FOR THE DISTRICT OF COLUMBIA
May 17, 2000
FOR THE DISTRICT OF COLUMBIA
PLAINTIFFS' REPLY MEMORANDUM
This Reply Memorandum addresses together the various motions and responses relating to
remedies filed by Microsoft on May 10, 2000.(1)
The Court's duty at the remedy stage, which Microsoft does not dispute, is to fashion equitable relief that will (i) end the unlawful conduct, (ii) prevent the defendant from engaging in similar unlawful conduct in the future, and (iii) restore the possibility of competition in the market damaged by the violations. Plaintiffs' Memo in Support at 13. Plaintiffs' proposed remedy serves each of these purposes. The proposed reorganization will leave Microsoft's assets intact but restructure them so as to remove the incentives that Microsoft now has to use those assets in anticompetitive ways to protect its operating system monopoly and will thereby lower the entry barriers Microsoft unlawfully raised, enhance competition, and foment an explosion of new innovation that will benefit consumers. The proposed transitional conduct remedies will restrict Microsoft's ability, until the reorganization takes effect, to further injure competition by committing the same kinds of violations in the future that the Court found it committed in the past. The proposed remedy will undo the harm to competition caused by Microsoft's illegal conduct without creating the costly inefficiencies of more burdensome regulation or, as plaintiffs' experts explained, risking material harm to Microsoft's shareholders.
Microsoft is of course entitled, as it evidently intends, to appeal this Court's conclusions that it violated the Sherman Act by a wide pattern of exclusionary conduct that injured rivals, raised entry barriers, impaired consumer choice, and retarded innovation. But in proposing a remedy to this Court, it was required to make a proposal that addressed the violations the Court found. It did nothing of the sort. Instead, it offered a cosmetic remedy that would have virtually no competitive significance: It would neither undo the harm that Microsoft inflicted on competition nor prevent Microsoft from illegally using its monopoly power to inflict similar harm in the future.
Microsoft attempts to elide the need for structural relief by pretending, contrary to the evidence at trial and this Court's Findings, that its conduct had no effect on competition. For example, the evidence at trial (including Microsoft's own internal documents) showed, and this Court found, that Microsoft's illegal conduct eliminated the serious threat that the browser and Java posed to become a middleware platform that would make applications available to multiple operating systems and thereby erode the applications barrier to entry. There is no doubt that the browser and Java posed a serious threat; certainly, Bill Gates and other Microsoft executives had no doubt at the time. Nor can there be any doubt that Microsoft's conduct eliminated that threat. The evidence also showed, and the Court found, that Microsoft used its control over its applications (including Office) to limit the ability of actual and potential competitors to compete effectively.
What remedy does Microsoft propose to undo the damage to competition caused by its past illegal conduct? Nothing. Moreover, despite months of analyzing potential remedies (including reorganizations similar to that proposed by plaintiffs), and despite the fact that information concerning potential issues with respect to such a reorganization is uniquely within Microsoft's knowledge, Microsoft offers no factual support for any of its objections. Instead, it offers only unsupported speculation as to possible problems -- speculation that is inconsistent with the trial record, with the detailed submissions made with plaintiffs' remedy proposal, and with Microsoft's own prior statements.
Microsoft does not, and cannot, rebut the basic points supporting the proposed spin-off:
(1) The separation of Microsoft's operating systems and applications businesses will undo the artificial preservation and enhancement of the applications barrier to entry caused by Microsoft's illegal conduct;
(2) the separation will reduce Microsoft's incentive and opportunity to manipulate the boundaries between applications and the operating system in order to prevent the development of cross-platform middleware;
(3) the separation will permit each business to compete with the other, both through internal development and through alliances; and
(4) there is no alternative that would restore competitive conditions without significantly greater burdens and potential inefficiencies.
Microsoft does not even pretend that the conduct remedies it proposes would undo the anticompetitive effects the Court found resulted from Microsoft's illegal conduct. Moreover, Microsoft's proposed remedy would leave it free in the future to continue to engage in the very conduct that the Court found unlawful under Section 2 of the Sherman Act, including:
(1) contractually tying two separate products that have no technological integration;
(2) binding together two products without any technological justification and for the express purpose of suppressing competition on the merits;
(3) making predatory expenditures with no expectation of profitability except from the elimination of competition;
(4) retaliating against OEMs for distributing non-Microsoft products or refusing to distribute Microsoft products;
(5) seeking agreement with actual and potential platform competitors to divide markets, or to induce exit from a market entirely, in exchange for access to Windows APIs or other consideration; and
(6) redesigning Windows for the express purpose of making the interconnection of non-Microsoft software "a jolting experience."
Buried in Microsoft's rhetoric are a handful of legitimate questions of interpretation concerning plaintiffs' proposed remedy. We address those questions below. However, the main points about the parties' proposed transitional conduct remedies are these:
(1) Plaintiffs' proposed remedy are directly related to repairing the damage to competition caused by, and preventing continuation of, conduct the Court has expressly found to violate Section 1 and/or Section 2 of the Sherman Act; and
(2) Microsoft's proposed remedy would leave it free to continue the very practices which the evidence at trial showed, and this Court found, to be unlawful and would do nothing to restore competition.
Given Microsoft's failure to come to terms with the Court's rulings, it is now more apparent than ever that structural relief (accompanied by transitional conduct remedies needed to allow that relief to work) is the only remedy that has a chance of ending Microsoft's persistent unwillingness to abandon its widespread use of unlawful practices to maintain and extend its Windows monopoly.
II. Microsoft's Proposed Remedy Is Neither Serious Nor Sensible
As plaintiffs explained in their opening memorandum, an effective Sherman Act remedy must serve three goals: ending the illegal conduct, preventing its recurrence, and healing the competitive harm to the marketplace. See National Society of Professional Engineers v. United States, 435 U.S. 679, 697 (1978); Ford Motor Co. v. United States, 405 U.S. 562, 573 (1972); E.I. DuPont de Nemours & Co. v. United States, 366 U.S. 316, 326 (1961) (cited in Plaintiffs' Memo in Support at 24). Microsoft's proposed remedy does not address the last of these goals at all and falls far short of the first two. Microsoft contends that there was not a proven connection between its illegal acts and its continued market dominance in operating systems and, thus, that the "only appropriate remedy" is a narrow, backward-looking order that would enjoin certain aspects of the illegal conduct in which it engaged in the past. See Summary Rejection Memo at 2-3. But Microsoft mischaracterizes the trial record and this Court's conclusions, and its skeletal proposed remedy does not even adequately address the precise conduct found to be unlawful, let alone restore competitive conditions or prevent similar violations in the future.
Microsoft focuses on the wrong target; it proceeds as if the issue in the case were whether Microsoft had illegally obtained its monopoly in the operating system market. In fact, however, the central violation proven at trial and found by this Court was that Microsoft illegally maintained its monopoly by obliterating middleware threats posed by Netscape Navigator, Java, and Intel's platform software initiatives and thereby increased entry barriers into the operating system market. Microsoft's illegal conduct thus postponed and reduced the likelihood of competition in the operating system market. As the Court concluded, "Microsoft's campaign succeeded in preventing - for several years, and perhaps permanently - Navigator and Java from fulfilling their potential to open the market for Intel-compatible PC operating systems to competition on the merits. Because Microsoft achieved this result through exclusionary acts that lacked procompetitive justification, the Court deems Microsoft's conduct the maintenance of monopoly power by anticompetitive means." Conclusions at 9 (citations omitted). Accordingly, because the Court found "a significant causal connection between the conduct and the . . . maintenance of monopoly power," Microsoft's violations warrant "more extensive equitable relief" than a mere prohibition on a continuation of the precise conduct it employed in the past. III Philip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 653b, at 91-92 (quoted in Summary Rejection Memo at 6-7).
While there can be no precise formula for an effective Sherman Act remedy, the Court has the power and the obligation to require "such orders and decrees as are necessary and appropriate" to give effect to the Sherman Act's prohibitions. Northern Securities Co. v. United States, 193 U.S. 197, 344 (1904). In a case like this one, the Court "has the duty to compel action by the [wrongdoer] that will, so far as practicable, cure the ill effects of the illegal conduct, and assure the public freedom from its continuance." United States v. United States Gypsum Co., 340 U.S. 76, 88 (1950). It is thus clear that simply forbidding the continued violation is not a complete remedy; the court's task "does not end with enjoining continuance of the unlawful restraints" but must also undo the achievements of the defendant and eliminate the possibility of their repetition. United States v. Paramount Pictures, 334 U.S. 131, 171 (1948). As Areeda and Hovenkamp explain, "[u]ltimately, of course, the purpose of the decree is to create a situation in which unrestrained competition can occur." Antitrust Law ¶ 345, at 162.
Microsoft, however, has proposed nothing that would "create a situation in which unrestrained competition can occur." To the contrary, Microsoft does not even address the issues of how to lower the entry barriers that were raised by its illegal conduct or of how to increase the likelihood -- unlawfully diminished by its actions -- of competition in the operating system market. Its proposed remedy is thus, on its face, inadequate as a matter of law and should be rejected for that reason alone.
While Microsoft's proposed remedy makes no pretense of repairing the damage to competition or reversing the increase to entry barriers caused by its conduct, it does purport to end and prevent a recurrence of its unlawful conduct. MS FJ Memo at 3. One might have expected Microsoft to undertake this task by fashioning a remedy that addressed all of its violations and left for determination only an assessment of the efficacy and suitability of the particulars. But here, too, Microsoft's remedy is grossly inadequate on its face.
(1) In the first place, Microsoft's proposal does not address some of the most important violations found by the Court. For example, a central part of the case was Microsoft's tying of Internet Explorer, beginning with IE 1.0, to its monopoly operating systems -- contractual tying in the case of Windows 95 and technological tying in the case of Windows 98. The Court found that that tying violated both Section 1 and Section 2 of the Sherman Act. See Conclusions at 10-11, 25-34; Findings ¶¶ 149-241. Microsoft's proposed remedy, however, is devoid of any prohibition on either contractual or technological tying. It does not even end the ongoing, unlawful tying of the browser to the Windows operating system,(1) much less prohibit Microsoft from illegally maintaining its monopoly in the future by tying other products to Windows in response to new middleware threats. To the contrary, it would expressly authorize Microsoft to tie its browser to Windows whenever any content provider -- including Microsoft itself -- required a Microsoft browser to display any Web page. See MS Proposed Final Judgment § 4(b). Microsoft's remedy ignores the Court's Findings and is, therefore, inadequate on its face.
Microsoft's remedy also fails to address other important elements of the case. These include its repeated efforts to induce Netscape, Intel, Apple, IBM and others to agree to divide markets or otherwise not to compete (Findings ¶¶ 110, 114, 132) and its designing Windows in order to create a "jolting experience" for the users of Netscape's rival browser (Findings ¶ 160).
(2) Moreover, even where Microsoft's proposal does address a violation, it does so in unduly narrow terms. The law is clear that, in crafting a remedy to prevent a "recurrence of the violation," a court is not limited to a "simple proscription against the precise conduct [the defendant] previously pursued." National Society of Professional Engineers, 435 U.S. at 697, 698. When, as the Court has found here, "the purpose to restrain trade appears from a clear violation of the law, it is not necessary that all the untraveled roads to that end be left open and that only the worn one be closed." International Salt Co. v. United States, 332 U.S. 392, 400 (1947). An injunction which simply bars repetition of the precise illegal conduct at issue in the trial would leave the defendant with the "full dividends of [its] monopolistic practices and profit from the unlawful restraints of trade which [it] had inflicted on competitors," Schine Chain Theaters, Inc. v. United States, 334 U.S. 110, 128 (1948). The Court may thus exercise its "broad power to restrain acts which are of the same type or class as the unlawful acts which the court has found to have been committed or whose commission in the future, unless enjoined, may fairly be anticipated from the defendant's conduct in the past." Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 132 (1969). Even otherwise lawful acts may be prohibited if necessary to prevent the resumption of the defendants' anticompetitive tactics. See National Society of Professional Engineers, 435 U.S. at 698; DuPont, 366 U.S. at 327.
Measured by these well-established legal standards, Microsoft's proposal cannot be taken seriously. Its many defects include the following:
(a) Section 4 of Microsoft's proposed remedy purports to rectify illegal restrictions Microsoft imposed upon OEMs (see Findings ¶¶ 231-238), but it falls far short. Section 4(a) would prevent Microsoft from refusing to grant or canceling an OEM's Windows license if it ships or promotes certain non-Microsoft software. Neither that Section nor any other in Microsoft's proposal, however, would limit Microsoft's use of other retaliatory or coercive devices -- such as withholding or threatening to withhold needed technical information or support or other consideration or imposing punitive price increases -- to force OEMs not to cooperate with Microsoft's rivals. The Court found that Microsoft used such devices in the browser war (see Findings ¶¶ 117-118, 120-130, 230), so any serious remedy must broadly prohibit retaliation against and coercion of OEMs.
