|This document is available in two formats: this web page (for browsing content), and PDF (comparable to original document formatting). To view the PDF you will need Acrobat Reader, which may be downloaded from the Adobe site. For an official copy, please contact the Antitrust Documents Group.|
January 28, 2002
Renata Hesse, Esq.
Dear Ms. Hesse:
Pursuant to the instructions in the Competitive Impact Statement in United States v. Microsoft Corp., we are submitting to the Department of Justice as an attachment to this e-mail the Comments of SBC Communications Inc. on the Proposed Final Judgment. We would appreciate your sending a reply to this email at your earliest convenience to confirm your receipt of SBC's submission.
In addition, to guard against the risk of a faulty email transmission, we are tonight sending a hard copy of SBC's Comments to you via U.S. Postal Service Express Mail.
Thank you for your consideration.
Very truly yours,
COMMENTS OF SBC COMMUNICATIONS INC.
January 28, 2002
TABLE OF CONTENTS
Ford Motor Co. v. United States,
Hartford-Empire Co. v. United States,
Int'l Salt Co. v. United States,
N. Pac. Ry. v. United States,
Nat'l Soc'y of Prof'l Eng'rs v. United States,
Otter Tail Power Co. v. United States,
Schine Chain Theaters, Inc. v. United States,
United States v. Am. Tel. & Tel. Co.,
United States v. American Broadcasting Co., Inc.,
United States v. Business Inv. & Dev. Corp.,
United States v. Crescent Amusement Co.,
United States v. Delta Dental of R.I.,
United States v. E.I. du Pont de Nemours & Co.,
United States v. General Elec. Co.,
United States v. Glaxo Group Ltd.,
United States v. Greyhound,
United States v. Grinnell Corp.,
United States v. GTE Corp.,
United States v. Microsoft Corp.,
United States v. Microsoft Corp.,
United States v. Playmobil USA, Inc.,
United States v. Republic Services, Inc.,
United States v. U.S. Gypsum Co.,
United States v. United Shoe Mach. Corp., 110 F. Supp. 295
United States v. United Shoe Mach. Corp.,
United States v. Western Elec. Co.,
United States v. Western Elec. Co.,
United States v. Western Electric Co.,
Zenith Radio Corp. v. Hazeltine Research, Inc. 395 U.S. 100 (1969)
15 U.S.C. § 16 11, 28, 29
42 U.S.C. §§ 251-59, 271
Benjamin Woodhead, Microsoft's Australian Monopoly? Let the U.S. Handle It,
Browser Bruiser, Chicago Sun Times, October 27, 2001
Byron Acohido, Challenging Microsoft? It Could Take Moxi,
Byron Acohido, Microsoft Memo to Staff: Clobber Linux, USA Today, Jan. 4, 2002
T. Capers Jones, Estimating Software Costs Function Point Analysis: Measurement
David Garmus and David Herron, Function Point Analysis: Measurement Practices
Department of Justice, Antitrust Division Manual
Don Clark, AOL Sues Microsoft Over Netscape in Case That Could Seek Billions, Wall Street Journal, Jan. 23, 2002
Experience the Connected Home: Share One or Many Computers (May 9, 2001),
Is Apple Out of the Running in the Operating Systems War? (Jan. 8, 2002), at http://www.websidestory.com/cgibinwss.cgi?corporate&news&press_1_163.
Jesse Berst, Office Suites for Free, ZDNet AnchorDesk (March 7, 1997), at http://www.zdnet.com/anchordesk/story/story_743.html
Lee Gomes, Linux Campaign Is An Uphill Battle For Microsoft, Wall Street Journal, June 14, 2001
Microsoft Unveils New Home PC Experiences with "Freestyle" and "Mira",
MSN Shuts Out Other Browsers, Associated Press, October 28, 2001
Rebecca Buckman, Microsoft is Suing Linux Start-up Over Lindows Name,
Wayne Epperson, NT Insurance at a Premium, HostingTech (August 2001), at www.hostingtech.com/security/01_08_nt
SBC Communications Inc. ("SBC") respectfully submits the following comments pursuant to Sections 2(b) and 2(d) of the Antitrust Procedures and Penalties Act, 15 U.S.C. § 16(b)-(h), relating to the revised proposed Final Judgment that was agreed to on November 6, 2001, by the United States and certain state plaintiffs in these actions on the one hand, and defendant Microsoft Corporation ("Microsoft") on the other (the "proposed settlement").
The history of Sherman Act enforcement has witnessed few unlawful monopolies as durable, resilient and exclusionary as Microsoft's. This much is clear from the trial record, the District Court's monopoly maintenance findings and the Court of Appeals' affirmance. Far from providing reassurance that changes in technology will end Microsoft's stranglehold over operating system and middleware competition, or that the company's monopoly will be subject to serious competitive pressures when the proposed settlement's five-year term expires, the record demonstrates the exact opposite. Microsoft's continuing ability to commingle its browser and operating system, which the settlement ignores, leaves Microsoft with the incentive and ability not only to destroy traditional middleware threats to its operating system monopoly, but also to exercise anticompetitive control over the Internet, where server networks currently not dependent on Windows pose the greatest threat to the Microsoft monopoly.
The consequences of failing to restrain an ever-expanding Microsoft operating system monopoly -- now at over 95% market share -- do not, however, fall solely upon software producers whose competitive assaults might erode that overwhelming market domination. Nothing in the proposed settlement would stop the threat that Microsoft's adjudicated and unlawfully-maintained monopoly poses to the very heart of consumer choice in the American economy. The settlement ignores Microsoft's ability to effectively destroy free consumer choice among the far greater array of businesses that use electronic means of communication -- such as telecommunications services (local, long distance and cellular), Internet access, voice messaging, instant messaging, video and music services, e-commerce, interactive games, to name a few. The settlement would allow Microsoft to abuse its illegally-maintained control of operating systems by becoming the ultimate "gatekeeper," controlling the bottleneck that both gives businesses in these critical related markets (whether established or still emerging) access to potential customers, and gives consumers the means to reach the providers that they choose to deal with.
Just as Microsoft has for years successfully imposed on consumers its own products and services, irrespective of the comparative merits of competing products it has excluded from the market, Microsoft will -- without the kind of strong relief required to break its operating system monopoly -- be in a position to repeat its anticompetitive strategy in other markets. Unchecked, Microsoft will favor its own and its partners' services, exclude competitors' products and services from access to consumers, and degrade its rivals' services and raise their costs (by charging a toll, imposing a fee for listing as an available service or creating an interoperability obstacle). Because potential customers will have to pass through a Microsoft operating system (whether embedded in a PC, a cellular phone, a set-top box or a PDA), Microsoft will retain the ability to exclude or marginalize all manner of telephone services, messaging products, video or music offerings, Internet services, and other "utilities" of modern life. In this way, the Microsoft monopoly threatens to destroy the vast panoply of consumer choice among the myriad sources that create and distribute communications and entertainment products and services. The proposed settlement does virtually nothing to lessen Microsoft's ability to maintain its operating system monopoly and to prevent its enhancement by Microsoft's impeding effective competition for all the products and services that will have to be accessed through Microsoft's monopoly platform.
