The most interesting news is that IBM isn't asking the court to alter its order in regard to production of the AIX and Dynix code. It's going to comply, and it seems to have no problem doing that, although it points out it has never before been asked to do such a thing, and it's a huge undertaking.

I'm guessing that when Judge Kimball said he was inclined to agree with Wells on that point, it may have convinced IBM that it probably wasn't going to be successful in getting them to change their minds, and you have to weigh out the advantages of an appeal as opposed to the expense and effort of filing one, and so I'm thinking they decided to focus on the one issue that they feel it's literally impossible to comply with by the deadline set, in the hopes that the matter can be settled without having to appeal. They are also, however, developing the record for appeal on the one issue they raise in this motion.

Sometimes it's easier to comply with an order than to argue about it, if it's not essential. We now see, by the decision IBM made about what to make an issue of, that IBM doesn't believe that SCO will find a thing in that code, onerous as the task is for IBM to produce it.

IBM also doesn't mind identifying the 3,000 individuals who made the most contributions and changes to AIX and Dynix and to identify the changes they made to AIX and Dynix, insofar as this information is available in CMVC and RCS. And IBM is fine with undertaking a reasonable search for white papers and design documents not found in CMVC or RCS.

The only relief it is asking for is that it asks the Court to reconsider the portion of the Order requiring IBM to search for and produce documents from the files of all those 3,000 individual developers. It says that realistically, it would take more than 6 months to actually do that, and it suggests an alternative: that it produce the massive discovery already agreed to, and then let SCO look through what it gets from IBM and then narrow the focus to those developers it particularly wishes to have discovery about.

In footnote 8, IBM brings to the court's attention two prior cases, where similar discovery was denied, because it was simply duplicative and onerous. It even attaches the cases as exhibits, United States ex rel. Regan v. Medtronic, Inc., where discovery was denied where the "burden ... outweighs the possible relevancy of such evidence ... [in that it requires] Medtronic to contact 5,300 employees and review millions of documents at 70 storage locations over a 20 year period" and Fed Ins. Co. v. Southwestern Wire Cloth, Inc., a case limiting the scope of document requests even where the "requested information was relevant ... [g]iven the potentially thousands of documents implicated." IBM believes it is duplicative and onerous here too, particularly because the CMVC materials already provide the name of the developers who made changes to the code or added code. From that, SCO should be able to figure out which of the 3,000 are the developers whose notes they wish to target.

"Assuming (unrealistically) IBM were able to search for and through the files of 20 people a day," IBM says, "5 days a week, it would require IBM more than 6 months just to search for responsive documents, independent of how long it would take IBM to review the documents for responsiveness, third-party confidential information (and to provide notice to third parties where required), and privilege and prepare them for production." The problem IBM faces is, they told the court how burdensome it would be to turn over the code, and now that the court ordered it anyway, they are doing it, so the court may view their current concerns as something they actually can comply with if ordered to do so, too.

In footnote 4, IBM brings to the court's attention a problem it is having with SCO, who "having been given an inch, . . . seeks to take a mile. SCO has suggested that the Order not only requires IBM to search for and produce documents relating to AIX and Dynix but also requires IBM to detail its Linux contributions." That wasn't in the recent order, IBM points out, and it's publicly available information already, and in the court's March 8, 2004 order, it specifically said IBM wasn't required to produce Linux information that was publicly available. In short, SCO is being unpleasant, and IBM would like the court to know about it and to make clear to SCO the limits of the discovery being required.

SCO will now oppose the motion, and by IBM's including the issue of publicly available Linux contributions in its motion, SCO is now in the position of having to explain to the court why it wants IBM to do its work for it or why it feels it is required. In short, the parties are simply unable to work a thing out between themselves on this point, and so now the court will have to intervene and settle this new dispute.