Section 4(b) purports to address the Court's determination that Microsoft denied OEMs sufficient latitude to modify the user start-up sequence and desktop (see Findings ¶¶ 202-229), but it is inadequate for several reasons: First, except for a provision regarding icons for certain non-Microsoft software, it applies only to browsers and thus does nothing to prevent Microsoft from using OEM restrictions in the future to crush new middleware threats. Second, Section 4(b) applies only to license agreements and claims under the copyright laws and thus does not restrict Microsoft's ability to coerce OEMs by other means, such as ancillary agreements, claims under other laws, or extralegal coercion. Third, Section 4(b) does not prevent Microsoft from continuing to prohibit OEMs from providing alternate user interfaces desired by their customers and necessary to enable cross-platform middleware to reduce the applications barrier to entry.
(b) Section 5 is Microsoft's proposed remedy for the exclusionary ICP, OLS and ISP agreements that the Court found violated Section 2 of the Sherman Act. Conclusions at 15-18. It would prohibit Microsoft from entering into a contract in which Microsoft agrees to promote another party's product or service on the Windows desktop in exchange for that party's agreement to limit the amount of non-Microsoft platform software it distributes. In effect, this provision says that Microsoft cannot use one particular form of payment -- placement on the desktop -- to induce exclusive distribution, but leaves Microsoft free to induce such exclusivity by every other form of consideration (including simple payment from its huge cash reserves). This proposed remedy misses the point: It is the exclusive distribution (and the concomitant limitation on the distribution of competing products) that is anticompetitive, not the form of payment. Moreover, Section 5 has no application to exclusivity that is induced or coerced by Microsoft but not embodied in a "contract." Microsoft's proposal thus fails even to prohibit the range of exclusive agreements the Court found illegal at trial. See Findings ¶¶ 230-234 (Compaq); 287-290, 293-297 (AOL); 317-322, 325-326, 332 (major ICPs); 337-340 (major ISVs); 350-352 (Apple).
(c) Section 6 deals with software developers' access to technical information. Two things are striking about this provision. First, Microsoft defines technical information (Section 1.i.), far too narrowly, to include only information regarding the identification and means to invoke selected APIs. This definition is so narrow that, at the press conference immediately following Microsoft's filing, Microsoft General Counsel Bill Neukom said that the provision would require no change in Microsoft's current practices ("This provision in the decree would essentially reaffirm and, if you will, codify the practice that Microsoft has always used . . . So this is consistent with what we have done.") (http://microsoft.com/presspass/trial/May00/05-10conference.asp.) (Remedy GX 39).
Second, and more important, Microsoft does not propose to create parity among third-party software developers with respect to access to the technical information needed for applications to operate effectively with Windows, much less parity between ISVs and Microsoft's own applications developers. Section 6(a) is little more than a tautology because it would merely enjoin Microsoft from denying a bona fide ISV access to only the technical information that Microsoft chooses to make available "to the software development community at large."
Section 6(b) implicitly admits the emptiness of Section 6(a) because it contemplates selective provision of additional information to selected ISVs. It restricts Microsoft only by requiring that, when Microsoft provides an ISV with access to additional technical information outside the scope of 6(a), it may not condition the provision of that information on the ISV's agreeing not to write applications for Microsoft's competitors. There is nothing in the provision to stop Microsoft simply from withholding the information from ISVs that do write applications for Microsoft's competitors.
These provisions do nothing to remedy the Court's determination that Microsoft engaged in anticompetitive conduct by using the carrot and stick of information disclosure to coerce or punish ISVs into taking or refraining from other kinds of action, nor do they even purport to ensure that Microsoft will not in the future disadvantage rival developers of products that threaten the Windows monopoly by withholding necessary information used by Microsoft's own application developers.
(d) Section 7 would bar Microsoft from conditioning the timely release of a product "that is ready for commercial release" on an agreement by the vendor, on whose product the Microsoft product is designed to run, to limit the development of competing software. Although this provision is apparently directed at Microsoft's threat regarding Office and Apple, it would not even address that problem because the product there was not ready for commercial release (see Findings ¶ 347) and the violation was Microsoft's threat to stop otherwise profitable activity, including further development, for an anticompetitive purpose (see Findings ¶ 355).
(e) Section 8 would require Microsoft to continue to license predecessor operating systems when major new ones are released, but not when important interim releases are made. This requirement is thus likely to be meaningless because, as the evidence at trial showed, Microsoft may make significant changes to interim releases that could coerce customers to migrate to the newest release. See part IV.C., below.(2)
III. Microsoft's Objections To Plaintiffs' Proposed Restructuring Remedy Are Unsound
Microsoft criticizes plaintiffs' proposed restructuring both in its Summary Rejection Memo and in its Summary Response. The former does not suggest any legal bar to the reorganization but instead makes prudential arguments to the effect that the remedy is disproportionate and unnecessary to redress Microsoft's unlawful conduct. It provides no basis for a summary rejection of the proposal but instead repeats the same kinds of equitable arguments set forth in the Summary Response. We therefore address all of Microsoft's arguments together.
The reasons for the proposed restructuring are straightforward: Microsoft injured competition in the operating system market by engaging in a widespread pattern of illegal conduct -- conduct that made no business sense but for the prospect that it would crush nascent middleware threats to its operating system monopoly. The effect of that conduct was to increase entry barriers and thereby reduce the likelihood of competition in the operating system market.
The restructuring will redress the harm to competition because it will put Office and Microsoft's other valuable non-operating-system products in a different company, so that they will be used in the future to benefit consumers and maximize their own profits, rather than strategically, at the sacrifice of profits, to protect the operating system monopoly. This reorganization will offset the harm from Microsoft's illegal campaign by increasing the likelihood of meaningful operating system competition. That is so both because Office and Microsoft's other assets are uniquely valuable operating system complements in their own right and thus could increase the likelihood of effective competition from any operating system (such as Linux) to which they are ported and because they could become middleware platforms and thus replicate the various middleware threats that Microsoft extinguished. In addition, by increasing the likelihood of operating system competition, the restructuring will reduce Microsoft's ability and incentive to engage in the future in the kinds of illegal conduct the Court found in this case.
Microsoft argues in its Summary Response that the reorganization will not increase the likelihood of operating system competition. Its argument is unsubstantiated and mischaracterizes the proposed remedy.
(1) First, Microsoft says that "not one of the government's experts opines that the extreme relief requested by the government will actually result in an immediate increase in competition in the market for 'Intel-compatible PC operating systems'" (Summary Rejection Memo at 14) (emphasis in original) and argues that the restructuring is predicated on "tentative and unsubstantiated speculation" (Summary Response at 10). But the restructuring is intended to redress the harm to competition caused by Microsoft's illegal acts, not to guarantee immediate operating-system competition. If the latter were the objective, the proper remedy would be a reorganization of Microsoft into multiple PC operating system companies, which would directly and immediately compete with one another. Instead, the proposed restructuring will create an increased likelihood of competition in the operating systems market by lowering the applications barrier to entry that Microsoft illegally raised. Instead of forcing immediate entry and competition, the restructuring will enable the market to determine the rate of entry and the form of new competition.
(2) Microsoft argues that Office is not cross-platform middleware that could evolve into a platform competitor to Windows. Summary Response at 9. The argument seems to be that ISVs "rely on Windows," not middleware, "for a full range of system services" and, thus, that Office is not a threat to Windows. Id.
There are three fallacies in this argument: First, it ignores the plain thrust of the Court's determination that an application like Netscape's browser, which itself did not offer the "full range of system services," could present APIs to which ISVs would develop applications and thus serve as middleware which, if ported to competing platforms, could threaten Microsoft's operating-system monopoly. Findings ¶¶ 69-70. Microsoft's argument that Office relies on Windows for some system services, "including basic functions like accessing files on the computer's hard disk" (Summary Response at 9), is little more than an argument that Office, like Navigator before it, is not an operating system. Of course not; that is the point. It is middleware, and middleware "relies on the interfaces provided by the underlying operating system while simultaneously exposing its own APIs to developers." Findings ¶ 28.
Second, Microsoft's argument ignores the role of Office as the kind of "killer application" that Netscape once was (Barksdale, 10/27/98pm at 71:6-11) and that could in its own right enhance the competitive vitality of competing operating systems. The value of Office and Microsoft's other non-operating-system products to competing operating systems will of course increase when they are owned by a firm whose incentives are to maximize their value and induce operating system competition, rather than to protect the Windows monopoly (see Plaintiffs' Memo in Support at 33-36).
Third, Microsoft's unsubstantiated argument ignores testimony in this case about the role of Office and the products in it as a platform. See Devlin, 2/4/99am at 41:21-42:3; Felten, 12/14/98am at 37:10-38:1; Felten, 6/10/99am at 58:22-59:4. Microsoft also ignores its own documents demonstrating that it believes Office to have the platform characteristics that made Netscape a threat to commoditize the operating system, that BackOffice is an important applications development platform, and that it markets BackOffice and Office together as platform for complex applications. There are more than two and a half million developers for Office because of "the benefits of Office programmability," and "that 2.6 million represents over 50 percent of all the professional developers in the world. Developing business solutions on the Office platform makes a lot of sense because Office is where knowledge workers spend most of their working day."(3) BackOffice is already a platform for ISVs, with, as of October 1998, "over 550 ISV applications that carry the Designed for BackOffice logo, ensuring compatibility and performance on the BackOffice platform." Putting Office and BackOffice together lets a developer create a "Digital Nervous System solution built from the Office and BackOffice platform" in four different networked applications areas, according to Microsoft.(4)
(3) Microsoft asserts that "it would not be a simple task to 'port' Microsoft Office to Linux because Linux does not provide system services analogous to those in Windows on which Microsoft Office relies." Summary Response at 9. Of course Linux does not use the same system services as Windows; it is a different operating system from Windows. That is true with the Macintosh as well, yet Microsoft offers a full-featured version of Office for the Macintosh. Findings ¶ 344. Indeed, Microsoft's unsubstantiated assertions about the inadequacies of Linux stand in sharp contrast to the trial testimony of its own expert, Dean Schmalensee ("One of the things, for instance, with Linux, which is, if you think about it, if the Linux platform doesn't have a feature that a particular applications writer would like, the applications writer can put it in, subject to some constraints." (Schmalensee, 6/22/99pm at 59:15-19)); to Microsoft's statements elsewhere that the WordPerfect office suite and Sun's StarOffice are already ported to Linux (Summary Response at 10); and to the internal assessment of one of Microsoft's top executives ("someone's going to do a decent office package for linux. and someone's going to do a decent web based solution. and when they do, watch out." Remedy GX 40 (Silverberg e-mail to Slivka, 2/14/1999).).
(4) Finally, Microsoft asserts that porting Office to Linux, which an independent applications company would have both the incentive and freedom to do, would not be an "instant panacea" that would ensure the success of Linux or any other incipient operating system rival. Summary Response at 10. But Plaintiffs have not suggested that porting Office would guarantee the success of Linux (or any other operating system); it is not the purpose of the remedy or the antitrust laws to guarantee any particular market outcome. The availability of Office will, however, give a unique and powerful boost to the viability of other operating systems (comparable to that which the Netscape browser might have offered but for Microsoft's illegal acts) and thus increase the likelihood of meaningful competition in the operating system market.
Office's "combination of assets" make it uniquely situated to provide such a boost (see Romer Declaration ¶ 27), as the Court implicitly recognized in its Finding regarding the critical importance of Office to the Macintosh. Findings ¶ 344. And if Office is owned by a separate firm that will have every incentive to aid Linux (or any other potential operating system rival to Windows) and keep the new version of Office up-to-date and of high quality, it should do far more to help that operating system compete than it was able to do for the Macintosh while it was owned and controlled by the Windows operating system company.
Most of Microsoft's response consists of arguments that the proposed restructuring will be excessive or inefficient. These arguments are largely wrong and overstated. They provide no good reason for the Court to forgo the procompetitive benefits of the restructuring.
Microsoft repeatedly argues that the Court did not find a causal connection between Microsoft's continuing monopoly power and its many illegal acts (e.g., "the Court did not find" that the "conduct contributed significantly to Microsoft's 'maintenance' of a monopoly." MS FJ Memo at 2). To be sure, the Court did state that "[t]here is insufficient evidence to find that, absent Microsoft's actions, Navigator and Java already would have ignited genuine competition in the operating system market." Findings ¶ 411 (emphasis added). But the Court went on to say, in the very same paragraph, that "[i]t is clear, however, that Microsoft has retarded, and perhaps altogether extinguished, the process by which these two middleware technologies could have facilitated the introduction of competition . . . ." Id. The Court thus concluded that "Microsoft's campaign succeeded in preventing - for several years, and perhaps permanently - Navigator and Java from fulfilling their potential to open the market for Intel-compatible PC operating systems to competition on the merits." Conclusions at 9.