SBC is one of the businesses that will be significantly impacted. Through its affiliates, SBC provides voice and data communications services throughout the United States and internationally. Some of these services are Internet-based; others are not. Some of SBC's services (such as its unified messaging service, discussed below) would erode Microsoft's operating system monopoly; others will not. All, however, are at risk if Microsoft is not prevented from maintaining and expanding its operating system monopoly. Thus, while SBC devotes a significant portion of its comments to explaining why curing the palpable deficiencies in the proposed settlement is essential to protect Internet-based services that could erode the Microsoft monopoly, including SBC's own ventures, those deficiencies are of equal importance to SBC's core communications businesses.
The reason why the effects of the Microsoft monopoly reach so far can be summed up in a single word -- "convergence." Convergence refers to the development, for home or office use, of devices or platforms that will provide consumers with multiple communications, computing and entertainment products and services. In order to perform these functions, all such devices or platforms -- including personal computers, PDAs, wireless phones and set-top boxes -- need to utilize operating systems, whether installed in the device itself or residing on Internet servers. By maintaining and expanding its operating system monopoly across platforms, Microsoft can establish its position as "gatekeeper" to all such forms of communications, computing and entertainment services. And as gatekeeper, Microsoft will be in a position to direct customers using these platforms toward its services, to degrade or block access to competitors' services, and to impose costs on those competitors it cannot completely eliminate. By controlling all of these communications gateways, Microsoft will not only preserve its operating system monopoly against all serious threats, it will substantially lessen competition in the provision of innovative new "convergent" services.
For example, competition is now growing to reach consumers, through "gateway" devices such as PCs or television set-top boxes, with broadband communications signals that can carry everything from TV programming to Internet content to telephone conversations. An estimated 10 million American homes may use such devices next year and 25 million by 2006. See Byron Acohido, Challenging Microsoft? It Could Take Moxi, USA Today, Jan. 16, 2002, at B-3. Microsoft has already announced that it is developing an extension to Windows XP that will allow PCs to function in this manner. Id.; Microsoft Unveils New Home PC Experiences with "Freestyle" and "Mira" (Jan. 7, 2002), at www.Microsoft.com/presspass/Press/2002/Jan02/01. Unfettered by the proposed settlement, Microsoft can thus use its illegal operating system monopoly to become the literal communications gateway into and out of the American home or office. It then will have enormous power over the products and services consumers use to communicate with each other, to do their work and to entertain themselves.
In this memorandum, SBC addresses the numerous ways in which the proposed settlement fails to meet a paramount goal of relief in this case: To "pry open to competition" in the PC operating system market that Microsoft has dominated for over a decade by using blatantly exclusionary tactics.
The following facts are now beyond dispute in this proceeding:
First, Microsoft's monopoly has been extraordinarily durable, having prospered for over a decade (D.Ct. at ¶ 35), having increased steadily to over a 95% share even during the litigation (CA at 54; D.Ct. at ¶ 35), and having enjoyed the continuing protection of significant barriers to entry. See CA at 54-56 ("Because the applications barrier to entry protects a dominant operating system irrespective of quality, it gives Microsoft power to stave off even superior new rivals"); D.Ct. at ¶¶ 36-44, 61 ("Microsoft could significantly restrict its investment in innovation and still not face a viable alternative to Windows for several years. . . .").
Second, Microsoft's monopoly has created not only the power, but also the incentive, to exclude competition: every technological innovation that emerged to challenge Microsoft's dominance was met with a successful strategy of anticompetitive exclusion. Microsoft was able to "extinguish," perhaps permanently, the two greatest innovative threats to its dominance that arose in the 1990's -- Netscape and Java. CIS at 16-17; see also CA at 76-80 ("Microsoft's ultimate objective was to thwart Java's [and Netscape's] threat to Microsoft's monopoly;" it adopted as strategic goals to "kill cross-platform Java" and interfere with the ability of Netscape's browser to interoperate with Microsoft products); D.Ct. at ¶¶ 68-77. So long as Microsoft retains the power and incentive to exclude the competitive threats of the 21st century, economic theory predicts, and history demonstrates, that it will seek to evade any regulatory barriers placed in its path. Thus, the prospect of innovation offers no solace to restoring competition, only a sure target for Microsoft's exclusionary conduct.
Third, Microsoft's incentive to engage in calculated predation is so strong that it readily harms consumers and degrades its own products to achieve anticompetitive exclusion. D.Ct. at ¶ 174 (finding that by commingling "Microsoft has unjustifiably jeopardized the stability and security of the operating system"), ¶¶ 408-12 (highlighting harm inflicted upon consumers); CA at 62, 65 (affirming district court findings of consumer harm). It is also revealed in a "take no prisoners" approach in which deception, threats, attempts to conspire and degradation of middleware connections were used to stifle competition. CA at 73, 75-77. Nothing in the foreseeable future, much less in the monopoly maintenance record, suggests that marketplace or technological developments alone will suffice to curb Microsoft's market power, its incentive to exclude and its proven ability and willingness to do so ruthlessly.
Finally, Microsoft's monopoly affects the country's most powerful engine of national economic prosperity and productivity -- the processing and communication of information. Where monopolization has injured industries of comparable importance, the future of competition has never before been entrusted to illusory promises by the offending firm or to uncertain marketplace developments, unprotected by judicial supervision from recurrent acts of exclusion. See United States v. Am. Tel. & Tel. Co., 552 F. Supp. 131, 215-17 (D.D.C. 1982), aff'd sub nom. Maryland v. United States, 460 U.S. 1001 (1983) ("AT&T") (rejecting proposed consent decree and ordering its modification based, in part, on the "complexity and magnitude" of the decree and the decree's effect "on the largest corporation in the world . . . the entire telecommunications industry, the computer industry . . . and thus the interests of literally millions of individuals").
That is why in monopolization cases the law demands that relief must decisively end the anticompetitive practices, prevent their recurrence and extension into new markets, and restore competition. "Antitrust relief should unfetter a market from anticompetitive conduct and `pry open to competition a market that has been closed by defendants' illegal restraints.'" Ford Motor Co. v. United States, 405 U.S. 562, 577-78 (1972) (citation omitted). If a decree does not effectively pry a market open to competition, "the Government has won a lawsuit and lost a cause." Int'l Salt Co. v. United States, 332 U.S. 392, 401 (1947). To restore competition, therefore, the relief must take account of all the factors relevant to the offense, including in particular the likely duration of the monopoly power, which, of course, is the wellspring of the incentive as well as the ability to exclude. See Ford Motor Co., 405 U.S. at 575 (affirming ten-year ban on Ford's manufacture of spark plugs; prohibition was a "necessary step toward the restoration of the status quo ante" in the market).
The government has repeatedly embraced the foregoing standards in this case (see, e.g., Govt. D.Ct. Memo. at 24; CIS at 3), but its proposed settlement fails their purposes. The government has abandoned, without explanation, injunctive relief that it urged upon the District Court as essential to curb Microsoft's appetite for anticompetitive conduct and has agreed to a decree filled with loopholes. For example, although the Court of Appeals found commingling of browser and operating system code to be unlawful acts of monopoly maintenance, and the government advocated that such commingling be prohibited as "an especially potent competitive weapon for Microsoft . . . to target competing middleware threats," Gov't D.Ct. Reply Memo at 61, the proposed settlement does not prohibit such conduct. Similarly, although the Department of Justice Antitrust Division Manual provides that the government "should not negotiate any decree of less than ten years' duration" and the government in this case objected to Microsoft's initial proposal for a four-year decree because "there is no sound justification for entering a decree of shorter duration," the remedies in the proposed settlement are to last only five years.
