UNIX SYSTEM LABORATORIES, INC., Plaintiff, v. BERKELEY SOFTWARE DESIGN, INC., and certain named individuals in their collective capacity as The Regents of the University of California, Defendants.

Civ. No. 92-1667


1993 U.S. Dist. LEXIS 19503; 27 U.S.P.Q.2D (BNA) 1721

March 30, 1993, Decided  
March 30, 1993, Filed


SUBSEQUENT HISTORY: Reported at: 27 U.S.P.Q.2d (BNA) 1721 at 1734.


COUNSEL: For Plaintiff: George L. Graff, Esq., James W. Kennedy, Esq., Paul, Hastings, Janofsky & Walker, NY, NY. Gary F. Werner, Esq., Crummy, Del Deo, Dolan, Griffinger & Vecchione, P.C., Newark, NJ.
For Regents of the University of California, Defendant: Joel Linzner, Esq., Carla J. Shapreau, Esq., Crosby, Heafey, Roach & May, Oakland, CA. James E. Holst, Esq., John F. Lundberg, Esq., Mary E. MacDonald, Esq., University of California, Oakland, CA. By: Frederick B. Polak, Esq., Post, Polak & Goodsel, Roseland, New Jersey. For Berkeley Software Design, Inc., Defendant: Leslie A. Fithian, Esq., Heller, Ehrman, White & McAuliffe, Palo Alto, CA. James H. Forte, Esq., Saiber Schlesinger Satz & Goldstein, Newark, NJ.

JUDGES: Debevoise


Debevoise, District Judge.

Plaintiff Unix System Laboratories, Inc. (sometimes referred to as "USL") instituted this action seeking relief from Defendants' past and prospective distribution of computer software in alleged violation of Plaintiff's proprietary rights in the UNIX operating system. On March 3, 1993, I denied Plaintiff's [*2]  motion for a preliminary injunction against further distribution of certain software by Berkeley Software Design, Inc. ("BSDI"), and I also denied various motions to dismiss by Defendant Regents of the University of California ("the Regents"). Plaintiff now moves for a new hearing pursuant to Rule 59(a), and to alter or amend my March 3 opinion and order pursuant to Rule 52(b) of the Federal Rules of Civil Procedure.


The facts and procedural history of this case are set forth at length in my opinion of March 3, and I shall presume knowledge with that opinion. Unix System Labs., Inc. v. Berkeley Software Design, Inc., No. 92-1667 (D.N.J. March 3, 1993). Plaintiff initially appealed that decision and, several days later, brought its Rule 59(a) and 52(b) motions by order to show cause. However, because of jurisdictional concerns, Plaintiff later moved to withdraw its appeal before docketing in the Court of Appeals pursuant to Appellate Rule 42(a). I granted Plaintiff's motion to withdraw its appeal on March 16, 1993, and this motion followed.

The principal factual issue here is whether "32V source code has now been distributed, without [*3]  notice, to literally thousands of licensees." Slip op. at 29. Plaintiff's first contention is that the record lacks evidence that any 32V source code was distributed without proper notice. Plaintiff does not actually assert that it affixed notice to its distributions of 32V source code, only that Defendants have not proved otherwise.

In fact, the record does contain evidence that AT&T distributed 32V without notice of copyright. In paragraph 5 of the McKusick Affidavit, Dr. McKusick explained: n1
it is my understanding that . . . the AT&T legal staff instructed AT&T programmers to place copyright notices on all the source files that made up the 32V distribution, and then instructed them to remove all the copyright notices. The source of this information is Steven Johnson, who was an AT&T employee involved in inserting and removing the copyright notices.

In addition, attached to the Second Shapreau Declaration is a photocopy of a magnetic tape labeled "UNIX/32V time-sharing system," with notice that "use, duplication or disclosure is subject to restrictions stated in your contract," but without notice of copyright.

- - - - - - - - - - - - - - Footnotes - - - - - - - - - - - - - - -

n1 Dr. McKusick's evidence is hearsay, but a trial court may consider hearsay in deciding a motion for a preliminary injunction. 11 CHARLES A. WRIGHT & ARTHUR R. MILLER, FEDERAL PRACTICE AND PROCEDURE 2949 (1973).

