of this
controversy." Id. at 1082 (emphasis added). SCO's need for the revision trees is of a kind with
Weahkee's, in that IBM believes SCO must compare UNIX System V, AIX, Dynix, and Linux side-by-side
and show derivation. Due to the constantly evolving nature of these operating systems,
mere snapshots in time (what IBM has provided) are not enough.
Finally, the requested materials are not available from any source other than from IBM. Of
[A] court may limit discovery where "the discovery sought is ... obtainable from
some other source that is more convenient, less burdensome, or less expensive."
Fed. R. Civ. P. 26(b)(2)(i). The court also may limit discovery if "the burden or
expense of the proposed discovery outweighs its likely benefit." Fed. R. Civ. P.
American Medical Systems, Inc. v. National Union Fire Ins. Co. of Pittsburgh, PA, 1999 WL
562738, at *2 (E.D. La. Aug 2, 1999). However, none of these apply to limit discovery here: the
discovery SCO has requested from IBM is not available from any other source, nor will IBM
bear any burden or expense in producing the requested materials that outweighs the benefit of the
discovery (See Section III.C, infra).
SCO Needs the Materials to Respond to IBM's Allegations.
IBM argues that SCO must list lines of code which were obtained from UNIX System V
(IBM's Second Set of Interrogatories to SCO, Interr. Nos. 12-13; Fourth Set, Nos. 15-16), and
that without providing such information, IBM is entitled to summary judgment on
noninfringement of SCO's copyrights. One way SCO can unequivocally establish derivation and
modification -- and defeat IBM's arguments -- is by examining AIX and Dynix source code
directly.(9) Without access to AIX and Dynix source code, SCO has little IBM code to compare to
UNIX System V source code. The AIX and Dynix source code that IBM has already provided
shows derivation and modification in the sense that UNIX System V code is present in those
versions. See Exhibits E and F to Letter from Brent Hatch to Todd Shaugnessy, dated April 19,
2004 ("Hatch April Letter") [Exh. 3]. To this extent, these files support SCO's position that AIX
and Dynix are modifications of, and derivative works based on, UNIX System V.
SCO alleges in its contract claim that IBM contributed AIX and Dynix to Linux in
violation of its Agreement. SCO also alleges that AIX and Dynix, considered as whole
programs, are derivative works of UNIX System V, and again -- under the Agreement -- cannot
be split into parts and distributed. To establish that AIX and Dynix are derivative works of
UNIX System V and that their source code was contributed to Linux, SCO must be able to
compare UNIX System V, AIX and Dynix, and Linux source codes. The source code that IBM
has produced has allowed SCO to make such a comparison, and to find tens of thousands of lines
of code from AIX and Dynix that were copied to places in Linux. See SCO's Revised
Supplemental Response to IBM's Interrogatory 1, filed Jan. 12, 2004 [Exh. 4], and Exhibits B
and C to Hatch April Letter [Exh. 5]. This establishes that IBM donated AIX and Dynix source
code to Linux, and supports SCO's breach of contract claim.
In short, the files that IBM produced pursuant to this Court's Order of March 3, 2004,
directly support SCO's defenses to IBM's contentions in this case. But those files are
incomplete, and the proof they offer is, in some cases, inferential. For example, while it is
possible to state that File A from AIX (or Dynix) was the origin of File B in Linux (see attached
code comparison of Dynix v4.6.1 and the patches IBM placed in Linux 2.4.1 [Exh. 6]), it is
nearly impossible to find the original source of File A (the code contribution to Linux) due to
IBM's incomplete production. Only complete production by IBM of AIX and Dynix source
trees will enable SCO to find the additional intermediary versions of code which must exist.
Additionally, complete production by IBM will enable SCO to tell the entire story of
AIX's derivation from UNIX System V. This story began when IBM obtained access to UNIX
System V through the licensing agreements it reached with AT&T in 1984. Also, full disclosure
by IBM will allow SCO to complete the tracking and identification of AIX-based contributions
to Linux. This task includes identification of direct code copying, identification of copying of
sequences and structures, and identification of methods and concepts traced from AIX (including
design documents) into Linux, or otherwise made available to Linux by IBM.
Comprehensive production by IBM will also allow SCO to complete the tracking of
Dynix and ptx-based contributions to Linux. This task includes identification of direct code
copying, identification of copying of sequences and structures, and identification of methods
and concepts traced from Dynix and ptx (including design documents) into Linux, or otherwise made
available to Linux by IBM. Id.
SCO will also be able to complete the following when in possession of the request
Provide a more detailed identification of IBM's improper contributions of AIX code,
methods and concepts to Linux.
Provide a more detailed identification IBM's improper contributions of Dynix/ptx
code, methods and concepts to Linux.
Provide a more detailed identification of IBM's improper use of SCO's copyrighted
information improperly contributed to Linux by others.
Conduct a detailed evaluation of the code similarities between System V ES/MP code
base and all produced versions of Dynix/ptx, which will yield a view of the subset of
System V code that is contained in the Dynix/ptx code.
Conduct a detailed evaluation of the code similarities between System V ES/MP code
base and all AIX/AIX5L code bases, which will yield a view of the subset of recent
SVR4 code recently added to AIX and AIX5L by IBM.
Accordingly, in order for SCO to fully trace the derivation of AIX and Dynix from UNIX
System V, IBM must produce their revision control systems in a format that will allow SCO to
track the derivation and modification of UNIX System V source code in AIX and Dynix. Id.