The government's retreat from established antitrust policy and from its prior opposition to Microsoft's remedial proposals has grave implications for a competitive economy and for SBC. Not only is Microsoft allowed to repeat conduct, previously found anticompetitive, to protect its operating system monopoly from middleware sources of competition, but it is free to do so where the courts have already recognized an even more powerful threat exists, namely from the Internet. D.Ct. at ¶¶ 56, 59-60 (cited with approval in CA at 79). Since Internet servers can perform computing functions formerly accomplished only by PCs, networks of servers and PCs that freely interoperate (or "talk" to each other) -- regardless of the type of operating system software that they use -- are a platform for applications not dependent on Windows. This means that the combination of inexpensive computers or handheld devices (like a "dumb" PC, a cellphone, or a PDA) and smart server networks connected to the Internet can break the monopoly power of Microsoft's PC operating system by offering a server network alternative that will work with any operating system and provide more and better application choices at less cost. D.Ct. at ¶¶ 22-27 (cited with approval in CA at 52), ¶¶ 56, 59-60 ("[T]he rise of the Internet. . . has fueled the growth of server-based computing, middleware, and open-source software development. Working together, these nascent paradigms could oust the PC operating system from its position as the primary platform for applications development and the main interface between users and their computers.").
Yet nothing in the government's settlement prevents Microsoft from turning an open Internet into a closed Microsoft environment simply by doing two things: (1) commingling its browser, Internet Explorer ("IE"), with its Windows operating system; and (2) changing the protocol its browser uses to "talk" to Internet servers to an undisclosed proprietary standard that will only work effectively with Microsoft servers. Because of the dominance of Microsoft's browser (currently 91% of all browser usage), all web servers would then be forced to have a Microsoft server operating system in order for the servers, and the web sites they host, to be accessible to the vast majority of users. In turn, all consumers and businesses that wish to access the Internet will be forced to purchase a Windows operating system in order to utilize Microsoft's browser. Nothing in the decree prevents this scenario, because Microsoft is free to use its illegally maintained monopoly power to force servers to interoperate only with Windows, such that Microsoft becomes the Internet gatekeeper of a once open and competitive system. Microsoft's operating system monopoly would thereby become still more powerful and durable, as another threat to its dominance is destroyed. In this way, the applications barrier to entry that protects the Windows monopoly will extend to the Internet.
The reality of this threat for the future competitiveness of Internet-based businesses has a direct bearing on a wide range of Microsoft's potential and actual competitors, including SBC. Through its affiliates, SBC provides Internet access and Internet services to customers. SBC is currently developing several new Internet-based businesses, most importantly its Unified Messaging Service ("UMS"), which will compete directly with specific Microsoft products and services. UMS will allow retrieval of voice, e-mail and fax messages from anywhere in the world, using any computer or device running on any operating system. The proposed settlement, however, allows Microsoft to make SBC's UMS product significantly less competitive by taking the two simple steps outlined above. In these circumstances, only Microsoft server operating systems would be interoperable with the vast majority of other devices that access the Internet, and Microsoft would be able to use its server control to discriminate against its competitors.
As this example shows, the omissions and loopholes in the proposed settlement are of no small importance; they have drastic consequences for a competitive economy. So too does the decision to limit the settlement to only five years. The trial court recognized in findings sustained by the Court of Appeals that competitive alternatives to the Microsoft operating system, such as web portals, servers and middleware, take years to develop as viable threats, yet the proposed decree ends almost as soon as it starts -- in only five years overall, with some provisions in effect for only four years. No sensible competitor would invest in technology improvements to the maximum extent necessary to challenge Microsoft -- innovations that require years to succeed absent predation -- when the decree is neither strong enough, nor long enough, to protect them. Yet the government breaks with its own policy of requiring decrees with ten-year terms, despite the fact that Microsoft's monopoly has existed for more than a decade and its unlawful conduct has spanned a period nearly as long.
Equally important, there is nothing in the decree to jump start competition, much less to "pry open" the monopolized market to give consumers the benefit of competition that would have existed from the likes of Netscape and Java had Microsoft's exclusionary conduct not "extinguished" them. See Schine Chain Theaters, Inc. v. United States, 334 U.S. 110, 128 (1948) (an injunction against future violations is inadequate when it allows the monopolist to retain its "unlawfully built empires"). Under the Tunney Act, the "public interest," see 15 U.S.C. § 16(e), is not served by a settlement that allows a monopolist to pursue conduct already adjudicated illegal, that leaves open easy escape routes from the proposed decree's proscriptions, and that utterly fails to restore competition to the monopolized market.
When, as here, there is an adjudicated record of serious competitive harm (monopolization) and wrongdoing (anticompetitive exclusion), the responsibility to protect the public from an inadequate settlement is high, and a reviewing court has broad power to do so. AT&T, 552 F. Supp. at 151-53. As the District Court has said, "The Supreme Court has vested this court with large discretion to fashion appropriate restraints both to avoid a recurrence of the violation and to eliminate its consequences." United States v. Microsoft Corp., Civ. Nos. 98-1232, 98-1233 (CKK), Transcript of Proceedings at 9 (Sept. 28, 2001).
For the reasons set forth below, approval of the proposed decree cannot be squared with ten years of government litigation that culminated in resounding appellate holdings of major antitrust offenses. The fact that adverse antitrust consequences will result is clear from the face of the proposed settlement, as well as by comparison to the injunctive provisions defended by the government in its earlier proposed litigated judgment. In fact, adoption of this proposed settlement would be worse than no decree at all, for its negotiated omissions and concessions allow conduct found illegal in the past to continue -- such as commingling of code -- and thus would appear to prevent even the government from attacking such decree-sanctioned behavior during its term. Such ambiguity surrounding the government's enforcement intentions is in itself affirmatively harmful to the public interest.
The proposed settlement in this case must be evaluated in light of the Court of Appeals' affirmance of the District Court's conclusion, supported by an overwhelming factual record, that Microsoft is guilty of a panoply of illegal activities to maintain and extend its monopoly in the market for Intel-compatible PC operating systems. Microsoft's conduct inflicted significant harm on consumers and competition in violation of Section 2 of the Sherman Act. CA at 50-80.
Microsoft is the world's largest supplier of computer software for PCs and, in particular, dominates the market for Intel-compatible PC operating systems software world-wide. Although it has the second-largest market capitalization among American companies, Microsoft's importance extends beyond its financial success, because it is a linchpin of the computer industry (including hardware, peripherals, software and data services), and the computer industry is critical to the functioning of a competitive American economy. See, e.g., Henderson Decl. ¶¶ 87-98; Romer Decl. ¶ 17.
Microsoft's operating systems monopoly is an enduring one, persisting for over a decade despite what the Court of Appeals has described as a "technologically dynamic market." CA at 49. Over that same period, the government has been forced to spend resources on a continuous basis to investigate, and then to prosecute, Microsoft for its illegal conduct. The FTC began investigating Microsoft's acquisition and maintenance of monopoly power in the operating systems market in 1990, although it did not bring charges against the company. United States v. Microsoft Corp., 56 F.3d at 1448, 1458 (D.C. Cir. 1995). Using the FTC's extensive investigation file as a starting point, the Antitrust Division of the Justice Department initiated its own investigation, and in July 1994 filed a civil complaint under Sections 1 and 2 of the Sherman Act, charging, inter alia, that Microsoft unlawfully maintained a monopoly of operating systems for Intel-compatible PCs. Id. That case was settled by a consent decree, thereby avoiding trial on the merits.