- - - - - - - - - - - - End Footnotes- - - - - - - - - - - - - - [*4] 

Defendants have now supplemented the record with further evidence in the form of Plaintiff's responses to BSDI's interrogatories. For example, in its response to interrogatories Nos. 5 & 6 of BSDI's second set of interrogatories, Plaintiff stated that "AT&T and USL have distributed the UNIX/32V source code without a copyright notice pursuant to the license agreements referenced in response to Interrogatory No. 1," and that "all files contained within the UNIX/32V source code have been distributed by AT&T or USL without a copyright notice." (Shapreau Recon. Decl., Ex. A, responses 5, 6; see also id., Ex. B, response 11.) These answers strongly support my earlier finding that Plaintiff distributed 32V without notice.

Plaintiff's second contention is that, even if AT&T did distribute 32V without notice, it did not distribute 32V without notice to thousands of licensees. Plaintiff argues that most of its many licensees received later versions of the code, complete with notice of copyright, rather than 32V.

It is difficult to know exactly how many licensees did, in fact, receive 32V without notice. Plaintiff states that "it was AT&T's written policy to affix a copyright notice on  [*5]  all of its UNIX source code distributed after July 16, 1984. Moreover, Plaintiff's records reveal that only 65 entities received commercial licenses for 32V prior to January 13, 1987." n2 (Pl.'s Supp'g Br. at 5; Earhart Decl. at PP 1-3.) AT&T distributed an additional 220 or so 32V licenses to educational institutions. The great majority, if not all, of these 32V licenses date from 1982 or earlier. (Shapreau Recon. Decl., Ex. C.)

- - - - - - - - - - - - - - Footnotes - - - - - - - - - - - - - - -

n2 AT&T's policy of affixing copyright notice apparently began about July 16, 1984. AT&T's "Memorandum For Record" of that date, the document that implemented the policy, discusses issues that are characteristic of the initial implementation of a new policy. (Earhart Decl., Ex. A at P 4, 5.)

- - - - - - - - - - - - End Footnotes- - - - - - - - - - - - - -

Defendants insist that there were considerably more than just 65 commercial licensees. AT&T had a source code exchange program that permitted licensees of higher-level UNIX versions, such as System III and System V, to obtain 32V source code from either AT&T itself or from other licensees. (Shapreau Decl.  [*6]  of Jan. 22, 1993, Ex. O at 54-57; id., Ex. L at 92-93; Forte Decl. of Jan. 8, 1993, Ex. C at 45-49; id., Ex. K at 58-62.) Obtaining 32V source code via the exchange program could benefit those with higher-level systems because "some of the functionality from the earlier release [(32V)] may have been dropped off" (Forte Aff., Ex. A at 166), and perhaps because of hardware limitations on unlicensed machines. By December 11, 1986, more than 1400 UNIX licensees were eligible to receive 32V via the exchange program. (Forte Decl. of Jan. 8, 1993, Ex. M; Regents' Amicus Supp'g Br. at 8.) It is not known how many of these licensees actually obtained copies of 32V. n3

- - - - - - - - - - - - - - Footnotes - - - - - - - - - - - - - - -

n3 Defendants also note that AT&T licensed its software products by central processing unit ("CPU"), not by licensee. (Shapreau Decl. of Jan. 8, 1993, Ex. N at 202.) Thus, an individual licensee might be licensed to use its particular UNIX product on hundreds of CPUs. For example, AT&T apparently licensed one university to use UNIX-S V2.OV2 on 872 different CPUs (Forte Decl. of Jan. 8, 1993, Ex. M at 219), and Berkeley itself entered into 78 32V letter agreements with AT&T and used 32V source code on at least 162 CPUs. (True Decl. at P 2.) Overall, Defendants calculate that AT&T directly licensed at least 899 different CPUs to use 32V. (Forte Decl. of Jan. 8, 1993, Ex. M; Mar Aff. at P 2.)

- - - - - - - - - - - - End Footnotes- - - - - - - - - - - - - - [*7] 

Plaintiff adds one further factual contention, that it never intended to market 32V as a commercial product. (Kennedy Recon. Decl. at P 3.a.) I made no findings of fact on this issue in my March 3 decision because, as discussed below, such finding were not necessary for that decision. Accordingly, I will not raise this issue now.

To summarize, I find that my determination that "32V source code has now been distributed, without notice, to literally thousands of licensees" was without adequate basis in the record. Instead, I find that there is substantial evidence that about 65 commercial licensees, about 220 educational licensees, and an unknown number of government licensees received 32V source code without notice of copyright. In addition, as of the end of 1986, more than 1400 licensees were eligible to obtain 32V source code by virtue of licenses to use that or other versions of UNIX.