IBM has made it obvious that they intend to appeal this matter, if necessary. In this motion, they carefully tell the court how they understand the order and what they will be providing. In doing that, they are in effect saying, This is how we understand your order. If you view it differently, please be explicit. That way, they are seeking to head off further disputes about how to interpret the order, and of course they are desirous of limiting the interpretation of the order so as to make it as reasonable as possible to comply with. SCO is now free to say it should be interpreted more broadly. Obviously, they already are, but now they will have to do it publicly and on the record.

IBM also has asked for oral argument on this motion, so they are developing the record for appeal at the same time that they are asking Judge Wells to do the right thing. The two cases IBM has offered the court as exhibits tell the court that IBM believes it will prevail in any such appeal, and it is giving the court an opportunity to make the demands for discovery a little more reasonable itself. This court has been very generous to SCO, way beyond what is usual, IBM seems to be saying, and it now wishes to draw a line in the sand.

******************************

SNELL & WILMER L.L.P.
Alan L. Sullivan (3152)
Todd M. Shaugnessy (6651)
Peter H. Donaldson (9462)
[address]
[phone]
[fax]

CRAVATH, SWAINE & MOORE LLP
Evan R. Chesler (admitted pro hac vice)
David R. Marriott (7572)
[address]
[phone]
[fax]

Attorneys for Defendant/Counterclaim-Plaintiff
International Business Machines Corporation

IN THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF UTAH

THE SCO GROUP, INC., Plaintiff/Counterclaim-Defendant, v. INTERNATIONAL BUSINESS MACHINES CORPORATION, Defendant/Counterclaim-Plaintiff.

IBM'S MOTION FOR RECONSIDERATION OF THE JANUARY 18, 2005 ORDER REGARDING SCO'S RENEWED MOTION TO COMPEL (ORAL ARGUMENT REQUESTED) Civil No. 2:03CV-0294 DAK Honorable Dale A. Kimball Magistrate Judge Brooke C. Wells

1

Pursuant to the Court's Order dated February 1, 2005, defendant/counterclaim-plaintiff International Business Machines Corporation ("IBM") respectfully submits the following request for reconsideration of the Court's Order Regarding SCO's Renewed Motion to Compel, dated January 18, 2005 (the "Order") (attached as Exhibit A hereto).¹ IBM also respectfully requests oral argument on this motion pursuant to DUCivR 7-01(f).

Preliminary Statement

The Order requires among other things, that IBM produce (1) CMVC and RCS data relating to IBM's AIX and Dynix operating systems (Order at 10); and (2) documents from the 3,000 "individuals who made the most contributions and changes to the development of AIX and Dynix" (collectively here, the "Ordered Production") (Id. at 16).

2

application, IBM asks the Court to reconsider the portion of its Order requiring IBM to search for and produce documents from the files of 3,000 individual developers.

Pursuant to the Order, IBM will be producing all of the information in its CMVC and RCS databases relating to IBM's AIX and Dynix operating systems, including source code, the change history of that source code, and programmer's notes and design documents. ³ IBM will also be identifying the individuals who made the most contributions and changes to the development of AIX and Dynix, and the changes they made to AIX and Dynix, insofar as this information is available in CMVC and RCS. IBM will also be undertaking a reasonable search for white papers and design documents not found in CMVC or RCS. Requiring IBM, on top of that, to search for and produce documents from the files of 3,000 developers would impose an extraordinary burden on IBM without any meaningful, offsetting benefit to SCO. Thus, we respectfully request that the Court reconsider this portion of the Order.⁴

3

Argument

As we understand the Order, the purpose of requiring IBM to search for and produce information from the files of 3,000 developers is to provide SCO with "programmer's notes, design documents, white papers, the comments and notes made by those who did the changes, the names and contact information of individuals who made changes and what changes they specifically made". (Order at 15-16.) However, a search and production of this magnitude is not, we respectfully submit, the best way to achieve this purpose and should not be required.