Three years later, the Justice Department filed a civil contempt action against Microsoft on the ground that it had violated the decree. On appeal from the grant of a preliminary injunction, the Court of Appeals ruled that Microsoft had not violated the relevant provision of the consent decree, but reserved the question of whether the company's bundling of Internet Explorer with the Windows operating system violated the antitrust laws. United States v. Microsoft Corp., 147 F.3d 935, 950 n.14 (D.C. Cir. 1998). The complaint that gives rise to the instant proceeding was filed in May 1998 by the Justice Department and a group of State plaintiffs, again alleging, inter alia, unlawful maintenance of a monopoly in the PC operating system market in violation of Sherman Act § 2. CA at 47.
The Court of Appeals affirmed the District Court's finding that Microsoft's Windows operating system accounts for over 95% of the Intel-compatible PC operating system market. CA at 54. As the District Court found:
D.Ct. at ¶ 35.1
The Court of Appeals held that not only was Microsoft's operating system monopoly virtually complete as measured by market share, but also that the monopoly's increasing power and scope derives from a structural barrier -- the "applications barrier to entry" -- that protects the company's future monopoly position even as against superior rivals. The Court held that this barrier to entry
CA at 55 (citations omitted). The Court of Appeals went on to hold that even if Windows may have gained its initial dominance through superior foresight or quality, Microsoft had maintained its position through means other than competition on the merits. "Because the applications barrier to entry protects a dominant operating system irrespective of quality, it gives Microsoft power to stave off even superior new rivals." CA at 56.
The Court of Appeals affirmed the District Court's findings regarding a variety of anticompetitive acts by Microsoft that were designed to maintain its monopoly by preventing the effective distribution and use of middleware products -- including Netscape's "Navigator" browser and the Java cross-platform technologies -- that might threaten the Windows operating system monopoly. The Court of Appeals noted with approval the District Court's conclusion that Microsoft's monopoly gives the firm incentives to perpetuate the monopoly by a pattern of exclusionary conduct. CA at 58. As the District Court concluded, "over the past several years, Microsoft has comported itself in a way that could only be consistent with rational behavior for a profit-maximizing firm if the firm knew that it possessed monopoly power, and if it was motivated by a desire to preserve the barrier to entry protecting that power." D.Ct. CL at 37.
In its successful efforts to thwart Netscape and Java, Microsoft demonstrated its ability to extinguish competitive threats to its monopoly as fast as they emerged in a rapidly changing technological environment. Microsoft's conduct also evidenced a remarkable willingness to hurt consumers and degrade its own products where necessary to accomplish the exclusion of competitive threats to its dominance.
Both Netscape and Java threatened to facilitate competition in operating systems by permitting software applications developers to write programs for the application programming interfaces (APIs) exposed by these middleware products, which in turn were capable of running not only on Windows, but on other operating systems. If such middleware were permitted to thrive, such "cross-platform" applications would have the potential to overcome the applications barrier to entry upon which Microsoft's operating system monopoly rests. CA at 53, 60.
The Court of Appeals upheld the District Court's findings and conclusions that Microsoft engaged in the following unlawful conduct in violation of Section 2 of the Sherman Act for the purpose of maintaining its PC operating system monopoly:
Sun Microsystems created Java2, a type of middleware that would support all applications regardless of the operating system they were written for. CA at 74. Programs calling upon Java's APIs will run on any computer that itself is configured for Java; thus, Java enabled software developers to write applications programs that could be run on different operating systems with relative ease. In May 1995, Netscape agreed with Sun to distribute Java with every copy of Navigator, which at that time was the dominant browser. Microsoft violated §2 in three separate ways in a successful effort to extinguish Java as a competing middleware platform:
As the government acknowledges in the Competitive Impact Statement, appropriate injunctive relief here must accomplish three things: "(1) end the unlawful conduct; (2) `avoid a recurrence of the violation' and others like it; and (3) undo its anticompetitive consequences." CIS at 24 (citing Nat'l Soc'y of Prof'l Eng'rs v. United States, 435 U.S. 679, 697 (1978); United States v. E.I. du Pont de Nemours & Co., 366 U.S. 316, 326 (1961); Int'l Salt Co., 332 U.S. at 401); CA at 107. See also Gov't D.Ct. Memo at 24 ("Permanent injunctive relief ordered in a Sherman Act case must be both forward-looking and remedial. The decree must (i) end the violation, (ii) `avoid a recurrence of the violation' and others like it and (iii) restore competition to the market."). Any remedy must be broad in scope and prophylactic in nature so that competition is restored and Microsoft is effectively precluded from further exercise of its monopoly power, even as new products are developed and circumstances in the market change.
CA at 103 (citations omitted) (quoting Ford Motor Co., 405 U.S. 562; United States v. United Shoe Mach. Corp., 391 U.S. 244 (1968)). In the proceedings on remand, the District Court has already recognized that any remedy, in order to be adequate, must go beyond merely prohibiting the conduct in which Microsoft has previously engaged:
Microsoft, Transcript of Proceedings at 9 (paraphrasing Nat'l Soc'y of Prof'l Eng'rs, 435 U.S. at 697; and United States v. U.S. Gypsum Co., 340 U.S. 76, 88-89 (1950)).
The public interest is not served merely by eliminating past anticompetitive practices; the remedy must eliminate the future recurrence of illegal conduct:
U.S. Gypsum Co., 340 U.S. at 90 (citations omitted); see also United Shoe Mach. Corp., 391 U.S. at 252 (relief should "render impotent the monopoly"); Nat'l Soc'y of Prof'l Engn'rs, 435 U.S. at 697 ("the District Court was empowered to fashion appropriate restraints on the Society's future activities to avoid a recurrence of the violation and eliminate its consequences").
In this case, the government has recognized the need to go beyond enjoining current violations to assure that Microsoft's violations do not recur. See Gov't D.Ct. Memo at 24 ("Forbidding the continuance of the violation -- here, for example, the anticompetitive bundling of Internet Explorer with the Windows operating system -- is necessary but not sufficient to rectify the harm caused and threatened by Microsoft's illegal conduct.").
As the government has acknowledged, "[r]estoring competition is the `key to the whole question of an antitrust remedy.'" CIS at 24 (quoting E.I. du Pont de Nemours & Co., 366 U.S. at 697); see also U.S. Gypsum, 340 U.S. at 90 ("The conspirators should, so far as practicable, be denied future benefits from their forbidden conduct."); United States v. Grinnell Corp., 384 U.S. 563, 577 (1966) ("We start from the premise that adequate relief in a monopolization case should . . . deprive the defendants of any of the benefits of the illegal conduct, and break up or render impotent the monopoly power found to be in violation of the Act."); CA at 103 (a remedies decree must "deny to the defendant the fruits of its statutory violation") (citations omitted). As the Supreme Court put it in a holding that is particularly cogent here:
Schine Chain Theaters, Inc., 334 U.S. at 128.
A decree must "break up or render impotent the monopoly power found to be in violation of the Act." Grinnell Corp., 384 U.S. at 577. It must "leave the defendant without the ability to resume the actions which constituted the antitrust violations in the first place." AT&T, 552 F. Supp. at 150.