C. The Burden on IBM is Negligible.
Without access to AIX's and Dynix's revision control systems, SCO will labor at a
substantial disadvantage in this case since it will be unable to adequately investigate and
independently substantiate IBM's improper contributions to Linux (and if necessary, to counter
IBM's copyright noninfringement argument). In sharp contrast to the potential burden to be
shouldered by SCO, IBM's production of all versions of AIX and Dynix and of the complete
revision control systems will be a small matter. As IBM explains in an internal document, all
levels of all files are stored on central servers, and should be easily available for downloading
and production to SCO. See IBM's Configuration Management Version Control (CMVC)
Introduction (1710058191-92) [Exh. 7].
This situation thus closely resembles the one in Dynamic Microprocessor Assoc. v. EKD
Computer Sales, 919 F. Supp. 101 (E.D.N.Y. 1996). One of the claims in that case required the
comparison of two versions of a program called pcAnywhere. The plaintiff, who controlled the
source code, did not want to produce it. The defendant's experts testified as to the prejudice this
refusal caused them:
We need the original source code for Versions 3 and 4 of pcAnywhere in order to
render a professional opinion on the issues posed to us. Without the source, we
can only speculate and the facts cannot be known with any certainty. In addition,
since [the plaintiff] and any experts it engages will obviously have access to the
source code, we would be placed at a considerable and unfair disadvantage in
evaluating any opinions tendered by them.
Id. at 104. Recognizing the defendant's "substantial disadvantage ... if it is unable to
adequately investigate and counter [plaintiff's] claims through its own experts," the court
ordered the plaintiff to produce the source code. Id. at 105.
IBM develops and maintains source code through what are arguably the world's most
sophisticated set of tools for that purpose. The task's complexity is not measured by the number
of lines of code, or the number of files to be produced: producing a million lines of code is not
materially different, in this context, from producing ten lines, or ten million lines. IBM keeps
them neatly organized in servers designated for that purpose. Gathering these materials is, for a
competent engineer, a rather trivial task. Considered alongside the fact that SCO has no other
possible source for this information and requires it for many aspects of this case, IBM's burden
AIX's and Dynix's revision control systems are the histories of those operating systems.
They trace their development in detail. They will reveal IBM's contributions to Linux; they will
show that AIX and Dynix are derivative works of UNIX System V. Thus, SCO respectfully asks
this Court to compel IBM to produce, in a usable format,(10) the revision control systems for AIX
WHEREFORE, SCO requests that pursuant to this Court's March Order, IBM be ordered
to produce all revision control systems for AIX and Dynix, all versions of AIX and Dynix code,
and other revision-related materials; specifically, the items listed in Section II, supra.
DATED this 28th day of May, 2004.
Brent O. Hatch
Mark R. Clements
HATCH, JAMES & DODGE
Robert Silver, Esq.
Stephen N. Zack, Esq.
Mark J. Heise, Esq.
BOIES, SCHILLER & FLEXNER LLP
Counsel for Plaintiff/Counterclaim Defendant
 Section 7.10 of the Agreement [Exh. 1] states:
Except as provided in Section 7.06(b), nothing in this agreement grants to
LICENSEE the right to sell, lease or otherwise transfer or dispose of a
SOFTWARE PRODUCT in whole or in part.
 In the mid-1980s, Dynix had a sister program called ptx. The two were later merged into
Dynix/ptx, sometimes referred to as simply "Dynix" or "PTX".
 IBM has been aware of SCO's position on this point for quite some time. SCO informed
IBM's counsel on April 19, 2004 that without interim versions of the AIX and Dynix/ptx code
and the revision control systems that tracked derivations and modifications, it was difficult if not
impossible for SCO to respond to IBM's theory of its noninfringement of SCO's copyrights. See
Letter from Brent Hatch to Todd Shaughnessy, dated April 19, 2004, ¶ 3 [Exh. 2]. See also
Letter from Mark Heise to David Marriott, dated April 7, 2004 [Exh. 2].
Despite SCO's clear statement of its position that IBM's production remained
incomplete, IBM has tried to use its own failure to respond adequately to SCO's discovery
requests as the basis for a Cross-Motion for Partial Summary Judgment on its Claim for
Declaratory Judgment of Non-Infringement, dated May 18, 2004. IBM should not be allowed to
rely on its own discovery shortcomings in order to obtain partial summary judgment against
SCO. To respond to IBM's charges in its Cross-Motion, SCO must have access to the materials
requested in this memorandum.
 IBM's Opposition to SCO's Motion to Amend the Scheduling Order ("IBM Opp.") at 14-
15 (emphasis original). IBM's Opposition also claims that SCO has engaged in "discovery
misconduct" and is the sole cause of delay in this case. Id. at 2.
 Examples of revision control systems include RCS (revision control system), SCCS
(source code control system), and CVS (concurrent versioning system) files and/or trees with
unexpurgated log information.
 Notably, in addition to needing this information to respond to IBM's allegations that SCO
has engaged in "discovery misconduct," (IBM Opp. at 2), SCO needs this information simply to
defend against IBM's copyright counterclaim (if it remains in this litigation).
 For example, IBM has refused to produce AIX version 5.0, without adequate explanation.
See Letters between Mark Heise and David Marriott/Christopher Kao, dated March 29, April 2,
April 7, and April 14, 2004 [Exh. 2].
 This list contains many items, but as explained in Section III.C, infra, the burden on IBM
to produce these materials is negligible.
 Of course, another way is for IBM simply to admit that AIX and Dynix are derived from
UNIX System V. In this litigation, IBM has declined to do make this admission, thereby
contradicting IBM's own public statements and internal documents. IBM's inconsistent
position need not be fully addressed here, but it will certainly be addressed in this
 IBM's initial production of AIX to SCO was in a non-standard format. It was not until
March 25, 2004 that IBM produced AIX source code in a format SCO could use. ISO-9660
DVDs or CD-ROMs are the current standard interchange format.