Microsoft's argument is thus fundamentally misfocused. The Court found that Microsoft illegally maintained its monopoly by increasing entry barriers, thereby postponing and reducing the likelihood of meaningful competition in the operating system market. The proposed restructuring, which does not directly end the Windows monopoly but simply makes competition more likely in the future by reducing barriers to entry in the operating system market, is thus appropriately and precisely related to the violation found by the Court. Microsoft's argument that the government has not proved that its monopoly would already have been dissipated but for its illegal conduct is irrelevant.
Microsoft's argument is wrong for a more profound reason as well. Its position in essence is that, because Microsoft was able through its "well-coordinated course of action" (Conclusions at 20) to eliminate the Netscape, Java and Intel threats to its monopoly in their infancy, the Court cannot impose a meaningful remedy. This bold assertion that nipping an incipient competitive threat in the bud deprives the courts of the power to order an effective remedy amounts to an argument for the antitrust equivalent of a license to commit infanticide. In fact, however, a very different inference should be drawn from Microsoft's successful illegal efforts to thwart competitive threats before they take hold -- what is needed is a structural remedy that will reduce the incentives of the owners of Microsoft's assets to use them again in the future for anticompetitive purposes.
In any event, Microsoft's argument is wrong as a matter of law. The courts have long recognized that "[t]he vagaries of the marketplace usually deny us sure knowledge" of what would have happened in the absence of defendant's unlawful conduct. J. Truett Payne Co. v. Chrysler Motors Corp., 451 U.S. 557, 566 (1981). Thus, lest defendants be able to use that uncertainty to avoid appropriate remedies, "[t]he most elementary conceptions of justice and public policy require that the wrongdoer shall bear the risk of the uncertainty which his own wrong has created." Bigelow v. RKO Radio Pictures, 327 U.S. 251, 256 (1946).
Microsoft cites no contrary cases. Instead, it relies on a quotation from the Areeda and Hovencamp treatise. MS FJ Memo at 3. But the quotation is taken out of context. In the cited section, the treatise addresses "[u]nconsummated exclusionary conduct" that "did not in any way impair the vitality, momentum or prospects of the rival." III Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 653b, at 92 (1996). That is obviously inapplicable here in light of the Court's extensive findings about the significant adverse impact of Microsoft's consummated illegal conduct on Netscape, Java, Intel, consumers and innovation.
The relevant section of the Areeda & Hovencamp treatise is the next one (¶ 653c), addressed to "partial causation." This section says that "equitable relief beyond a mere injunction against repetition of the act is generally appropriate" and includes divestiture among the appropriate remedies. Id. at 94-96.(5)
Microsoft argues that the restructuring is "punitive." This is wrong both as a matter of fact and as a matter of law.
As a factual matter, the proposed relief is not punitive for two basic reasons. First, it is the least burdensome means of achieving the requisite objectives of repairing the harm to competition and ensuring that similar antitrust violations do not recur in the future. See Plaintiffs' Memo in Support at 2, 30-31. The remedy will preserve all of Microsoft's assets and enable the management of the resulting firms to use those assets in a manner consistent with the antitrust laws, free of long-term intrusive governmental oversight.
Second, the proposed reorganization is not punitive to shareholders. Microsoft is a publicly-held corporation, and significant assets of such entities are spun-off in a manner proposed by plaintiffs all the time. See Greenhill Declaration ¶¶ 49, 53-55. All but three shareholders will continue to own both entities; not even the three restricted "Covered Shareholders" will lose the value of their shares because they will transfer or convert their shares in one of the new companies to equivalent holdings in the other. See Greenhill Declaration at ¶ 48. The restructuring is unlikely over the intermediate to longer term to have a materially adverse impact on shareholder value and could be positive. See Greenhill Declaration ¶ 88; see also id. ¶¶ 52, 55.
Moreover, as plaintiffs have previously noted, the law is clear that an antitrust remedy must redress the wrong and restore competition. Plaintiffs' Memo in Support at 24-27. Structural relief is appropriate where it is likely to achieve those purposes. Id. at 31-32. An antitrust violator should not be permitted to profit from its wrongdoing, either by enjoying the benefits of an illegally enhanced monopoly or illegally increased entry barriers or by using the uncertainty created by its illegal conduct as an excuse to avoid an effective remedy. Id. at 25.
Microsoft argues that the Supreme Court's decision in United States v. National Lead Co., 332 U.S. 319 (1947), "exemplifie[s]" federal courts' "general reluctance to order divestiture in a Sherman Act case" (Summary Rejection at 10). To the contrary, far from indicating such a "reluctance," National Lead reaffirms the principle that "an understanding of the findings of fact is essential to an appreciation of the reasons" for the proper remedy. 332 U.S. at 338. It also recognizes, regarding "lines of precedent" in the remedy context, that "in this field, such lines cannot be much more than guides. The essential consideration is that the remedy shall be as effective and fair as possible in preventing continued or future violations . . . in the light of the facts of the particular case." Id. at 335. The Court found that, under the particular facts there -- which concerned violations committed primarily through unlawful agreements -- "the cancellation of such agreements and the injunction against the performance of them by the appellant companies" was a sufficient remedy. Id. at 351. The Court emphasized that there was "vigorous and effective competition" between National Lead and its major competitor that needed only to be freed from restrictive agreements. Id. at 352-53. It said nothing about the appropriateness of divestiture under other facts.
Microsoft makes the remarkable argument that it should escape serious remedies for its pervasive pattern of illegal behavior because, it asserts, the illegality of that behavior was "unsettled." Summary Rejection Memo at 3. This argument is specious; the applicable law has been well settled for many years. The prevailing standard for illegal monopolization in this Circuit is clearly set forth in Neumann v. Reinforced Earth Co., 786 F.2d 424, 427 (D.C. Cir. 1986); see Conclusions at 8, 21. And the principles underlying that standard have repeatedly been articulated and applied by the Supreme Court and other courts. See Plaintiffs' Proposed Conclusions of Law at 15-19.
Only the legality of Microsoft's binding or hard-wiring of the browser and the operating system under Section 1 of the Sherman Act could be described as involving unsettled law. The uncertainty about that issue was manifest when the Court of Appeals suggested in dicta in the consent decree case, after Microsoft released Windows 98 with its hard-wired browser (see Findings ¶¶ 160-161, 164-171), that rules of special deference might be appropriate in matters of product design. United States v. Microsoft Corp., 147 F.3d 935, 948 (D.C. Cir. 1998). But that issue has little bearing on the proposed reorganization. First, the product design issue applies only to the narrow matter of Microsoft's hard-wiring of its browser to Windows 98 and has nothing to do with the other aspects of Microsoft's broad, anticompetitive campaign that form the basis for the Court's conclusion that Microsoft violated Section 2 of the Sherman Act. Second, the Court of Appeals' statements on hard-wiring concerned only an analysis of tying under Section 1, which raises very different issues from those under Section 2.
It is hardly surprising that, as Microsoft notes (Summary Rejection Memo at 16-17), there are no cases involving the specific type of corporate reorganization proposed by plaintiffs here under the precise competitive conditions that are present in this case. But Microsoft's efforts to suggest that the proposed remedy is unprecedented is wrong, and its efforts to distinguish prior divestiture cases are misplaced.
Microsoft contends that the Supreme Court has approved divestiture as a remedy for monopolization only in cases involving mergers. See Summary Rejection Memo at 17-19. To the contrary, the Supreme Court made clear by the United Shoe case that divestiture can be an appropriate remedy in a Section 2 case that did not involve mergers. See Plaintiffs' Memo in Support at 32. And it is not the case, as Microsoft's appears to suggest, that the Court has rejected divestiture generally as an appropriate remedy in monopolization or nonmerger cases.
Moreover, while Microsoft tries to distinguish some of the leading divestiture cases on the formalistic ground that they included, among other things, mergers, the defendants in those cases certainly did not view their circumstances as being different from Microsoft's. Instead, like Microsoft, they insisted that their companies were indivisible units that could not sensibly be reorganized. Thus, for example, Standard Oil argued in 1909 that "[t]he inherent vice of this decree is that it seeks to create an artificial division which never existed before; it does not seek to compel members who were formerly independent to resume that independence, but it seeks to compel different subcompanies, which have never been independent, which have never been more than mere agencies created for certain purposes, to sever their allegiance with the principal, and to stand apart, independent and hostile to that principal and to each other." Brief on the Law on Part of Appellants, Standard Oil Co. v. United States, No. 725, Oct. Term, 1909, at 120-121. Further foreshadowing Microsoft's position, Standard Oil also said of the proposed units of divestiture: "All of these . . . are naturally a part of one whole -- all operated together and to and with each other -- all are useful to the other, and to be so useful must have a connection with one or more of the others. . . . There are many parts, but each part has its place, and if a part is taken out, the whole structure is disintegrated." Id. at 284. And, as Microsoft does, Standard Oil insisted that the remedy, "if carried to its logical conclusions attacks the very foundations of the modern business world." Id. at 127.
In any event, the Supreme Court has made clear that the appropriate remedy in a monopolization case depends on the particular facts of the case. See, e.g., Ford Motor Co. v. United States, 405 U.S. 562, 573 (1972) (the means a court uses to restore the competition eliminated by illegal conduct depend on "the special needs of the individual case" (internal quotations omitted)); United States v. National Lead Co., 332 U.S. 319, 335 (1947) (remedy evaluated "in the light of the facts of the particular case"). The cases Microsoft cites reflect nothing more than the application of this principle.
Microsoft's elaborate efforts to distinguish United States v. Paramount Pictures, Inc., 85 F. Supp. 881 (S.D.N.Y. 1949), aff'd sub nom Loew's, Inc. v. United States 339 U.S. 974 (1950), and United States v. United Shoe Machinery Corp., 110 F. Supp. 295, 398 (D. Mass. 1953), aff'd 347 U.S. 521 (1954) (Summary Rejection at 19-20), are unpersuasive. Microsoft notes that Paramount involved "concerted action" among five competitors, as opposed to its "purely unilateral" illegal actions in this case. But Microsoft does not explain why this distinction should make a structural remedy less appropriate to correct the particular competitive problems found by the Court here. In fact, given Microsoft's broad monopoly power, the distorted set of competitive incentives and opportunities created by that power, and its ability "unilaterally" to engage in the anticompetitive conduct found by the Court, a structural remedy is particularly appropriate and necessary to "effectively pry open to competition" the market. International Salt, 332 U.S. at 401.
Microsoft seeks to distinguish United Shoe by quoting at length the District Court's initial misgivings about the need for and practicalities of dividing up a single unitary company, "with one plant, . . . one set of jigs and tools, one foundry, one laboratory for machinery problems, one managerial staff, and one labor force . . ." (Summary Rejection at 20-21, quoting 110 F. Supp. at 348). But the real lesson of the case is that the conduct remedies were not effective and the government thus had to return to court seeking to divide the company up. That the defendant ultimately agreed to the breakup indicates, not an important distinguishing fact from this case, but instead that the original rhetoric about the impossibility of breakup was unfounded.
Finally, there is the ATT case -- a monopolization case that did not involve mergers and culminated in a divestiture. Microsoft attempts to distinguish ATT on the ground that the divestiture was ordered by a consent decree, but that fact undermines neither the appropriateness of the remedy nor the Court's authority to order it.