Because an antitrust remedy, in order to be adequate, must neutralize the monopolist's power to resume the action constituting the adjudicated violation, any remedy "must effectively foreclose the possibility that antitrust violations will occur or recur." Id. at 150. Again, the Supreme Court has given instruction that is directly relevant here:
Int'l Salt Co., 332 U.S. at 400. As the District Court has recognized, even practices not found to be unlawful should be prohibited where necessary to avoid recurrence of monopolization. AT&T, 552 F. Supp. at 150 n.80 (citing United States v. United Shoe Mach. Corp., 110 F. Supp. 295, 346-47 (D. Mass. 1953), aff'd, 347 U.S. 521 (1954)). Similarly, the court must impose additional restraints to allow development of new competition in the relevant market. Id. (citing Ford Motor Co., 405 U.S. at 575).
Given the record in this case, the remedy must anticipate new forms of exclusion such that, in view of Microsoft's incentive to exclude and demonstrated willingness to do so, the company may not further restrain trade illegally and is prevented from repeating its past unlawful practices in new contexts.
Where the monopoly in question is as powerful and persistent as that maintained over the last decade by Microsoft, there is a real danger that the monopolist will evade the particular provisions of any consent decree that is entered. In order to cope with the threat of regulatory evasion, antitrust judgments must contain broad proscriptions of anticompetitive conduct that will, by their generality, cover new forms of exclusion. See, E.I. du Pont, 366, U.S. at 1254 (An "injunction can hardly be detailed enough to cover in advance all the many fashions in which improper influence may manifest itself."); AT&T, 552 F. Supp. at 167 (approving consent decree ordering divestiture, preclusion from specific markets, and compulsory, royalty-free licensing) ("it is unlikely that, realistically, an injunction could be drafted that would be both sufficiently detailed to bar specific anticompetitive conduct yet sufficiently broad to prevent the various conceivable kinds of behavior that AT&T might employ in the future"); Zenith Radio Corp. v. Hazeltine Research, Inc. 395 U.S. 100, 132 (1969) (court may 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 be committed or whose commission in the future, unless enjoined, may fairly be anticipated from the defendant's conduct in the past"); CA at 103 (court must "ensure that there remain no practices likely to result in monopolization in the future"). The "broad power" the Court has to fashion an effective remedy includes the authority to prohibit exploitation of monopoly power in any manner and to order provisions designed to create and foster new competition, including the disclosure of proprietary information, mandatory licensing, exclusive dealing bans and many other remedies. Gov't. D.Ct. Memo at 26 (citing United States v. Crescent Amusement Co., 323 U.S. 173 (1944); Hartford-Empire Co. v. United States, 323 U.S. 386 (1945); United States v. Glaxo Group Ltd., 410 U.S. 52 (1973); Int'l Salt Co., 332 U.S. 392; Ford Motor Co., 405 U.S. at 572).
In order to prevent evasion of antitrust proscriptions put in place by a consent decree, courts routinely retain jurisdiction in order to modify decrees, resolve disputes, and ensure there is a forum for timely adjudicating whether defendants are in compliance. See, e.g., Int'l Salt Co., 332 U.S. at 401-02; Otter Tail Power Co. v. United States, 410 U.S. 366, 381-82 (1973); United Shoe Mach. Corp., 391 U.S. at 251-52; AT&T, 552 F. Supp. at 215-17 (ordering modification of proposed consent decree to include provisions relating to Court's continuing ability to enforce decree).
The applications barrier to entry that Microsoft enjoys through its operating system monopoly will, as the District Court found (and the Court of Appeals agreed), make it extraordinarily difficult for a new operating system to attract enough developers and consumers to be a viable alternative to Windows in any reasonable time frame. D.Ct. at ¶¶ 30-31; D.Ct. CL at 36; CA at 54-56. The overwhelming majority of consumers will only use Windows because there are already a large variety of applications written for that operating system. Given that it is expensive to port applications from one operating system to another, software developers will generally write applications only for the operating system that is used by the dominant share of PC users.
Software developers and ISPs are now forced, given the economics of the industry, to use Windows, an operating system that they would not necessarily choose, but that is virtually the sole conduit available to deliver their product to the end-user. Given these circumstances, "it remains to be seen whether server or middleware-based development will flourish at all." D.Ct. at ¶ 32.
In order to allow alternative operating systems to develop, the public interest demands a decree that will "pry open to competition a market that [is] closed" by the enormous applications barrier to entry and by Microsoft's continuous course of illegal conduct. See Int'l Salt Co., 332 U.S. at 401. Given the time necessary for a competitive operating system or middleware product to overcome the applications barrier to entry (if it is possible at all), any sustainable decree must assure consumers, programmers and potential competitors of a lengthy time frame in which to develop new products that can compete with Windows. Without an adequate time frame for competing products to take hold, consumers will be unwilling to scrap the investment in applications, training, and hardware that they have already made in Windows.
Without a decree that is broad enough to ensure that Microsoft does not continue to benefit from its past practices and erect new barriers to market entry, the very purpose of antitrust relief in monopolization cases will be thwarted. Without a strong and long-lasting decree, Microsoft's entrenched dominance and the threat of further exclusionary conduct will preclude entrepreneurs and other innovators from improving products and services. As the government has acknowledged, "an injunction which simply bars the precise illegal conduct proven at trial would leave the defendant with the full dividends of [its] monopolistic practices and profit from the unlawful restraints of trade which [it] has inflicted on competitors." Gov't D.Ct. Reply Memo at 10 (quoting Schine Chain Theaters, 334 U.S. at 128 (internal quotation marks omitted)).
If the decree leaves any room for doubt whether Microsoft will retain its freedom and power to exclude competitors, then software developers will, in their economic self-interest, continue what they have been doing for years -- writing applications that operate solely on Microsoft's platform -- thereby perpetuating the very monopoly that this case has found to be illegal. Such a result violates the fundamental tenet that an antitrust remedy must effectively "restore future freedom of trade." See U.S. Gypsum, 340 U.S. at 90 (reversing an injunction limited to sale of gypsum board in Eastern United States and directing entry of injunction covering all gypsum products throughout the country because the "relief, to be effective, must go beyond the narrow limits of the violation"); see also Glaxo Group Ltd., 410 U.S. at 64 (ordering compulsory patent licensing on appeal where necessary to assure "the public freedom from . . . continuance of the illegal conduct").
Rather than being narrowly drawn, the remedy in this case must be broad, prophylactic, flexible and forward-looking in order to provide competition a safe harbor from Microsoft's exclusionary power.