As stated, IBM will provide SCO with an extraordinary quantity of information regarding the development of AIX and Dynix in producing information from CMVC and RCS.⁵ CMVC and RCS include, among other things, changes made to AIX and Dynix, comments about the changes, and information about who made that change and when. CMVC and RCS also include extensive programmer's notes and design documents relating to AIX and Dynix. IBM intends also to produce any white papers and design documents found outside of CMVC and RCS after a reasonable search. Thus, IBM's planned production will provide SCO large quantities of the type of information the Court has ordered IBM to search for in the files of 3,000 developers.⁶

IBM's planned production will not include every scrap of paper that might be found in the files of the 3,000 developers who made the most contributions and changes to the development of AIX and Dynix. However, rather than require IBM to search for and produce

4

documents from 3,000 developers in an effort to capture what might not be included in CMVC and RCS and a reasonable search for design documents and white papers not in CMVC, the Court should defer the production of additional information relating to the development of AIX and Dynix until after SCO has had an opportunity to review the extensive data that IBM will be providing. After reviewing that data, SCO should be able to identify with specificity a reasonable number of developers, if any, from whose files it would like additional production and IBM can then provide the information SCO wants without having to search for, collect and produce redundant and cumulative discovery or discovery in which SCO has no interest. Approaching developer-specific discovery in this way is entirely consistent with the Court's decision early in this case to stage the discovery process.

Moreover, as a practical matter, searching for, reviewing (for responsiveness, third-party confidential information, and privilege) and producing documents from the files of 3,000 people would be a Herculean task. It would be impossible to complete in 6 months, let alone in 60 days, as presently required by the Order.⁷ Assuming (unrealistically) IBM were able to search for and through the files of 20 people a day, 5 days a week, it would require IBM more than 6 months just to search for responsive documents, independent of how long it would take IBM to review the documents for responsiveness, third-party confidential information (and to provide notice to third parties where required), and privilege and prepare them for production. A search for

5

documents from the files of 3,000 developers would yield millions of pages of paper, much of it duplicative or irrelevant.

It is well settled that discovery should not be allowed, even if relevant, where it is unreasonably cumulative or duplicative or where its production would impose an undue burden. See Fed. R. Civ. P. 26(b)(2)(i) (stating that discovery shall be limited by the Court if "the discovery sought is unreasonably cumulative or duplicative, or is obtainable from some other source that is more convenient, less burdensome, or less expensive"); Fed. R. Civ. P. 26(b)(2)(iii) (stating that discovery shall be limited by the Court if it determines the "burden or expense of the proposed discovery outweights its likely benefit, taking into account the needs of the case, the amount in controversy, the parties' resources, the importance of the issues at stake in the litigation, and the importance of the proposed discovery in resolving the issues").⁸

In sum, to require IBM to produce documents from 3,000 developers in 60 days, on top of the discovery IBM is already providing, is to compel the impossible, at enormous expense, with no meaningful advantage to SCO. There are far better ways for SCO to obtain the information it seeks, such as by permitting targeted discovery of individual developers to the extent it makes sense after SCO has had an opportunity to review the information IBM is going to provide.

6

Conclusion

For the foregoing reasons, IBM respectfully requests that the Court reconsider and vacate the portion of the Order requiring the production of documents from 3,000 developers in 60 days, without prejudice to SCO seeking reasonably tailored discovery from the files of individual developers after SCO has reviewed the information IBM is going to produce.

DATED this 11th day of February, 2005

SNELL & WILMER L.L.P.

(signature)

Alan L. Sullivan
Todd M. Shaughnessy
Peter H. Donaldson

CRAVATH, SWAINE & MOORE LLP
Evan R. Chesler
David R. Marriott

Attorneys for Defendant/Counterclaim-
Plaintiff International Business Machines
Corporation

Of Counsel:

INTERNATIONAL BUSINESS MACHINES CORPORATION
Donald J. Rosenberg
Alec S. Berman
[address]
[phone]

Attorneys for Defendant/Counterclaim-Plaintiff
International Business Machines Corporation