Indeed, Microsoft draws precisely the wrong lesson from the breakup of the Bell System. Contrary to Microsoft's assertions, dealing with the "unitary" and "integrated" nature of the pre-divestiture Bell System was a crucial concern in the Court's review of both the proposed AT&T Modification of Final Judgment ("MFJ") and the details of the AT&T Plan of Reorganization ("POR"). The court addressed key issues identical to those raised by Microsoft -- how to deal with the defendant's corporate-wide research and development and its intellectual property and concluded that ATT's corporate-wide research efforts should not preclude its reorganization into separated businesses. As Judge Green noted in approving the MFJ, "[c]onsiderable evidence was adduced during the AT&T trial concerning the central role of Bell Laboratories . . . in innovation in the telecommunications industry and, more broadly, in industrial research," United States v. American Tel. and Tel. Co., 552 F. Supp. 131, 167 (D.D.C. 1982) aff'd. sub nom. Maryland v. United States, 460 U.S. 1001 (1983). But, while the Court permitted AT&T technically to retain Bell Laboratories, the POR authorized the newly-created Regional Bell Operating Companies ("RBOCs") to establish an 8,800-employee "Central Staff Organization" (later known as Bellcore (see United States v. Western Elec. Co., 675 F. Supp. 655, 666, n.50 (D.D.C. 1987)), whose employees would be, inter alia, transferred from AT&T, Bell Laboratories and other AT&T affiliates upon divestiture, to replace support previously received from Bell Laboratories and other parts of AT&T. United States v. Western Elec. Co., 569 F. Supp. 1057, 1113-1118, 1114 n.247 (D.D.C.), aff'd. sub nom. California v. United States, 464 U.S. 1013 (1983). Similarly, in addressing AT&T's intellectual property, the court expressly required in the POR that the RBOCs be granted compulsory, royalty free, sublicensable licenses to all AT&T patents, including those currently in force or to be issues within five years after divestiture. Id. at 1082-1091.(6)
A crucial function of Bellcore was to ensure technological interoperability so that "telecommunications [will] continue to interoperate in an engineering sense as one national network and to prevent "balkanized regional networks" and "fragmentation" to the "detriment of all users." Id. at 1118. The Bellcore example thus demonstrates that the reorganization proposed here can be implemented in a way that will avoid the specter of balkanization and fragmentation suggested by Microsoft. Summary Rejection Memo at 16.
Microsoft argues that it is a unified company that cannot readily be divided into "applications" and "operating systems" components, that the development of operating systems and applications can be achieved only by a single company, and that a break-up would cause harmful departures of employees. None of these assertions is factually supported; and Microsoft has, in fact, publicly disclaimed each of them in the past.
(a) Microsoft argues that the restructuring will cause a reduction in its ability to innovate because it will undo the vertically integrated structure that it asserts is critical to its innovation. Summary Response at 11-17. This argument is inconsistent with the facts in several respects.
In the first place, Microsoft's argument that there are unique synergies between the operating system and applications businesses is inconsistent with its prior public statements. Microsoft has repeatedly said that it makes available to third-party applications developers all of the Windows technical information that those developers need to develop and release in a timely manner applications that run well on Windows. See, e.g., CNET article, 3/31/99, at http://news.cnet.com/news/0-1003-200-340558.html?tag=st.ne.1002 (quoting Microsoft Spokesman Mark Murray: "Microsoft has said in the past it publishes all of its APIs and makes extensive technical information on its products available freely to every developer. We have a staff of 2,000 people to work with developers to provide them information they need to write Windows applications.") (Remedy GX 56). If third-party applications developers can now obtain from Microsoft the Windows technical information they need to develop Windows-based applications, there is no good reason why, after the reorganization, applications company developers cannot similarly obtain such information from the operating systems company.
Second, Microsoft has long asserted that it consults intensively with third-party applications developers about new features that may be added to Windows and that it often adds features at their request; in effect, Microsoft has repeatedly said, it can effectively collaborate across corporate boundaries. See, e.g., Schmalensee, 6/22/99pm, at 59:19-60:1 ("Microsoft . . . it talks to developers about what features they would like in new versions. Other applications, operating system and platform vendors proceed the same way. They involve developers, they invest in information, they involve them in the design process, and work to improve the functionality of the platform."); Microsoft website, http://www.microsoft.com/TechNet/winnt/ Winntas/prodfact/NTBOOST.asp ("Microsoft devotes far greater resources than any other vendor to supporting independent developers who write applications for its platforms, and solicits developer input years before new operating systems are released.") (Remedy GX 41).(7) In this respect, Microsoft is like other operating system vendors. See, e.g., Allchin, 2/1/99am, at 61:3-8; DX 2161 ("the Caldera operating system provides a built-in web browser for viewing both local hard disk and internet web sites, a series of common functions across both local hard disk resources and internet web sites, and the use of the web browser to enhance other operating system functions"); Allchin 2/1/99pm, at 72:18-73:23 (the "built-in" browser is designed by KDE, a German firm not affiliated with Caldera).
Third, Microsoft has long been organized, and reorganized, into separate operating systems and applications divisions, with little dependency between them. The story Microsoft now tells of dependency between the operating systems and applications is, at the very least, greatly exaggerated. Dean Schmalensee's written direct testimony at trial cited approvingly two book-length studies of Microsoft's software development practices and internal organization, Cusumano, Michael A. and Richard W. Selby, Microsoft Secrets, Harper Collins Publishers: London, 1996 (hereafter "C&S"), and Stross, Randall E., The Microsoft Way, Addison-Wesley: Massachusetts, 1996 (hereafter "Stross"). Schmalensee Direct ¶ 195. Both studies had extraordinary access to Microsoft documents, managers, and engineers. Both quote Microsoft documents and people at length. Both speak highly of Microsoft's capabilities in developing and marketing software. Both are focused on the company's internal structure as it relates to the development and marketing of new products. And both repeatedly and convincingly belie the story of internal structure now asserted by Microsoft:
Indeed, far from valuing integration into a single business, Microsoft's own officials have increasingly come to realize -- outside the litigation -- that it needs to create separate business units. See Remedy GX 40 (Silverberg e-mail) ("-- since the goal is to have a completely reinvented version of windows, it's clear the team has to be completely separate and independent from windows. because the windows team's goal is to make new thing completely unnecessary . . . . really needs to be a separate company within the company.") (emphasis added); id. ("steve needs to do something so that the company ends up with an org that essentially is a separate company within the company. it has to be free to do what it thinks best. it has to be so that its energy can be 95% focused externally. rather than 80% internally, as is the case today. the company is so wrapped up in its shorts that it can't get anything done. it has an incredible amount of iq yet is getting only pennies on the dollar -- so much iq is wasted. the best example i can think of of what can happen when people are motivated and externally focused is ie3. . . . so much done is so little time by so few people.") (emphasis added).
To be sure, there will be short-term dislocations caused by the reorganization. But such dislocations are not unfamiliar to Microsoft. As Bill Gates has stated, "[r]eorganizations are expected around Microsoft. But that doesn't mean they don't create anxiety. They do, for almost everyone affected--including me. My concern is whether or not we're making the right decisions, and whether key employees will be enthusiastic about their new roles. I gain confidence about a potential reorganization when I see that it makes clear what every group is supposed to do, minimizes the dependencies and overlap between groups, and offers developing employees larger responsibilities." Remedy GX 43. (Gates column, 2/29/96; http://www.microsoft.com/BillGates/ columns/1996essay/essay960202.htm) (emphasis supplied).
Admittedly, some forms of communication are probably easier within a single firm than across firm lines. But those benefits are modest, especially in this era of endemic collaboration across corporate lines. And they are far outweighed by the increased likelihood that the resulting businesses will collaborate with a wider range of other firms, without being inhibited by the incentive to pass up profitable opportunities in order to protect the Windows monopoly.(8) In fact, as Microsoft acknowledges elsewhere in its papers, "[t]he history of innovation in the computer industry is replete with examples of companies coordinating their development efforts, each focusing on its core competencies, even when that meant that each 'refrained' from doing some things they would have done if proceeding alone." Summary Response at 53.
Indeed, Microsoft's argument not only overstates the benefits of integration, but also understates the benefits of the restructuring for innovation by the new companies. After the restructuring, the new companies will have the incentive and ability to collaborate with innovative entrants. Romer Declaration ¶ 34. The applications company will no longer have the incentive to prevent innovations that threaten the applications barrier to entry and will, instead, have a strong incentive to collaborate with innovative entrants into the PC operating system market or with new alternatives to the PC operating system monopoly, such as applications-hosting technologies. Henderson Declaration ¶¶ 100-101. Similarly, the operating systems company will lack many of the tools used to disadvantage rivals' innovative efforts either directly or by rewarding or punishing potential collaborators with the rival and will have added incentive to innovate alone or by collaboration with others. Romer Declaration ¶¶ 30-31.
(b) In any event, Microsoft's exclusive focus on its own innovation is misplaced. It is axiomatic that the antitrust laws are intended to protect, not individual competitors, but competition in the market as a whole. Brown Shoe v. United States, 370 U.S. 294, 320 (1962); Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447, 458 (1993). Thus, even if restructuring causes some short-term dislocation at Microsoft, that would not be a sufficient answer to a remedy that promises to reverse the stifling effect of Microsoft's illegal conduct on innovation by others (see Findings ¶ 412) and thereby to increase innovation in the market as a whole.
(c) Microsoft also raises the specter of a hemorrhaging of valuable employees because of the breakup, arguing that "people will not work where they do not want to work, and they will not be productive if forced to work in conditions they find unsatisfactory." Summary Rejection Memo at 24; see also Summary Response at 18-19. But Microsoft is now organized into small teams that are configured into larger technology and product teams. The proposed reorganization will keep the product teams intact.
Each of the successor companies will be among the largest and richest in the software industry. As Microsoft's Chief Financial Officer recently explained: "When you think about Microsoft and our heritage, we've really had two enormously successful franchises. The Windows franchise and the knowledge worker franchise. . . . Microsoft Office."(9) Each of those "franchises" will be kept intact under the reorganization and will have one of Microsoft's "most profitable products". Findings ¶ 115. Apart from its vague rhetoric about "uncertainty and chaos" (Summary Rejection at 24), Microsoft suggests no reason why its employees will choose not to continue to work in those franchises.
The "uncertainty and chaos" of a reorganization are not a sufficient basis for rejecting the restructuring for another reason as well: Microsoft and its employees have repeatedly proven that they can successfully reorganize. Bill Gates said this in 1996:
About every two years in its 20-year history, Microsoft has undertaken a major reorganization. We changed the structure of the company at the beginning of 1994 and again in February of 1996. I'm sure we'll change it again many times. Reorganizations are expected around Microsoft.(10)
Microsoft reorganizations have continued apace since then; as Brad Silverberg put it in his 1999 email to Ben Slivka, "there are three things you can count on in life at msft: death, taxes, and another reorg." Remedy GX 40 (Silverberg email, emphasis added).
(d) Microsoft's final plaint is that new products now in the pipeline will be able to be developed only if the company stays in its current form. It asserts generally that "Microsoft needs all of its existing resources (and then some) to solve the hard problems it now faces, including the need once again to transform its entire product line to keep pace with dramatic changes in the Software industry . . . If Microsoft is deprived of those resources by being broken in two, innovation will be slowed and consumers will suffer." Summary Response at 8.
That is just empty rhetoric. The reorganization will not deprive individual development projects of needed resources because the resources will go with the projects.
In an effort to bolster its rhetoric, Microsoft argues more specifically that it needs to be an integrated company in order to respond to "to the latest paradigm shift in the computer industry, namely, the transformation of software from standalone products . . . that run locally on a personal computer to Web-based services . . . that perform similar functions but reside on a server and are accessible from many different kinds of devices. Microsoft is seeking to provide a range of new system services to Software developers, now called Next Generation Windows Services [NGWS], that will enable them to develop Web-based applications that are accessible from a wide range of devices." Summary Response at 17. But there is no basis to think that only an operating system monopoly can develop Web-based services. To the contrary, the technical vision behind NGWS, that of server-based computing, is one that this Court found may eventually develop into a competitive threat to the Windows monopoly (Findings ¶ 27) -- one that Microsoft has a unique incentive and ability to delay or distort.(11)
6. Microsoft's Concerns About The Ancillary Restrictions On Collaboration Between The Applications Company And The Operating Systems Company Are Unfounded
Microsoft makes a few rather ill-defined objections to the post-reorganization constraints on interactions between the applications company and the operating systems company set forth in Section 2.b. of the proposed Final Judgment. None is well-founded.
(a) Microsoft contends that placing the development tools in the applications company and the operating system in the operating systems company, combined with "the severe restrictions on cooperation between the two companies," will slow innovation by making it more difficult to develop Windows applications. Summary Response at 17; see also id. at 3. Relatedly, Microsoft asserts that, in the past, technologies developed by applications developers have been included in and distributed with the operating system (id. at 17-20), that some future products will similarly draw on technologies developed by both groups (id. at 3, referring to the Tablet PC, the E-book, the Pocket PC, and the X-box game console) and that, unless operating system and tools developers remain in the same firm, it will be impossible to develop Next Generation Windows Services (id. at 17-18).
Microsoft fails, however, to identify the "severe restrictions" that will prevent legitimate cooperation in the future. The reason is that the proposed Final Judgment contains no such restrictions. To the contrary, the applications company and the operating systems company will be free to work cooperatively in genuine efforts to develop new software technologies. Developers in the two firms will be able to communicate and cooperate with each other, just as Microsoft communicates and cooperates with third-party developers today.