Pursuant to the Tunney Act, 15 U.S.C. § 16, in evaluating an antitrust settlement, a court may not "rubber stamp" a proposed consent decree, but must instead "make an independent determination as to whether or not entry of a proposed consent decree [is] in the public interest." Microsoft Corp., 56 F.3d at 1458 (quoting S. Rep. No. 298, 93d Cong., 1st Sess. 5 (1973)); accord AT&T, 552 F. Supp. at 149 & n.74.3
In determining whether the consent decree is in the public interest, the Court must begin by defining the public interest in accordance with the antitrust laws, AT&T, 552 F. Supp. at 149 (citing S.Rep. No. 93-298 at 3; H.R. Rep. No. 93-1463 11-12), and ensure that the provisions of the decree will "preserve free and unfettered competition as the rule of trade." Id. (citing N. Pac. Ry. v. United States, 365 U.S. 1 (1958)). The consent decree's provisions must "break up or render impotent the monopoly power found to be in violation of the Act." Id. at 150 (quoting Grinnell Corp., 384 U.S. at 577) and "must leave the defendant without the ability to resume the actions which constituted the antitrust violation in the first place," id. Not only must the decree remedy past violations, "it must also effectively foreclose the possibility that antitrust violations will occur or recur." Id.; see also id. at 151 ("[I]t does not follow that courts must unquestionably accept a proffered decree as long as it somehow, and however inadequately, deals with the antitrust and other public policy problems implicated in the lawsuit.").
In its first decision involving Microsoft, the Court of Appeals recognized that a more deferential review standard is appropriate under the Tunney Act in cases where there has been no trial and hence "there are no findings that the defendant has actually engaged in illegal practices." Microsoft Corp., 56 F.3d at 1460-61. It follows, therefore, that where there are express findings based on a full trial record "that the defendant has actually engaged in illegal practices," id., a more intensive Tunney Act review is required. Accord AT&T, 552 F. Supp. at 152. In the instant case, there have been both a lengthy trial on the merits and exhaustive findings of illegal monopoly maintenance by Microsoft -- findings that the Court of Appeals expressly affirmed. Thus, unlike in more routine Tunney Act proceedings involving settlements without adjudicated findings of liability, the proposed consent decree in this case is subject to a more searching standard of review by the trial court. See also U. S. Gypsum Co., 340 U.S. at 89 ("[C]ourts should give weight to the fact of conviction as well as the circumstances under which the illegal acts occur. Acts in disregard of law call for repression by sterner measures than where the steps could reasonably have been thought permissible.").
The AT&T case provides strong support for applying a higher degree of scrutiny in this case than in the typical Tunney Act proceeding. In AT&T, while noting that ordinarily a degree of deference to the Department of Justice's view that a settlement is in the public interest is appropriate, the District Court held that such deference was not warranted where the court had heard "what probably amounts to well over ninety percent of the parties' evidence both quantitatively and qualitatively, as well as all of [the parties'] legal arguments." 552 F. Supp. at 152. The District Court thus concluded that it was "in a far better position than are the courts in the usual consent decree cases to evaluate the specific details of the settlement." Id. The Court of Appeals, in its first Microsoft opinion, embraced this distinction and specifically contrasted the AT&T consent decree proceeding with the first Microsoft decree, which was presented before any evidence had been taken. See Microsoft Corp., 56 F.3d at 1461.
The circumstances now justify a searching and demanding review of whether the decree is in the public interest. The settlement here is not before the Court "in the first instance," or even with "ninety percent of the parties' evidence" presented (as in AT&T,) but rather after a full trial on the merits and multiple findings that Microsoft violated the Sherman Act. The District Court now has before it all of the trial evidence, as well as Findings of Fact and Conclusions of Law, affirmed by the Court of Appeals, regarding the relevant market and Microsoft's illegal, anticompetitive conduct. The Court may therefore make a fully informed and independent determination concerning whether the settlement is truly in the public interest.
As in AT&T, close scrutiny of the settlement is also necessary because of its importance to the national economy. In refusing to narrow the scrutiny given the consent decree, the District Court in AT&T noted that given the "potential impact of the proposed decree on a vast and crucial sector of the economy and on such general public interest as the cost and availability of local telephone service, the technological development of a vital part of the national economy, national defense, and foreign trade, the Court would be derelict in its duty if it adopted a narrow approach to its public interest review responsibilities." AT&T, 552 F. Supp. at 152.
The proposed settlement here is of no less importance. This settlement has broad ramifications for the national economy, especially in technology development, and impacts millions of American consumers -- ramifications with little precedent in the history of antitrust jurisprudence. In such circumstances, the Court's careful, independent review is essential to ensure the decree serves the public interest.
Finally, the proposed settlement also requires heightened scrutiny because half of the States that joined in prosecuting the case do not agree that the settlement would protect the interests of their citizens. The government is now expressing views substantially inconsistent with its expressed positions at earlier stages of the case. Where elected representatives of the public are sharply divided on whether the settlement actually serves the public interest, any questions concerning whether the settlement is fair to the public must be subject to exacting scrutiny. "None of this means, of course, that the Court would be justified in simply substituting its views for those of the parties. But it does mean that the decree [should] receive closer scrutiny than that which might be appropriate to a decree proposed in a more routine antitrust case." AT&T, 522 F. Supp. at 153.
The Court of Appeals affirmed findings that Microsoft extinguished all tangible threats to its operating systems monopoly. CA at 79; D.Ct. at ¶¶ 68-77. The findings also support the conclusion that if Microsoft had pursued competition on the merits rather than anticompetitive conduct, significant erosion of its monopoly would have occurred. See generally CA at 58-79. Certainly, that is what Microsoft's CEO believed when he envisioned the Windows operating system being "commoditized" by Netscape. D.Ct. at ¶ 72. The proposed settlement does nothing to deprive Microsoft of either the "fruits" or the source of its successful strategy of extinguishing competition, nor does it restore to consumers the benefits of the choices that they would have had if Microsoft's illegal conduct had never occurred.
At this stage of the proceedings, the government states that its goal is merely to "restore the competitive threat that middleware products posed prior to Microsoft's unlawful undertakings." CIS at 3. These were, as the government admits, merely "nascent threats," id. at 24, 25, not the fully-developed alternatives that would have existed today but for Microsoft's conduct. The competitive threats to the Microsoft monopoly were stillborn, not as a result of fair competition but, as the government acknowledges, because of Microsoft's predation:
CIS at 16-17. Although the CIS acknowledges that merely prohibiting future instances of Microsoft's past exclusionary, monopolistic conduct is not sufficient to restore competition, in reality that is all the proposed settlement attempts to do, and even those minimal efforts are unavailing.
Indeed, in the earlier remedy proceedings, the government characterized Microsoft's view of appropriate relief (which the government has now largely adopted) as a "crabbed view of antitrust remedies:"
Gov't D.Ct. Reply Memo at 49. It is therefore ironic that the government now embraces in the proposed settlement many of the same substantive decree provisions it earlier dismissed as woefully inadequate.
Presaging the current dispute over remedies, the government stated in a pleading before the District Court almost two years ago:
Gov't D.Ct. Memo at 27-28. The government then went on to describe as potential middleware or platform threats to Microsoft's operating system monopoly such products and technologies as Microsoft's own Office suite; applications such as voice recognition software, media streaming technology and email programs; server operating systems (and the need for interoperability between PCs and servers); and non-PC devices such as PDAs and hand-held computers. See id. at 28-29.
A settlement such as this one, which limits itself to protecting the next generation of emerging threats instead of "prying open" the monopolized market (thereby effectively blessing the extinction of the first generation and the preservation of Microsoft's monopoly), cannot claim to serve even this minimal goal without anticipating and prohibiting, with both specificity and generality, the many ways in which Microsoft can thwart new forms of competition from novel or different technologies, such as those listed by the government.