7

CERTIFICATE OF SERVICE

I hereby certify that on the 11th day of February, 2005, a true and correct copy of IBM's
MOTION FOR RECONSIDERATION OF THE JANUARY 18, 2005 ORDER
REGARDING SCO'S RENEWED MOTION TO COMPEL was served by U.S. Mail, postage prepaid, on the following:

Brent O. Hatch
Mark F. James
HATCH, JAMES & DODGE, P.C.
[address]
[phone]

Stephen N. Zack
Mark J. Heise
BOIES, SCHILLER & FLEXNER LLP
[address]

Robert Silver
Edward Normand
Sean Eskovitz
BOIES, SCHILLER & FLEXNER LLP
[address]

8

¹ In making this motion, we are mindful of Judge Kimball's Order dated February 8, 2005. Although Judge Kimball indicated that he agrees, in essence, with the Order, we do not understand his February 8 order specifically to address or to foreclose the relief requested in this motion.

² IBM understands this portion of the Order, relating to the production of all material relating to AIX and Dynix found in CMVC or RCS, to require the production of material concerning the AIX and Dynix operating systems themselves, not other material in CMVC that only indirectly related to AIX, such as hardware system designs, firmware, manufacturing-related components, and middleware and other software designed to run on top of the AIX or Dynix operating systems.

³ IBM is making every effort to meet the March 18 deadline for complying with the portions of the Order as to which IBM is not seeking reconsideration. While we are hopeful that IBM will be able to meet the deadline for the data in CMVC and RCS, it is very likely that IBM will need to seek additional time to search for and produce white papers and design documents not in CMVC or RCS. IBM has never before had to do what it has been ordered to do, and preparing the material for production is very complicated and time-consuming.

⁴ We endeavored without success to resolve this matter informally with SCO before bringing this motion. SCO apparently believes that IBM should be required to search for the files of 3,000 AIX and Dynix developers, notwithstanding the size and redundancy of the task. Moreover, having been given an inch, SCO seeks to take a mile. SCO has suggested that the Order not only requires IBM to search for and produce documents relating to AIX and Dynix but also requires IBM to detail its Linux contributions. Not only is that not what the Order provides, but also in an order dated March 3, 2004, the Court expressly ruled that IBM is not required to produce its Linux contributions to SCO insofar as they are publicly available. Nothing about the Order indicates that the Court intended to overrule the March 3, 2004 Order.

⁵ CMVC and RCS identify, by login ID, developers who made changes to the source code; IBM will separately search for and produce, to the extent it can, the names and contact information of the persons whose login IDs appear in the produced CMVC or RCS data.

⁶ CMVC has been used in AIX development since 1991. IBM does not maintain revision control information for AIX source code prior to 1991. IBM continues to search for, but has to date been unable to find, any revision control information for AIX source code prior to 1991. In addition, some early versions of Dynix are available only from archived media, and therefore do not include source control information.

⁷ The magnitude of the task of collecting and producing documents from 3,000 persons becomes clear when contrasted with the document production that has taken place thus far from both IBM and SCO, over the past year and a half of discovery. To date, IBM has produced documents from the files of 174 individuals, while SCO has produced documents from the files of 63 individuals. That is, in a year and a half, SCO and IBM combined have produced responsive documents from the files of 237 persons, less than 10 percent of the 3,000 persons whose documents IBM has been ordered to produce within 60 days.

⁸ See United States ex rel. Regan v. Medtronic, Inc., No. 95-1236, 2000 U.S. Dist. LEXIS 22565, at *11 (D. Kan. July 7, 2000) (denying discovery where the "burden ... outweighs the possible relevancy of such evidence ... [in that it requires] Medtronic to contact 5,300 employees and review millions of documents at 70 storage locations over a 20 year period.") (attached as Exhibit B hereto); Fed Ins. Co. v. Southwestern Wire Cloth, Inc., No. 95-C-689, 1996 U.S. Dist. LEXIS 22001, at *40 (N.D. Okla. Dec. 9, 1996)(limiting the scope of document requests even where the "requested information was relevant ... [g]iven the potentially thousands of documents implicated") (attached as Exhibit C hereto).