Plaintiff contends that I have mischaracterized certain material facts relevant to its distribution and licensing of 32V. Consequently, Plaintiff now moves for a new hearing on its preliminary injunction motion pursuant to Rule 59(a), and to amend or make additional factual findings [*8]  pursuant to Rule 52(b). In order to evaluate Plaintiff's motion, it is first necessary to place the findings challenged by Plaintiff into their proper legal context.

Plaintiff has challenged my findings concerning the number of licensees because Plaintiff believes that these findings played a key role in my prior decision denying Plaintiff's motion for a preliminary injunction. A plaintiff ordinarily bears the burden of proof in a motion for a preliminary injunction, but here the burden shifted to Defendant BSDI on one key issue: whether Plaintiff had a valid copyright in 32V. The burden shifted because Plaintiff, having copyrighted 32V in 1992, benefitted from a presumption that the copyright was valid. To challenge this presumption, BSDI sought to prove that AT&T had already published 32V by distributing it without notice of copyright in a general, rather than a "limited," publication. A limited publication is one in which a work is distributed to "a definitely selected group and for a limited purpose, and without the right of diffusion, reproduction, distribution or sale . . . .'" Rexnord, Inc. v. Modern Handling Systems, Inc., 379 F. Supp. 1190, 1197 (D. Del. 1974) [*9]  (quoting White v. Kimmell, 193 F.2d 744, 746-47 (9th Cir. 1952), cert. denied, 343 U.S. 957, 96 L. Ed. 1357, 72 S. Ct. 1052 (1952); see also American Visuals Corp. v. Holland, 239 F.2d 740, 744 (2d Cir. 1956) (quoting White v. Kimmell). In other words, BSDI sought to prove that AT&T either had distributed 32V to a general group, or had distributed it for general purposes.

I examined only the first element of this test in my earlier decision: whether AT&T had published 32V to a limited group. The key measure of a limited group is not the composition of the group, but rather the criteria by which the group is selected. For example, if the selection criteria are sufficiently broad, "'a general publication may be found when only one copy of the work reaches a member of the general public . . . .'" Ford Motor Co. v. Summit Motor Products, Inc., 930 F.2d 277, 299 (3d Cir.), cert. denied, 116 L. Ed. 2d 324, 112 S. Ct. 373 (1991) (quoting Brown v. Tabb, 714 F.2d 1088, 1091 (11th Cir. 1983)); [*10]  accord Burke v. Nat'l Broadcasting Co., 598 F.2d 688 (1st Cir.), cert. denied, 444 U.S. 869, 62 L. Ed. 2d 93, 100 S. Ct. 144 (1979). Similarly, "even if the persons receiving copies in fact constitute only a select group, the publication is nevertheless general if copies were available to persons not included in the group." Brown, 714 F.2d at 1092 (citing 1 NIMMER ON COPYRIGHT 4.13[A]). Thus, the immediate significance of the findings challenged by Plaintiff must be measured by the weight these findings carried in my determination that AT&T's distribution of 32V was not a limited publication.

A. Rule 59(a)

Rule 59(a) states that a "new trial may be granted . . . on all or part of the issues . . . (2) in an action tried without a jury, for any of the reasons for which rehearings have heretofore been granted in suits in equity in the courts of the United States." Courts have generally limited the "reasons" referred to in the rule to "manifest error of law or fact," and "newly discovered evidence." 6A JAMES W. MOORE & JO D. LUCAS, MOORE's FEDERAL PRACTICE [*11]  P 59.07; accord Imprisoned Citizens Union v. Shapp, 461 F. Supp. 522, 523 (E.D. Pa. 1978) (citing MOORE'S FEDERAL PRACTICE).

As discussed above, I erred in concluding that "thousands" of people had received distributions of 32V without notice. The number is probably closer to 65 commercial and 220 educational licensees, although over a thousand licensees may, in fact, be eligible to receive 32V. However, not all errors of fact merit a new hearing under Rule 59(a). If an error of fact does not undermine the grounds of the prior decision, then a new hearing is inappropriate because it would interfere with the "just and speedy determination" of the action. FED. R. CIV. P. 1.

I find that my factual error here does not warrant a new hearing. The decisive factor in determining whether Plaintiff's publication of 32V was limited was not the number of licensees, but rather the criteria by which AT&T selected the licensees. Plaintiff has not moved to amend my findings with regard to AT&T's licensing practices, which are covered at length in deposition testimony submitted by Plaintiff. (Kennedy Reply Decl., Exs. C, D, E.)