In this regard, it is noteworthy that the Court of Appeals, like the District Court, found that Microsoft's commingling of its browser and operating system codes constituted illegal monopoly maintenance. CA at 64-67. Yet the settlement would allow such conduct to continue. And as long as such commingling is allowed, Microsoft has the power to prevent the next generation of computing on web and network servers, nascent or otherwise, from overcoming its operating system monopoly. Thus, the decree does not even bar "repetition of the precise acts in the precise contexts that were at issue in the trial." Gov't D.Ct. Reply Memo at 49.
One of the principal threats to the dominance of Microsoft's operating system monopoly was middleware, which refers to "software products that expose their own APIs." CA at 53; D.Ct. ¶¶ 28, 68. Since middleware exposes APIs for which software developers can write programs, it can provide a less time-consuming and cheaper means of writing applications that can run on various operating systems. Id. Anything that reduces the need to adapt, or "port," an application to competing operating systems threatens to overcome Microsoft's monopoly in the PC operating systems market by eliminating the applications barrier to entry. CA at 54-56; D.Ct. at ¶¶ 68-78.
Unfortunately, the provisions that address middleware are so limited and rife with exceptions as to be virtually meaningless. Sections III.C and III.H of the proposed settlement are inadequate in at least the following respects: (a) the definitions of key terms invite easy evasion and make Microsoft's compliance virtually discretionary; (b) while the settlement is fairly specific in limiting Microsoft's ability to restrict OEMs from promoting competing software, it is silent on a crucial tactic -- technological binding4-- that Microsoft was proven to have used to the same exclusionary ends; and (c) the settlement undermines its own purported goals by including exceptions to each prohibition that largely negate the relief ordered. In its narrowness, the settlement also ignores new products, the potential for future innovation, and novel methods by which similar anticompetitive results may be achieved. As such, the decree fails either to remedy past effects or prevent future anticompetitive acts from occurring.
The definition of "Microsoft Middleware" is of crucial importance because, if a program constitutes "Microsoft Middleware," Microsoft is then subject to requirements that it disclose the programming interfaces and communications protocols by which the middleware interoperates with the Windows operating system. The definition also triggers Microsoft's obligation to allow OEMs to re-configure the PC desktop to give purportedly equal access to competing middleware. See CIS at 17-18; RPFJ §§ III.C, III.D, III.E. As shown below, these disclosures and obligations are not adequate to accomplish their avowed purpose. The government's stated goal is to ensure the viability of the OEM distribution channel for competing middleware products and the ability of those products to achieve "seamless interoperability" with the Windows operating system. CIS at 38.
The proposed settlement defines "Microsoft Middleware" as software code that: (1) is distributed separately from the Windows Operating System Product; (2) is trademarked; (3) provides functionality similar to a Microsoft Middleware Product; and (4) has the code necessary to be considered a self-contained product. See RPFJ § VI.J. Because each element of this definition is too narrow or too easily evaded by Microsoft, the obligations that are triggered by the definition are largely illusory.
The first part of this definition bears directly upon Microsoft's practice of initially distributing a middleware product separately, then bundling it for sale with Windows, and finally binding it to the operating system. See D.Ct. at ¶¶ 155-74 (discussing employment of these tactics with IE). Binding, or commingling of source code was held by the Court of Appeals to be illegal conduct used by Microsoft to eliminate the browser threat. See CA at 64-67; D.Ct. at ¶¶ 159, 170-74. It is unnecessary technically and has no procompetitive justification. Id.
The practical effect of the settlement's definition, however, is to allow Microsoft to achieve the same anticompetitive results merely by omitting the first of the three steps mentioned above. Simply by bundling middleware applications with the operating system from the outset (so that they would not be "distributed separately"), Microsoft may render any provision regulating its use of "Middleware" a nullity. Because the settlement contains no limitations on bundling or commingling of Microsoft middleware with the monopoly operating system, the definition actually encourages Microsoft to engage in anticompetitive practices -- i.e., commingling of code -- in order to avoid application of the decree.
Second, the definition of "Microsoft Middleware" requires that the product must be trademarked. Simply by not seeking a trademark, Microsoft can ensure that its middleware will not be covered by the settlement's provisions.5 This means that Microsoft can distribute any product that may have other intellectual property protections, such as copyright or patent protection, but that is not trademarked, without the product being considered "Middleware." Of course, if Microsoft chooses to bind a product to the operating system and not distribute it separately, there would be no need to trademark the product.
The third requirement, which refers to the functionality of a "Microsoft Middleware Product," further limits the scope of the "Microsoft Middleware" definition. This is because the definition of "Microsoft Middleware Product" lists by name several products traditionally considered middleware, including Internet Explorer, Microsoft's Java Virtual Machine, Windows Media Player, Windows Messenger, Outlook Express, and their successors in the Windows operating system. See RPFJ § VI.K. But limiting the definition of Microsoft "Middleware" only to those products which in the past were distributed as middleware fails to account for future development of new products. In addition, the definition omits important existing products, such as Microsoft Office6 and Internet telephony products, that perform functions analogous to the listed "middleware" products.7
The fourth requirement is that the code must be "self-contained." This too encourages commingling of Microsoft middleware with the operating system, because it allows Microsoft to create cross-dependent products solely to avoid complying with the provisions applicable to "Middleware." If Microsoft is allowed to commingle the code for the products in such a way as to create cross-dependencies between the operating system and middleware (as it did illegally for IE), it can avoid compliance with many of the substantive provisions in the decree.8
In the CIS, the government explains that the definitions of "Microsoft Middleware" and "Microsoft Middleware Product" include the "functionality" of a number of existing Microsoft middleware products, including IE, Windows Media Player, and Outlook Express. See CIS at 17-20. What is not mentioned, however, is that the government previously advocated a definition of middleware that was truly based on the function of middleware and, as such, there was no need to distinguish between "Microsoft Middleware" and "Non-Microsoft Middleware." See Final Judgment § 7(q).9 Nor does the CIS discuss the fact that in the event that a particular item of software code fails to meet any one of the four definitional requirements in the settlement, it will not be regulated at all by sections III.C, III.D, and III.E of the decree. This is significant, because the definition as it stands now neither comports with the traditional definitions of middleware, nor with the way the courts in this case have used the term. See, e.g., CA at 53.
From the outset, the government has supported injunctive relief designed both to give OEMs control over how to configure the PCs they sell and to provide end-users with the ability to remove Microsoft middleware from their computers. See Gov't D.Ct. Reply Memo at 45-47, 60-64. Sections III.C and III.H rely on the definition of "Non-Microsoft Middleware" to identify the competing software products that Microsoft must allow OEMs to include on the Windows desktop if they so choose and to distribute to consumers. The intention was to open the OEM channel to distribution of competing software and thereby remove one of the barriers Microsoft had erected to protect its Windows monopoly. Indeed, the definition of "Non-Microsoft Middleware Product" encompasses those technologies that Microsoft "extinguished" (such as the Netscape browser) as it defines the products entitled to protection. Before a new program receives this protection, however, the settlement's definition of "Non-Microsoft Middleware Product" requires that at least one million copies of the product must have been distributed in the previous year. RPFJ § VI.N. This onerous requirement defeats the government's express purpose of giving new products an adequate chance at the OEM distribution channel.