Section 2.b. is intended to bar the successor firms from actions that would have the effect of re-integrating them into a de facto single firm, thereby undermining the reorganization, and to prevent the firms from colluding, rather than competing. Section 2.b.i., addresses the former concern by banning a direct re-integration by merger or joint venture; this is a standard restriction in reorganizations.
The other provisions of Section 2.b. ban other arrangements which could have the same anticompetitive impact on the incentives of the applications company and the operating systems company. Thus, under Section 2.b.ii., neither firm will be able to act as the sales, licensing or distribution agent for the other's products or services because that kind of marketing collaboration could resurrect the incentives of the applications company to protect the operating systems company's monopoly. That is, of course, not an onerous restriction because both firms are well poised to achieve whatever distribution, sales or licensing they need acting independently.
Section 2.b.ii., will not prevent either firm from licensing technologies (as opposed to products) from the other on a nondiscriminatory basis. Thus, it would not have prevented the inclusion in Windows of the long list of technologies, such as the Clipboard, the Toolbar, and OLE/COM, that Microsoft now claims would have been lost to society had the reorganization been undertaken in the past. To be sure, under Section 2.b.iv., the applications company could not license these technologies on terms more favorable than those available to other similarly situated firms, but this antidiscrimination provision does no more than ensure the obviously procompetitive goal of preventing the applications company simply from becoming a de facto co-owner of the Windows operating system.
Section 2.b.iii., which mandates nondiscrimination in technical discussions, will similarly prevent the firms from treating each other as preferred partners rather than competitive rivals. But it will permit the firms to have legitimate technical discussions for procompetitive purposes, whether to ensure that the applications company's tools, middleware, or applications work with Windows operating system or to ensure that Windows developers take into account input from the applications company's developers. Microsoft has long asserted that it is very effective at supporting and cooperating with ISVs outside its own firm, including applications, tools, and middleware vendors, and will under this provision need do no more than continue this practice in a nondiscriminatory way.
(b) Microsoft argues that the reorganization will also force the operating systems company to create inconsistent versions of the technologies now distributed with the operating system but allocated to the applications company because, Microsoft says, Section 2.b.ii., will prevent the applications company from licensing these technologies back to the operating systems company for distribution with its products. Summary Response at 22. This argument makes a cluster of related errors. First, as explained above, it misconstrues and overstates the restrictions on technical coordination between the two companies. Moreover, if the technology is in a separate product provided by the applications company, it can be offered for distribution with the applications company's other products (including Office), and it can be distributed through ISVs, OEMs and IAPs. Surely Microsoft cannot now argue that bundling with the Windows operating system is essential to effective software development. In addition, the restriction should encourage both the applications company to cooperate with other operating systems -- thereby increasing the likelihood of meaningful operating system competition -- and the operating systems company to develop its own, competing, technologies -- thereby encouraging competition in these categories of software as well. The result will be that the technologies will be chosen by the market, instead of being forced upon it by bundling with a monopoly operating system.
(c) Finally, Microsoft says that plaintiffs' proposed remedy would prevent the operating systems company from enhancing or improving Internet Explorer functionality, even for purposes of improving purely operating system functions, such as the user interface or Windows Help (see Summary Response at 3, 16). But this argument overlooks Section 1.c.ii. of the proposed Final Judgment, which makes clear that the operating systems company, will receive a perpetual license to the Internet Explorer intellectual property and is prohibited only from developing or distributing modified or derivative versions of "the Internet browser," not merely non-browsing functionality related to it.
Not surprisingly, Microsoft prefers its skeletal and short-term remedy to the transitional conduct remedies proposed by plaintiffs. But its objections to plaintiffs' proposal are unfounded and based almost entirely on tortured readings or mischaracterizations of the language of the proposed remedy.
Section 3.a of Plaintiffs' proposed remedy will prohibit Microsoft from again misusing its many tools, including license terms, access to information and marketing support, and control of the Windows desktop, that the Court found it wielded against OEMs as part of its pattern of conduct in violation of Section 2 of the Sherman Act. See, e.g., Conclusions at 11-15, 20-21, 24-25. Microsoft protests that this Section will destroy the value of the Windows trademark, allow OEMs to disassemble Windows and substitute third-party software for its components, and fragment the Windows platform. Summary Response at 24-30. These dire predictions are groundless.
(1) Ban on Adverse Actions for Supporting Competing Products (§ 3.a.i.). This provision will prohibit Microsoft from discriminating against any OEM for supporting a competing product or service or exercising its rights under the Final Judgment and will thus prevent Microsoft from continuing to use its monopoly power to harm competition by forcing OEMs to exclude and impede competitors. See, e.g., Findings ¶¶ 64, 115-132, 139, 175-177, 191-192, 203-208, 230-241. Contrary to Microsoft's assertion (Summary Response at 24-26), this provision will not prohibit Microsoft from taking adverse actions against OEMs for legitimate reasons such as software piracy. To the contrary, by its terms the prohibition will apply only to Microsoft actions that are based in whole or in part on an OEM's support of non-Microsoft products.
(2) Uniform Terms (§ 3.a.ii.). This provision will require Microsoft to adopt and employ a uniform license and schedule of royalties for only twenty Covered OEMs for Windows Operating System Products, including (if Microsoft so desires) volume discounts. Its purpose is to prevent Microsoft from discriminating among those OEMs and using differences, or threats of differences, in licensing terms, technical information, and other valuable benefits to force OEMs to disadvantage Microsoft's competitors. Although Microsoft objects that the provision "would require Microsoft to treat every one of the top twenty OEMs exactly the same way" (Summary Response at 28), the uniformity requirement is necessary to prevent Microsoft from employing the myriad forms of coercion and reward that the Court found it used to injure competition. (see, e.g., Findings ¶¶ 64, 115-132, 139, 175-177, 191-192, 203-208, 230- 241). Such coercion is difficult to detect, and the mere threat of its use may be sufficient to accomplish the desired, anticompetitive result. The limitation to the top twenty "Covered OEMs" is intended to ensure equal treatment for similarly situated OEMs, and the requirement that the volume discount schedule be "reasonable" is intended to ensure that Microsoft does not manipulate the discount schedule to achieve a purpose prohibited by the Final Judgment (by, for example, carefully crafting the purported volume discounts to reward or punish selected OEMs for their treatment of Microsoft's rivals).
Section 3.a.ii., will not require, as Microsoft suggests (Summary Response at 26-27), unlimited licensing of source code to OEMs. It will apply instead to licenses in the "form in which Microsoft distributes its Windows Operating Systems for Personal Computers." See Section 7.dd. Only if Microsoft chooses to license source code, rather than binary code, to an OEM will it have any obligation to make source code available to other Covered OEMs on a non-discriminatory basis.
Microsoft argues that the provision would also eliminate joint development efforts. It bases this argument on strained speculation about the government's enforcement intentions and the Court's response. Summary Response at 27. In fact, however, Section 3.a.ii., expressly provides that the uniformity requirement regarding technical and other information "shall not apply to any bona fide joint development effort by Microsoft and a covered OEM with respect to confidential matters within the scope of that effort." Moreover, the decree will not prohibit Microsoft from paying OEMs for technology development unrelated to Windows, so long as those payments are not structured so as to amount to a de facto change in the price of Windows.
Microsoft's assertion that the remedy would require renegotiation of all its contracts with OEMs (Summary Response at 28) is wrong. Section 3.a.ii., will simply prohibit Microsoft from enforcing "any provision in any Agreement with a Covered OEM that is inconsistent with this Final Judgment."
(3) OEM Flexibility in Product Configuration (§ 3.a.iii.). This provision will prevent Microsoft from restricting OEMs' ability to customize their PCS in certain ways to promote non-Microsoft software. Microsoft responds by invoking the same unfounded parade of horribles -- infringement of its copyright, reduction in the value of Windows, consumer confusion, and balkanization of the platform -- that it offered at trial and the Court rejected. Compare Summary Response at 28-31 with Findings ¶¶ 209-227, 410 (by restricting OEM flexibility, "Microsoft foreclosed an opportunity for OEMs to make Windows PC systems less confusing and more user friendly, as consumers desired"); see Conclusions at 12-14.
Microsoft's bleak forecasts rest on the premise that OEMs will act irrationally and thrust "Frankenstein's monster" (Summary Response at 30) on their customers. OEMs offer their computers into a highly competitive market (Findings ¶¶ 193, 222); and they are likely to make only those changes that reflect consumer demand and increase consumer satisfaction, lest they quickly feel the pain of lost sales and increased support costs. Findings ¶¶ 222, 410. Indeed, most of Section 3.a.iii., expressly provides that OEMs may do only what Microsoft has permitted users (e.g., invoke an alternate interface, change the default browser, uninstall a feature or an icon) (Findings ¶¶ 171, 222) or selected OEMs to do (insert their own splash screen screens, registration or Internet sign-up wizards, etc.). Findings ¶¶ 219, 223.
Microsoft's overreaching is illustrated by its assertion that Section 3.a.iii.(4), which will permit OEMs to configure Windows to "launch automatically any non-Microsoft Middleware, Operating System, or application," will "inevitably lead to balkanization of the Windows platform" (Summary Response at 29-31). Microsoft asserts that that provision will "permit OEMs to perform radical surgery on Windows, ripping and replacing large blocks of software code, while still entitling them to market the resulting Frankenstein's monster using Microsoft's valuable Windows trademark and logos" (id.). But Section 3(a)(iii)(4) permits no such thing; it will not authorize OEMs to "rip and replace" anything in Windows, but will simply enable them to configure their systems so that non-Microsoft software can launch automatically, OEMs can offer their own Internet access provider or other start-up sequence, and non-Microsoft Middleware can be made the default. Indeed, the provision expressly confines OEM removal of Windows features to "the means of End-User Access for Microsoft's Middleware Product."
Microsoft's assertion that "[a]s more and more OEMs modified Windows to include non-Microsoft middleware, . . . the Windows platform would lose all consistency" (Summary Response at 30) is similarly unfounded. It is remarkable that Microsoft now contends that the mere addition of non-Microsoft middleware to PCS also including Windows is bad for consumers. The contrary is manifestly true, and the government's proposed relief seeks to ensure that consumers can reap the benefits of increased OEM freedom. OEMs currently are free to install additional software on their PCS; providing OEMs with the ability to permit such software to start automatically or be the default, rather than requiring manual invocation by the user, will not cause any fragmentation. Because little if any underlying software code and no APIs will be removed in the process of exercising this flexibility, Microsoft's references to OEMs "substituting" third party middleware for that shipped with Windows, and to developers' critical need to know that software code will be present on a PC (Summary Response at 31), are misplaced.
Finally, Microsoft complains that the proposed limited OEM flexibility in configuring PCS to meet their customers' needs would somehow authorize the creation of "derivative works" that would infringe Microsoft intellectual property rights. But Microsoft has never been able to articulate, either at trial or in its remedies papers, precisely how any of the covered actions would create derivative works or otherwise infringe its copyrights. The proposed OEM flexibility is crafted to address the types of actions that OEMs have proven themselves capable of undertaking and will leave undisturbed the workings of Microsoft operating system products.
The purpose of Section 3.b is simple and straightforward: It is intended to prevent Microsoft from maintaining barriers to entry in the operating system business by requiring it to disclose information needed to permit competing middleware to interoperate effectively with Windows. The provision will eliminate Microsoft's ability to hamper middleware threats by failing to disclose all the interfaces and information necessary for effective interoperation. Microsoft has used such tactics in the past and, absent this provision, could continue to use them to thwart nascent middleware threats in the future.
The evidence at trial and the Court's Findings provide ample basis for this requirement. For example, in the section entitled "Withholding Crucial Technical Information," the Court found that Microsoft expressly conditioned the timing of disclosure to Netscape of critically important information about Windows 95 on Netscape's agreeing to Microsoft's proposal not to compete. Findings ¶ 90. When Netscape refused to comply, Microsoft delayed disclosing the information for several months, thereby delaying Netscape's release of its browser and causing it to miss most of the holiday selling season. Findings ¶ 91. Microsoft similarly withheld a scripting tool needed by Netscape. Findings ¶ 92.
Elsewhere in its Findings, the Court recognized that independent software developers are "highly dependent" on early and predictable disclosure of technical information by Microsoft in developing their software products (Findings ¶ 338) and that this dependence gives Microsoft considerable power over ISVs. In fact, Microsoft conditioned early disclosure of important technical information to some ISVs on their agreeing to favor two Microsoft middleware products, the Internet Explorer browser and the Java Virtual Machine, conduct that the Court found to be "another area in which [Microsoft] has applied its monopoly power to the task of protecting the applications barrier to entry." Findings ¶ 340; see also id. ¶¶ 338-339.