This is not to say that the number of licensees [*12]  is completely irrelevant. Had AT&T licensed noticeless 32V to only a few entities, then this might have constituted circumstantial evidence of a stringent selection process. But AT&T did not license 32V to only a few entities n4 -- it licensed 32V to 65 commercial licensees. Although 65 licensees is a relatively small number, and hence is some evidence that AT&T distributed 32V to a limited group, this evidence is not so overwhelming as to rebut the detailed descriptions of AT&T's licensing procedures contained in the record. See Brown, 714 F.2d at 1092 (three purchasers of an advertising jingle "did not constitute a select group because the jingle was available to others"). This evidence is especially unpersuasive because of the large number of entities eligible to receive 32V via their licenses for other AT&T products.

- - - - - - - - - - - - - - Footnotes - - - - - - - - - - - - - - -

n4 In my March 3 opinion, I focussed only on AT&T's 32V commercial licensees because the deposition testimony of record primarily discusses AT&T's commercial licensing practices. (SeeKennedy Reply Decl., Exs. C, D, E.) AT&T also has numerous educational and government licensees, as recorded in Exhibit M to the Forte Declaration and Exhibit C to the Shapreau Reconsideration Declaration. I did not decide whether AT&T's distribution of 32V source to these other licensees was a general or a limited publication.

- - - - - - - - - - - - End Footnotes- - - - - - - - - - - - - - [*13] 

Assuming arguendo that the number of AT&T's commercial 32V licensees is circumstantial evidence for a limited publication, this evidence is nonetheless much less probative than other evidence in the record. The most probative evidence is the deposition testimony submitted by Plaintiff, in which various employees have described the licensing procedures in some detail. As set forth in this testimony, Plaintiff licensed 32V source code to all those who wanted it, could pay for it, reasonably needed it, and would protect it from redistribution. Plaintiff's proposed findings of fact do not change this conclusion.

Plaintiff also contends that AT&T never intended 32V to be a commercial product. AT&T's intent is relevant, if at all, only to whether AT&T distributed 32V for a limited purpose. Burke, 598 F.2d at 692 ("the distinction between 'general' and 'limited' publication depends less on the creator's intention than on his action"); H.W. Wilson v. Nat'l Library Service Co., 402 F. Supp. 456, 458 (S.D.N.Y. 1975) (intent is irrelevant to whether work has entered public domain). Because I never reached the issue of whether [*14]  AT&T distributed 32V for a limited purpose, I need not reach the issue of AT&T's intent.

Accordingly, Plaintiff's motion for a new hearing pursuant to Rule 59(a) is denied. A new hearing would be futile and, as discussed below, is not necessary in order to correct errors of fact in my earlier decision.

B. Rule 52(b)

The final question is whether I should amend the record pursuant to Rule 52(b). Rule 52(b) states that, upon motion, "the court may amend its findings or make additional findings and may amend the judgment accordingly." A motion to amend pursuant Rule 52(b) has a different purpose than a Rule 59(a) motion. "The primary purpose of Rule 52(b) is to enable the appellate court to obtain a correct understanding of the factual issues determined by the trial court as a basis for the conclusions of law and judgment entered thereon." 9 CHARLES A. WRIGHT & ARTHUR R. MILLER, FEDERAL PRACTICE AND PROCEDURE 2582 (1971). Consequently, a court may amend its findings without amending its judgment, although it need not amend its findings to include immaterial matters.

The evidence offered by Plaintiff here, although not so substantial as to warrant a new hearing, is nonetheless [*15]  material to an understanding of the facts of the case. In addition, at least some of it is not exactly new. For example, a person can cull the number of 32V licensees from a 500-page affidavit exhibit already in record. (Forte Decl. of Jan. 8, 1993, Ex. M.) Accordingly, I shall grant Plaintiff's motion to amend my findings.

My preliminary findings of fact shall be amended to conform with the following findings: (1) the record holds substantial evidence that AT&T distributed 32V source code without copyright notice to about 65 commercial and 220 educational licensees; and (2) more than one thousand licensees of other UNIX products were eligible to receive 32V source code under the AT&T source code program.


Plaintiff's motion for a new hearing pursuant to Rule 59(a) is denied. Plaintiff's motion to amend the findings of my prior opinion pursuant to Rule 52(b) is granted. An appropriate order shall issue.

Dickinson R. Debevoise U.S.D.J.
Date: March 30, 1993