In light of these Findings, Microsoft's assertions that Section 3.b's required disclosure of APIs, communications interfaces, and technical information has "nothing to do" with any issue in this case (Summary Response at 38) are hollow and inaccurate. In addition to the Court's Findings, there was a variety of other evidence presented at trial about the importance of protocols and of Microsoft's efforts to manipulate interface information to thwart competition.(13)
Microsoft's comment that there was no mention at trial of either protocols for communications between desktop and server versions of Windows 2000 or Kerberos (Summary Response at 38-39) is both unsurprising and unremarkable; Windows 2000 had not yet been released during the trial. Microsoft's comment reflects its crabbed view of antitrust remedies; especially in an industry like the software industry, which as Microsoft has repeatedly emphasized is rapidly changing, a remedy limited to barring repetition of the precise acts in the precise contexts that were at issue in the trial could not possibly serve the required purposes of preventing recurrence of the violations and restoring competition.
Moreover, Microsoft can hardly argue that client-server interoperability issues are unrelated to the trial. In the first place, its own expert, Dean Schmalensee, testified that control over the browser could enable a firm to "severely" affect the functionality of server applications. Schmalensee, 6/24/99pm at 46:19 - 47:10. Second, having argued during the trial that Microsoft lacked monopoly power in the operating-systems market because of the future potential of server-based applications, Microsoft can hardly contend now that it should be free to frustrate the threat to the Windows monopoly posed by such server-based applications by withholding critical information needed for those applications to interoperate with Windows.
Microsoft attempts to justify its continued selective disclosure of critical information by overheated rhetoric. Microsoft raises the specter of the "confiscation of Microsoft's intellectual property" (Summary Response at 32) that supposedly would occur if it were required to disclose to outside developers the same information actually used by its own non-PC operating system developers. In Microsoft's extraordinary interpretation, this basic equality-of-disclosure provision would require it "to hand over billions of dollars of intellectual property to companies like IBM and Sun Microsystems -- which will use such proprietary information to gain an unwarranted competitive advantage vis-a-vis Microsoft." Id. Microsoft would even end up with "no ability to protect itself against patent infringement claims" (id.) and "could soon find itself enjoined from distributing many of its products." Id. at 33. Along the way, Microsoft claims that Section 3.b., would lead to "forced disclosure of every detail about the way in which Microsoft's operating systems work" (id. at 32) and would "order Microsoft to disclose all proprietary information concerning its operating systems." Id.
Notably, Microsoft bases these assertions, not on the language of the provision itself, but on mischaracterized and selective quotations from the definitions. Perhaps the best way to appreciate Microsoft's exaggerations is to recite the language of Section 3.b, which expressly limits the scope and purpose of disclosure. It requires Microsoft to --
The language of the proposed Final Judgment makes clear that Microsoft will be required to disclose only "external interfaces used by Microsoft's own products to obtain services from Microsoft operating systems," Summary Response at 33 (citing Felten Declaration ¶ 69).
In the face of this language, it is untenable for Microsoft to assert (citing only the definition of "API") that the proposed relief would force it to "disclose all of the internal interfaces in the operating system . . . despite the fact that such interfaces were never intended to be called by other software products," (Summary Response at 34);(14) to the contrary, Section 3.b, is expressly limited to information that Microsoft itself "employs to enable" interoperability between three well-defined classes of software external to the Windows platform software and the Windows platform software itself. Microsoft's other references to having to disclose "millions" of internal interfaces, to "blocks of software code [that] are not equipped to deal with unknown software code from third parties," and "random internal interfaces," id. , are similarly dispelled by the plain language of the provision. In fact, nothing in the proposed remedy requires disclosure of any internal interfaces, unless they are actually used by Microsoft applications, separate Microsoft Middleware Products (as narrowly defined by Section 7.p), or Microsoft software installed on a different computer (such as a server or computing device) to interoperate with the Windows platform software running on a personal computer.(15)
Microsoft's statement that Section 3.b., would require disclosure of proprietary information, such as codecs, of a type that is used and kept confidential by other firms like Apple and RealNetworks (see Summary Response at 36), does nothing to demonstrate that the provision is inappropriate. In the first place, disclosure is required only to permit effective interoperation with the Windows operating system product and thereby to solve a serious competitive problem. The referenced other firms, unlike Microsoft, do not have control of the monopoly operating system with which their competitors' products must interoperate, and they thus do not have the ability or incentive to favor their company's own applications and middleware products and exclude competitors'. Second, and more important, Microsoft's obligation to disclose codecs and related information for interoperability purposes arises only if and when Microsoft chooses to incorporate them into Windows. If it does not do so, it will be able to keep its codecs or other information proprietary. Third, Microsoft will be able to protect any legitimate interest it may have in confidentiality, in the context of a remedy safeguarding competition, by disclosing the information subject to confidentiality agreements otherwise consistent with the decree that prohibit dissemination of the information and limit its use to the intended purposes. The evidence shows that Microsoft has in the past been willing to rely on contractual provisions to protect its intellectual property, such as when it disclosed the Internet Explorer source code to Compuserve and AOL. See GX 1125, 804.
Microsoft fares no better with the law than with the facts. It relies on the Areeda & Hovenkamp discussion of Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263 (2d Cir. 1979), to argue that the disclosure provision is a "departure" from existing law. But the Areeda discussion concerned issues of antitrust liability for nondisclosure, not the appropriateness of requiring disclosure as a remedial step for proven violations. See IIIA Philip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 776b2 at 238 (1996). Moreover, the discussion concerns predisclosure of information, not parity of disclosure at the same time that it is used by the monopolist. Id. Furthermore, the treatise makes the point that "the problem is more complex" where an integrated firm operating in complementary markets uses monopoly power to gain an advantage in both markets from an innovation in one. Id.
Finally, Microsoft argues that the compliance mechanism, the creation of a secure facility where qualified representatives of OEMs, ISVs, and IHVs can study Microsoft's source code for the limited purpose of ensuring interoperability, will lead to competitors having "free rein to appropriate Microsoft's intellectual property" and "unfettered access" they can use to "clone innovative features and functionality" of Windows. These contentions are at odds with the plain language of the provision, which limits disclosure to "qualified representatives" in a "secure facility" for "the sole purpose of enabling their products to interoperate effectively with Microsoft Platform Software." Moreover, Microsoft's stated concerns are belied by the fact that it currently discloses source code to other firms for particular purposes and with appropriate safeguards. See, e.g., GX433, sealed, (Compaq) GX1519 (Bristol), GX1125 and GX 804 (Compuserve and AOL, license for IE Source). Microsoft was even prepared to share source code with IBM, a firm it now asserts would clone Windows if offered such access (GX 2164, 2167; Norris 6/8/99am at 26:25-27:24.
In sum, nothing about the disclosure provisions of Section 3.b., will result in the confiscation of billions of dollars of Microsoft intellectual property or give its competitors the right to use such intellectual property to compete with Microsoft on a royalty-free basis to compete with Microsoft. Instead, the provision will simply ensure that Microsoft is not able to withhold from others, particularly developers of potential middleware threats, the same APIs and technical information actually used by Microsoft's own developers to make their software work with Windows.
Section 3.c will prohibit Microsoft from taking any action that will interfere with or degrade the performance of non-Microsoft middleware without notifying the supplier in advance so that the supplier can take efforts to ameliorate the problem. Relying again on its erroneous assertion that "middleware" includes "virtually all" software written for Windows, Microsoft asserts that this provision is overbroad and impractical. Summary Response at 40. In particular, Microsoft says that the provision obliges it to "help all developers of 'Middleware' fix their products." Id.
Microsoft's objections are again unfounded. First, "Middleware" as defined in plaintiffs' proposed remedy (Section 7.o) is far less than all software. Second, Microsoft is obliged only to notify middleware developers about problems that it knows its operating system modifications will create, and Microsoft is thus wrong when it complains that this provision will require it to seek out latent bugs in other firms' software that might be exposed by Microsoft's changes, to apportion blame if such bugs are exposed, or to ensure that developers of all software products know "how to rewrite them to make them work equally well with the new operating system." Summary Response at 40-41.
The obligations to inform imposed by Section 3.c., are entirely consistent with Microsoft's existing developer relations program, the purpose of which is, after all, to promote the proper functioning of third-party software on the Windows operating system. The provision simply ensures that Microsoft will not choose to undermine that purpose selectively to further an anticompetitive objective by undermining a rival.
This provision will prohibit Microsoft from discriminating against ISVs and IHVs on the basis of their support of any Microsoft product or service or non-Microsoft product or service or the exercise of any options provided under the Final Judgment. The record at trial demonstrates the need for this safeguard to prevent Microsoft's continued use of the wide array of opportunities presented by its monopoly position to bribe and coerce third parties to favor its own products and exclude others and the significant effects that such conduct has already had. See, e.g., Findings ¶¶ 64, 79, 83-103, 115-132, 139, 203-208, 230-335, 337-356, 401, 403-406, 410-12; Conclusions at 19.
Microsoft reads this provision misleadingly and selectively to prohibit it from "providing any information to software developers or hardware vendors if doing so might 'affect' their decision whether to 'use, distribute, promote or support" any Microsoft product or service." Summary Response at 42. In fact, the provision says that "Microsoft shall not take or threaten any action affecting any ISV or IHV . . . based . . . on any actual or contemplated action by that ISV or IHV to -- i. use, distribute, promote, or support any Microsoft product or service . . . ." § 3.d. (emphasis added). Thus, the provision plainly will not prohibit Microsoft from providing technical information and other benefits necessary for developers to support its products. Nor will it prohibit Microsoft from "affecting" other firms by competing (Summary Response at 41-42).
Section 3.d., also will not require, contrary to Microsoft's suggestion, that information be provided in any particular form or forum, "town hall" (Summary Response at 42) or otherwise, or indeed that any information be provided at all (beyond that required by the disclosure provisions of Section 3.b. discussed above). Section 3.d. simply prohibits Microsoft from favoring or disfavoring particular ISVs and IHVs depending on whether they support Microsoft or its rivals.
Nor will Section 3.d. "block" joint development efforts. Summary Response at 42. Section 3.d.i, which will prohibit Microsoft from rewarding or punishing ISVs and IHVs based on their actions with regard to distribution and promotion of Microsoft products or services, does not involve joint development in any way; and Section 3.d.ii., will prohibit Microsoft only from rewarding or punishing ISVs/IHVs depending on how they treat non-Microsoft products. Neither provision restricts Microsoft from entering into bona fide joint development efforts.
Finally, Microsoft's allegation that Section 3.d., will prevent third parties from sharing information with it (Summary Response at 42) is fantasy. Nothing in the Section or elsewhere in the proposed remedy prevents Microsoft from entering into and honoring legitimate agreements to maintain the confidentiality of other firms' confidential information.
Section 3.e., will bar Microsoft from continuing or repeating the types of exclusive agreements that Microsoft used in its illegal campaign to maintain its Windows monopoly in violation of Section 2 of the Sherman Act. Microsoft's objections to this Section refer only to the language of Section 3.e.i, which prohibits agreements with a third party to "limit its development, production, distribution, promotion or use of, or payment for, any non-Microsoft Platform Software." Summary Response at 44-45.(16)
Microsoft first contends that the restriction is too broad because it applies to Platform Software, which is defined to include Middleware (Summary Response at 44, 39). As explained above, however, the term Middleware as defined in the Final Judgment is nowhere near as broad as Microsoft says. See Section 7.o., discussed in Part IV.B., above.
Moreover, Microsoft combines its overbroad reading of Middleware with a misreading of the plain language of Section 3.e.i, to argue that the provision will prevent it from entering or maintaining any deals with third parties; Microsoft reasons that all deals to promote or use Microsoft products give third parties an incentive to "limit" their promotion or use of competing products. Summary Response at 44-45. But, by its terms, Section 3.e.i., applies only to "any agreement . . . to limit" activities with Microsoft's competitors. It bans "agreements" with Microsoft about the scope of a firm's engagement with Microsoft's competitors, not independent decisions by the firm about whether to buy or promote non-Microsoft products. Microsoft's unsubstantiated assertion that "the government would contend" otherwise (Summary Response at 44) is baseless conjecture.
Microsoft also argues that its agreements with OLSs, ISPs, and ICPs were found to be unlawful because Microsoft used the "currency" of placement on the Windows desktop as payment for the virtual exclusivity provided by those agreements and, therefore, that Section 3.e's prohibition on Microsoft paying or offering any other form of consideration for exclusive contracts is overbroad. Summary Response at 45. But the agreements challenged at trial and held to be unlawful were illegal, not because of the form their payment took, but because the exclusivity they demanded in return for that payment injured Netscape and Java and thereby entrenched Microsoft's operating system monopoly. See Conclusions at 15-17. There is thus no reason to restrict the prohibition on such exclusive arrangements to those that involve desktop placement.
Section 3.f., is a straightforward provision that will prevent Microsoft from contractually forcing OEMs and end users of Windows to take other software products, whether they want them or not, as a condition of receiving a Windows operating system license. Such tying is at the heart of the case; the Court found that Microsoft used such "contractual . . . shackles" (Conclusions at 11) in violation of both Section 1 and Section 2 of the Sherman Act. Conclusions at 19-21.
Microsoft responds with several unfounded and erroneous assertions. First, it says that Section 3.f., requires that "any software code separately distributed from Windows -- no matter what its nature -- would have to be made optionally removable." Summary Response at 48. This is simply wrong. Section 3.f, both by its use of the term "Microsoft software product" and by contrast with Section 3.g, which specifically addresses instances in which a product is bound into Windows "in such a way that either an OEM or end user cannot readily remove or uninstall" it (Section 7.d), makes clear that it will apply only when the separate, tied product is already removable. (Remarkably, directly contrary to the Court's Findings, Microsoft repeats its familiar misstatement that "Internet Explorer . . . cannot be made optionally removable by OEMs -- the operating system will not work without" it. Summary Response at 47; compare with Findings ¶¶ 166-167, 182-184, 187, 191.) Moreover, the reference in Section 3.f., to "software product" dispels the notion that this provision applies to any software code, "no matter what its nature."
Second, Microsoft contends that this provision will forbid Microsoft from improving Windows and distributing upgrades, such as bug fixes and other "improvements to Windows, that Microsoft makes available free of charge for downloading from its Windows Update Web site" (Summary Response at 48). Microsoft can make this argument only by ignoring critical language from Section 3.f., which makes clear that the provision applies only to "any other MS software product that MS distributes separately from the Windows OS Product in the retail channel or through Internet access providers, Internet content providers, ISVs or OEMs . . . ." Summary Response at 46 (highlighted language omitted by Microsoft). By its terms, the provision will not restrict Microsoft's ability to continue to distribute any Windows improvements by downloading from the Windows Update site.
Third, Microsoft once again invokes its fragmentation argument, contending that Section 3.f., "contemplates actual removal of software code." Summary Response at 47. To the contrary, just as plaintiffs challenged Microsoft's contractual tying of the browser only because Microsoft would not use, or permit OEMs to use, the Add/Remove utility or a similar means to remove the browser -- and did not call for removal of the shared code -- so Section 3.f., will not require wholesale removal of code or lead to fragmentation of the Windows platform. See Findings ¶¶ 165, 184-185; Felten Declaration ¶¶ 89, 94.(17)
Section 3.g. will prohibit Microsoft from binding or "hard-wiring" separate middleware products to its operating system. Microsoft will be free to offer such products bundled with its operating systems, but it will be required to permit OEMs and end users that wish to do so to license the operating system without being forced also to take unwanted other products.
This provision will prevent Microsoft from repeating the illegal conduct that the Court found it undertook with respect to the browser. See, e.g., Findings ¶¶ 164, 166-174, 176; see also Zenith, 395 U.S. at 132 (a remedy should prevent defendant from repeating the "same type or class" of unlawful conduct). Forced bundling injures consumers directly and injures competition by increasing the costs rival software vendors must incur to get their products distributed effectively. It is an especially potent competitive weapon for Microsoft because Microsoft is able to target competing middleware threats -- like the browser -- by bundling its own version with its operating system monopoly, thereby protecting that monopoly.
Once again, Microsoft's response predicts dire consequences, this time sufficient to bring the "worldwide computer industry to its knees." Summary Response at 50. And once again its response is based on a distorted reading of the proposed Final Judgment:
(1) Microsoft reads "Middleware" (§ 7.o) far too broadly. See Summary Response at 49 ("'Middleware Product' encompasses anything that Microsoft might choose to add to one of its operating systems.") Microsoft's misreading is discussed above, in part IV.B.
(2) Microsoft ignores the definition of "Middleware Product" (§ 7.p), which is the term to which Section 3.g., applies and which is much narrower than "Middleware" (§ 7.o). That definition ensures that the anti-binding provision will apply to only a small group of products: (i) five specifically enumerated types of middleware (browsers, e-mail clients, multimedia (e.g., streaming media) viewers, instant messaging software, and voice recognition software) and (ii) other software that both (a) is or has been in the past year distributed separately from the operating system in the retail, IAP, ICP, ISV, or OEM channels by Microsoft (or by another company if Microsoft acquired the product from that company) and (b) provides functionality similar to that of competing, non-Microsoft Middleware. Microsoft is wrong when it says that the prohibition applies to "any software" that performs the same function as software offered separately "by anyone in the preceding year." Summary Response at 49 (emphasis added).(18)
(3) Microsoft suggests that Section 3.g.'s requirement of removal of "end user access" dramatically increases the scope of what is a "Middleware Product." But only if a product first meets the definition of "Middleware Product" is Microsoft required to provide the means of removing access to it. Microsoft's suggestion that Windows DLL files such as MSHTML.DLL could not be included in Windows and therefore could not provide, for example, user interface functionality (Summary Response at 50) is thus misplaced. Similarly, Microsoft's statement that features like the user interface, HTML Help, and Windows Update would be "precluded" because they "are dependent on Internet Explorer" is erroneous. Section 3.g., requires that OEMs and end users be able to remove access only to the middleware product -- in this case the browser -- not to APIs or code. See Felten Declaration ¶¶ 92, 94; Findings ¶¶ 183-185.
(4) Without reference to the actual language of Section 3.g, Microsoft asserts that it will be prohibited from binding products to Windows regardless whether there is separate consumer demand for such products. Summary Response at 6. But Section 3.g., applies only to products that have been distributed separately from operating systems, and it is perfectly appropriate for a remedy to use a simple, unambiguous standard like that, rather than requiring a full-blown assessment of the Jefferson Parish(19) separate demand test in order to determine where the remedy applies.
(5) Microsoft argues that, if it had to abide by this modest "unbinding" requirement, over a million (220) different configurations of Windows would result and a host of intractable problems, including testing and product support, would ensue. This is a remarkable claim in light of Microsoft's routine, existing practice. As the evidence at trial showed, Microsoft already makes about 80 components removable from Windows 98 through an Add/Remove utility. Felten Declaration ¶ 94, citing Felten 6/10/99pm at 6:7-16; GX 1700; see GX 1366 (showing dozens of functions that can be added or removed by the user, including, among other things, internet tools, desktop wallpaper, mouse pointers, dial-up networking, virtual private networking, and hyper terminal); Allchin, 2/2/99pm, at 5:2-5 (Microsoft provides a ready means of removing many files and features that Microsoft considers to be "integrated" features of Windows); 2/2/99pm, at 10:3 - 11:11. Under Microsoft's existing practice, it already permits well over 280 -- or somewhere in excess of 1 followed by 24 zeros -- different configurations.
(6) Finally, Microsoft asserts that offering the straightforward "unbinding" option for OEMs and end users for the few Middleware Products in existing operating systems would take "several years and hundreds of millions of dollars" for "redesign" and would be "completely impossible." Summary Response at 49. This unsubstantiated assertion cannot be reconciled with the record in this case and the Court's Findings. Professor Felten's removal program for Microsoft's browser, which achieves just the sort of removal of access that Microsoft would be required to provide under Section 3.g., without degrading any other part of the operating system, was developed in a very short time and at minimal expense. See Findings ¶¶ 177, 181, 183-184. And, as the Court found, "[g]iven Microsoft's special knowledge of its own products, the company is readily able to produce an improved implementation of the concept illustrated by Felten's prototype removal program." Findings ¶ 182, 177. It will be equally easy to "unbind" the other middleware products that are currently in Microsoft's existing operating systems, and six months from the effective date of the remedy order "is ample time for Microsoft to make the necessary changes . . . ." Felten Declaration ¶ 96; see also id. ¶¶ 93-95.
Microsoft continues its pattern of strained interpretations of straightforward language in its response to Section 3.h. According to Microsoft, prohibiting it from offering or agreeing to provide an actual or potential competitor any consideration "in exchange for such competitor's agreeing to refrain or refraining . . . from developing, licensing, promoting or distributing" competitive operating system or middleware software will prevent it "from even discussing, much less engaging in, co-development of promising new technologies," and bar "routine technical exchanges" and, indeed, will constrain Microsoft "from talking with virtually anyone in the software industry." Summary Response at 53, 55. But Section 3.h., is intended only to prohibit naked bargains not to compete, like Microsoft's attempted market allocation with Netscape and Intel, not joint development agreements, the exchange of technical information or agreements that are ancillary to lawful joint ventures, employment arrangements or corporate acquisitions.
Microsoft also objects that there is "no finding that Microsoft agreed with anyone 'to refrain or refrain[ed] in whole or in part from developing, licensing, promoting or distributing' a product that competed with a Microsoft product." Summary Response at 53. But Microsoft repeatedly sought such anticompetitive agreements, see Findings ¶¶ 80-132, so the Court has more than ample basis to prohibit them by its remedy. Zenith, 395 U.S. at 132.
Finally, Microsoft argues that the provision is overbroad because it covers competitors not only in operating systems, but in middleware as well. This argument ignores the Court's findings that Microsoft succeeded in preventing a competitive browser from emerging as an alternate platform. Findings ¶¶ 377-385. Prohibiting anticompetitive activity that could stifle the emergence of other forms of middleware as potential platforms is necessary both to prevent recurrence of past misconduct and to restore competitive conditions.
Microsoft's only objection to Section 3.i., is that it will apply to so-called "interim" operating system releases such as OSR 2.0 and OSR 2.5 and will thus be burdensome for Microsoft. Summary Response at 55. But the evidence at trial showed that Microsoft may make significant changes in interim releases that could provide a compelling basis for OEMs to want to continue distributing, and customers to continue licensing, the previous interim release (e.g., the inclusion of IE 4 with OSR 2.5). Subjecting these releases to the requirements of this provision is also necessary to ensure that Microsoft does not evade the purpose of the provision by manipulating whether it puts significant changes in "interim" releases or "major" releases in order to coerce OEMs to move to the newest release by, for example, effectively forcing OEMs to take an "interim" version that contains bundled middleware that the OEM may not want.
Microsoft protests that the Internal Antitrust Compliance provisions of the Plaintiffs' proposed remedy are unwarranted and burdensome, but in fact they are fully warranted by the pervasiveness of Microsoft's "well-coordinated course of action" (Conclusions at 20) and predatory campaign (Conclusions at 21) to protect its monopoly. Microsoft has shown "that it will use its prodigious market power and immense profits to harm any firm that insists on pursuing initiatives that could intensify competition against one of Microsoft's core products." Findings ¶ 412. There is no basis to conclude that the proposed compliance measures are unwarranted.
(1) Apart from its generalized complaint, Microsoft makes only two specific objections. First, it argues that the email retention requirement will be "burdensome to the point of absurdity" because it will cover thousands of employees and tens of millions of e-mail messages. Summary Response at 56-57. This argument is greatly exaggerated. Even if Microsoft's assumptions about the requirement's coverage are correct, and if the average size of an e-mail message is 5000 bytes (5kB), then the amount of data that would be required to be retained is on the order of, say, 20,000,000 messages x 5kB = 100 gigabytes. The current price of storage media holding this amount of data is approximately $600, hardly an "absurd" burden.
Second, Microsoft says that the compliance requirements would cripple its ability to develop products. But this argument is based largely on unsupported speculation about the purported reaction of IBM engineers to the very different consent decrees under which that company operated. Microsoft asserts, without any documentation, that these engineers "steered clear of entire fields of study" because they feared that engaging in those fields might be seen by the government as "potential violations" (Summary Response at 56), and it implies that Microsoft employees, who would otherwise "think outside the box," would behave similarly. That is nonsense.
(2) Microsoft argues that the inspection provisions of the proposed remedy are unreasonably broad and would permit "perpetual, essentially unbounded" investigations of Microsoft. MS FJ Memo at 11. But it neglects to mention that substantially identical provisions are standard is almost all antitrust consent decrees and that, in fact, the proposed provisions are virtually identical to the Inspection provisions contained in the 1994 Microsoft consent decree, agreed to by Microsoft and entered by this Court.
The only provision that is materially different from the 1994 decree is Section 5.a.i(2), which specifies that employees being interviewed may have their individual counsel present. This language is becoming standard in recent antitrust decrees and gives the employees choice about whether to have corporate counsel or their own counsel present at any interview.
(3) It is also appropriate to grant the States parity with the United States to perform compliance inspections with respect to matters relating to the decree. The States are here suing in their parens patriae capacity, as protectors of quasi-sovereign interests. The States' authority to bring such suits to protect their citizens and economies by enforcing the antitrust laws in matters of "grave public concern" is well established. E.g., Georgia v. Pennsylvania R.R. Co., 324 U.S. 439, 450-451 (1945) (conspiracy in violation of antitrust laws is a wrong "of grave public concern in which Georgia has an interest apart from that of particular individuals who may be affected"); Hawaii v. Standard Oil Co., 405 U.S. 251, 257-60 (1972) (summarizing cases establishing the right of a State "to sue as parens patriae to prevent or repair harm to its 'quasi-sovereign' interests); In re Insurance Antitrust Litigation, 938 F.2d 919, 927 (9th Cir. 1991) ("The state's interest in preventing harm to its citizens by antitrust violations is, indeed, a prime instance of the interest that the parens patriae can vindicate by obtaining . . . an injunction.").
The States have been granted full rights under standard inspection provisions substantially identical to those at issue here, both when they have joined with the United States as plaintiffs, e.g., United States v. Sony Corp. of America, 1998 U.S. Dist. Lexis 20815, 2000-1 Trade Cas. (CCH) ¶ 72, 787 (New York and Illinois, joining as plaintiffs with the United States, granted full rights to inspect and copy documents, interview witnesses, and require written reports under oath for compliance purposes in consent decree resolving merger action) and when separately exercising their independent authority to enforce state and federal antitrust laws.
Microsoft expresses vague "doubts" as to whether such standard visitorial provisions, if based solely upon the States' authority to enforce their own state antitrust laws, would be "consistent with constitutional principles" where a corporate defendant is incorporated and headquartered out-of state. Microsoft's Summary Response at 59. But such doubts are inapposite here, where the States seek to enforce claims under federal as well as state antitrust laws. Indeed, such provisions would be proper even under state laws which, as this Court has found, reach Microsoft's unlawful conduct because that conduct has "significantly hampered competition" in each State. Conclusions at 41.
V. Microsoft's Proposed Schedules And Procedures Are Unwarranted And Unreasonable And Can Serve No Purpose Other Than Prolonging The Remedy Process
Microsoft's Position as to Future Proceedings on the Issue of Remedy is a transparent effort to delay the determination and implementation of a remedy for its illegal acts as long as possible. Microsoft suggests that this Court should select a remedy first and then set a hearing schedule accordingly, with ever more burdensome discovery and more distant hearing dates as the selected relief becomes more significant. Meanwhile, Microsoft's ongoing harm to competition and consumers would go uncorrected, and Microsoft would continue to enjoy the fruits of its illegal conduct while further raising entry barriers and entrenching its monopoly.
Microsoft's violations have been established. Liability is not in doubt, and relief should be as prompt as possible. Having refused to engage effectively on the merits of plaintiffs' remedy, Microsoft should not be permitted to deprive it of force by pointlessly delaying its implementation.
Microsoft says that it needs extensive discovery in order to address plaintiffs' proposed remedy. But there is little in plaintiffs' proposal that Microsoft can really say is unexpected. Microsoft has known for several months about plaintiffs' interest in structural relief, information disclosure and the other forms of transitional conduct relief included in the proposed remedy; yet it is unable to specify any facts about which it needs discovery, from whom it could discover those facts or why it needs such discovery. The reason, of course, is that it does not need discovery because the relevant information needed to evaluate plaintiffs' proposed remedy -- whether about corporate structure or information disclosure or OEM or ISV relations -- is within Microsoft's particular knowledge and control.
Microsoft's proposal, with its alternative schedules tied to a pre-determination of the appropriate remedy by this Court, is merely an attempt to delay the day when the law will hold it accountable for its illegal acts. Because Microsoft has demonstrated no legitimate need for the substantial delays and wide-open discovery it seeks, its various requests for them should be denied.
The United States, seventeen of the Plaintiff States and the District of Columbia request
that the Court enter the proposed Final Judgment.(20)
1. Microsoft's various filings are cited as follows: Memorandum in Support of Motion for Summary Rejection of Government's Breakup Proposal ("Summary Rejection Memo"); Summary Response to Plaintiffs' Proposed Final Judgment ("Summary Response"); Microsoft's Memorandum in Support of Its Proposed Final Judgment ("MS FJ Memo"); Position as to Future Proceedings on the Issue of Remedy ("Remedy Proceedings"); and Microsoft's Proposed Final Judgment ("MS Proposed Final Judgment"). Plaintiffs' Memorandum in Support of Proposed Remedy, dated April 28, 2000, is cited as "Plaintiffs' Memo in Support." This Court's Findings of Fact, dated November 5, 1999, are cited as "Findings;" and its Conclusions of Law, dated April 3, 2000, are cited as "Conclusions."
2. Section 4(b)(iii) would prohibit Microsoft from entering into an agreement or asserting a copyright claim to prevent an OEM from deleting the Internet Explorer icon from the desktop. But neither it nor any other provision bans other forms of OEM coercion or requires Microsoft to permit OEMs to remove all ready means of access to the Internet Explorer browser.
3. There are numerous other loopholes, omissions, and inadequacies in Microsoft's proposal (including, for example, its limitation to a term of 4 years). In seeking to excuse the shortcomings of its proposal, Microsoft suggests that its proposed order could be entered promptly and thus provide "substantial immediate relief" (Remedy Proceedings at 4). But that promise is illusory because Microsoft proposes, in Section 9, that the Court's order provide that "Microsoft may seek a stay" of the entire decree pending appeal.
4. See GX 2214; "Microsoft Office 2000 Enables Developers to Quickly Deliver Custom Business Applications That Meet Business Needs" and "Microsoft Office 2000/Visual Basic Programmer's Guide" http://www.microsoft.com/PressPass/features/1999/06-14o2k.asp and http://msdn.microsoft.com/library/officedev/odeopg/deovrthebenefitsofofficeprogrammability.htm (Remedy GX 55).
5. See "Microsoft® BackOffice® Integration with Microsoft Office 2000 White Paper" at http://www.microsoft.com/Office/evaluation/solutions/IntgBack.htm (Remedy GX 57).
6. Indeed, Areeda and Hovenkamp assert that "a monopoly to which plainly exclusionary conduct appears to have made a significant contribution is itself unlawful." Id. at 97. That statement may well fit this case, but the Court need not decide that issue because plaintiffs are seeking only a structural remedy that will lower entry barriers, not the "complete extirpation" of Microsoft's monopoly. Id., quoting United States v. United Shoe Mach. Co., 391 U.S. 244, 250-252 (1968).
7. Like Microsoft today, AT&T complained that this requirement in effect forced it to "give away its technology." Western Elec. Co., 569 F. Supp. at 1085 (emphasis by Court). But the court reasoned that, because RBOC revenues had funded much of the Bell System's research, "that technology is as much theirs as it is AT&T's." Id.
8. See also Devlin Direct ¶ 17 ("Microsoft often seeks input from ISVs and other sectors of the software and computer industry when it develops new APIs. Microsoft widely publishes its APIs, conducts seminars for developers and provides ISVs with detailed, inexpensive technical support in writing programs that use Windows APIs. Microsoft's unmatched commitment to the developer community is another reason for Rational's success, and, no doubt, for Microsoft's success as well."); Ballmer press conference, http://www.microsoft.com/ PressPass/trial/apr00/04-03conference.asp ("I had an opportunity this morning to talk to a wonderful software developer building on our platform and the list and range of new things that they want to see in the basic platform was daunting, but we are dedicated to delivering on the kinds of things that folks like that want.") (Remedy GX 42).
9. Microsoft's chief executive officer, Mr. Ballmer, recently explained the key to the success of Microsoft and the personal computer: ". . . a whole different industry structure. A structure, which still is maintained today. A structure of specialization. You have chip companies, you have communication companies, you have systems software companies, you have applications companies. People tend to specialize. Now, we've been called out because we participate in two sectors of those, but, heck, it's still a very specialized business. ... We've got to have a way of communicating with software developers who now number almost three million strong in the United States alone. ....... From a Microsoft perspective, we think this represents a world of quite a bit broader partnerships, quite a bit broader set of partnerships. And I don't just mean partnerships like we and you need to be exclusive together. I mean little "p" partnerships. . . . [for disclosure] unless we're communicating clearly about our common vision and sharing technical specifications in an open way, we will all fail to achieve that which is very much upon us and possible." Remedy GX 51(George Washington University, April 19, 2000, http://www.microsoft.com/PressPass/exec/steve/04-19gwu.asp).
12. Not surprisingly, Microsoft's Ballmer described NGWS in less sweeping terms when he took over as CEO in January of this year: "As our company did with Windows in offices, we will create some specific key services that use our platform [NGWS]. Those services will focus in on a very specific set of user scenarios. Important, but certainly very narrow compared to the broad set of opportunities. . . . Most of the services, as I said, will be delivered by other company. [sic] And the platform evolution that we're talking about will be very open." Remedy GX 54 (http://www.microsoft.com/PressPass/features/2000/01-13ballmerceotranscript.asp).
13. See, e.g., GX 282 at 00126 (11/7/95 Intel memo quoting Paul Maritz: "Agreed to release RL DDI [Reality Labs device driver interface] under gentlemen's agreement that we [Intel] pull-out of 3DR [Intel's competing 3d rendering effort]); McGeady 11/9/98pm at 38:1 - 39:3; Findings ¶ 377 (citing Mehta email that Microsoft set out "2 years ago not to let netscape dictate standards and control the browser api's"); GX 911 (Christian Pierry acknowledging that Microsoft did not want to fix IE for numerous file types); DX 1786 (acknowledging that information relating to the Windows registry's routing of media data is undocumented).
14. Microsoft's purported reliance on the definition of "API" (Section 7.b) in this argument is based on a mischaracterization of what that definition actually says. Microsoft writes that, by supposedly requiring disclosure of interfaces that "might" enable products to "'benefit from'" the "'resources, facilities, & capabilities of' the operating system" (Summary Response at 34), this provision is "so vague and all encompassing" that it would force Microsoft to disclose millions of internal interfaces. In fact, the actual language of Section 7.b., says nothing about "might enable", but instead reads "that enable . . . ." That language, combined with the express terms of 3.b., itself -- that it covers only APIs, Technical Information, and Communications Interfaces "that Microsoft employs to enable . . . ," -- makes clear that the provision would require the disclosure only of what Microsoft actually uses for the interoperation of the three enumerated categories of plainly operating system-external software. Moreover, while Microsoft cites only the "benefit from . . . resources, facilities, and capabilities" of the operating system language (Summary Response at 34), it omits the other half of the definition, which makes clear that the definition is limited to interfaces, etc., that enable non-operating system products "to obtain services from" PC Platform software "and to benefit from . . . resources, facilities, and capabilities, etc." Section 7.b.
15. See discussion of "Middleware Product" in the treatment of Section 3.g., below. Because "Middleware Product" is narrowly limited to a small number of products that Microsoft distributes separately from the operating system, this definition provides no basis for the contention that requiring disclosure of the interfaces and information about interoperability between Microsoft Middleware Products and Microsoft Operating Systems requires disclosure of anything internal to the operating system.
16. Microsoft does not mention Sections 3.e.ii - 3.e.iv, which cover exclusive deals (Section 3.e.ii; see Findings ¶¶ 233, 320, 402); agreements by third parties to degrade the performance of non-Microsoft products (Section 3.e.iii; see Findings ¶ 322); and agreements to trade placement by Microsoft in Windows for adoption or promotion of Microsoft products (Section 3.e.iv; see Findings ¶¶ 255, 257-258, 322).
17. Of course, as Microsoft consistently ignores whenever it invokes the specter of fragmentation, "Microsoft itself precipitates fragmentation of its platform by continually updating various portions of the Windows installed base with new APIs" (Findings ¶ 193) and by permitting users to remove some 80 different "components" of or products from Windows 98. See Felten 6/10/99pm at 6:7-16; GX 1366.
18. The definition does include software that has, in the past year, been distributed separately by a third party from which Microsoft acquired that software. Otherwise, Microsoft would be free to acquire a promising, separately distributed middleware product, quickly bind it to Windows without ever distributing it separately itself, and then claim an exemption from the anti-binding remedy.
19. Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2 (1984).
20. Ohio and Illinois are in full agreement with the Department of Justice, the 17 States and the District of Columbia that Microsoft's Proposed Final Judgment is inadequate to remedy the serious violations found by this Court and that the delay sought by Microsoft in its Position as to Future Proceedings on the Issue of Remedy is excessive. Those States remain reluctant, however, to propose the imposition of structural relief before there is an opportunity to determine whether significant conduct relief alone would be sufficient to end Microsoft's anticompetitive behavior and restore competition to this vital market. For this reason, they urge the Court to impose the conduct relief proposed by the Plaintiffs immediately and in its entirety.