The CIS asserts that this level of distribution is "minimal" and "necessary" so that Microsoft's affirmative obligations will not be triggered by "minor" or "non-existent" products. CIS at 20-21. There is no support in the record, however, or in antitrust law generally for the notion that only large competitors deserve protection. "Minor" new products, i.e., the nascent competition that the CIS claims will be restored, deserve protection no less than older, more significant ones. One thing that the history of the software industry proves is that some of the most popular products and services were created by the ingenuity of small firms working alone without means of distributing their products. Most, even with the OEM distribution channel opened to them, failed to distribute one million copies the first year on the market, and the CIS cites no evidentiary support for setting the distribution trigger at the extraordinary level of one million copies.
Through this definition, the settlement creates a major obstacle to new products or competitors being able to obtain wide release and distribution of innovative products. Moreover, it has the additional pernicious effect of allowing Microsoft ample time to develop and promote or announce a preemptive offering before the non-Microsoft product reaches the one million distribution mark.
The settlement defines "Windows Operating System Product" as a closed universe of past operating system products that is comprised of the software code of Microsoft's currently-distributed versions of its PC operating system, including Windows 2000 Professional and Windows XP Home and Professional, and their successors. RPFJ § VI.U. The definition also leaves in Microsoft's "sole discretion" the determination of what software code constitutes future versions of the Windows Operating System Product. Id. The CIS fails to explain why the definition in the proposed settlement does not establish an objective standard, but instead entrusts such determinations to Microsoft's "sole discretion." Additionally, rather than explaining how the definition impacts upon the objectives of the decree and why it was drafted in this way, the CIS merely states that the definition leaves "packaging" (read: bundling or binding) decisions in Microsoft's hands. CIS at 23-24.
The government fails to reconcile this definition with the Court of Appeals' finding that Microsoft utilized commingling of code to maintain its monopoly. Nor does it explain how the definition meets the government's avowed goal that the settlement put an end to Microsoft's past monopolistic conduct. The definition gives Microsoft incentives to integrate middleware into its operating system to avoid having middleware products classified as such.
Of particular importance for the future, the definition fails to take into account that Microsoft manufactures non-PC and non-desktop PC operating systems, such as an operating system for personal digital assistants (PDAs) and other handheld devices. These systems include Windows CE 3.0, Windows NT ® Embedded 4.0, Windows CE for Automotive, Windows 2000 with the Server Appliance Kit, Windows for Smart Cards, Windows CE .NET and Windows XP Embedded. Any settlement that serves the public interest must cover new products that Microsoft can and will use to protect its PC operating system monopoly. There is an extensive set of devices which are the target for these systems beyond PDAs and pocket PCs, including smart phones, smart TVs, gaming devices, web pads, Internet appliances, media appliances, digital cameras, printers, scanners, retail point of sale devices, Windows based thin-client terminals, set-top boxes, residential gateways, automobile computing systems, home servers, industrial control devices and smart cards.
In short, the proposed settlement's definition ignores both past and future operating system products. A proper definition of "Windows Operating System Product" would both recognize Microsoft's past product releases and include all Microsoft operating systems for any hardware device, including PCs, servers and handheld computing devices.10
The proposed settlement effectively endorses, through its silence, tactics previously employed by Microsoft to prevent OEMs from becoming an effective distribution channel for competing middleware products. Among the deficiencies in section III.C of the settlement are: (i) its failure to prevent Microsoft from binding middleware to its operating system; (ii) its failure to require Microsoft to set meaningful price differentials between "fully loaded" and "stripped down" (without Microsoft Middleware) versions of the Windows operating system that could "pry open the market" for competing bundles of software and middleware offered by OEMs and third-party customizers; and (iii) the inclusion of limitations and loopholes that undermine the purpose of the decree provisions.
The Court of Appeals affirmed the District Court's findings that Microsoft's commingling of the code for IE with the code for Windows and its refusal to allow end-users to remove the IE browser from the Windows desktop constituted exclusionary acts in violation of Section 2. See CA at 66-67. Binding the IE middleware product to the Windows operating system injured both Netscape and consumers by degrading the ability of Netscape to effectively interoperate with Windows, thus reducing consumer options in browser choice, and by ensuring that deletion of files containing browser-specific functions would also delete vital operating system routines, thus crippling Windows. CA at 65-66 (citing D.Ct. at ¶ 164). Microsoft's anticompetitive purpose so dominated its business decisions that it degraded its own products by binding, since commingling of code decreased the security and reliability of Windows. CA at 62, 65; D.Ct. ¶ 174.11
In response to these acts, the government initially advocated a prohibition against the binding of software to the operating system, in order to prevent
Gov't D.Ct. Reply Memo at 60-61 (emphasis added). Indeed, the government's chosen remedy on this issue in the Final Judgment not only required abolition of commingling, but required the price of Windows to be reduced in proportion to the amount of unbundled programming that was removed by an OEM:
See Final Judgment § 3(g). In the CIS, the government acknowledges that the Court of Appeals found that Microsoft unlawfully "integrated its web browser into Windows in a non-removable way while excluding rivals," CIS at 3, but then makes no further mention of the commingling issue.
Notwithstanding the government's stated conviction (backed by the Court of Appeals' holding) that binding violates Section 2, the proposed settlement gives a green light to Microsoft's continuing to bind middleware products to its operating system. This gap in the settlement's coverage, coupled with the definitions of Microsoft Middleware and Microsoft Middleware Product, not only allows Microsoft to continue its past anticompetitive conduct, but also provides Microsoft with an incentive to use the same techniques to extend its monopoly into other areas.12
The settlement's failure in this respect is underscored by Microsoft's recent introduction of Windows XP, which plainly demonstrates its intent to continue defending the Windows monopoly by binding even more applications and services to its new operating systems, notwithstanding the determination that doing so is illegal. Windows XP has more Microsoft middleware products and services bound to or included with the operating system than any previous version of Windows. One of the services integrated into XP is Passport, a web authentication, security and credit card verification service that allows consumers, using a single log-in, to shop on thousands (and ultimately, Microsoft hopes, millions) of websites that accept Passport. Because Microsoft's past unlawful conduct allowed it to maintain a PC operating system monopoly and acquire a de facto monopoly in the browser market (IE is used to access the Internet by approximately 91% of consumers),13 Microsoft is in a uniquely advantaged position to encourage subscription to Passport whenever a user connects to the Internet from her XP desktop. This is so because XP comes fully loaded with prominently displayed prompts for Passport throughout the program, starting with the initial boot sequence and continuing each time the user logs on to her computer.
As Microsoft succeeds in generating Passport subscriptions through its monopoly distribution of Windows XP, retailers with web portals selling products and services on the Internet will be forced to accept Passport as their authentication system. In this way, Microsoft will be able to nullify threats to the Windows monopoly by precluding other web-based alternatives to Passport. Furthermore, by defending its PC monopoly with Passport, Microsoft will also insert itself on both sides of a web transaction. Because of the "network effect," the final outcome -- absent strong and effective injunctive relief -- is likely to be that most e-commerce will be conducted with either the consumer or vendor, or both, paying a fee to Microsoft for the use of Passport.
To the extent Passport gains a foothold as an authentication gateway to Internet commerce, this will erect a new barrier to entry for competing operating systems. Consumers will be reluctant to switch to a non-Windows PC operating system, because the personal information stored on Passport is readable only by Microsoft web servers, which in turn can be designed to interact most effectively with the Windows operating system and its embedded middleware, such as IE. At the same time, by erecting a fence (Passport) between PC users and the Internet generally, Microsoft will make it far less likely that a competing middleware platform, such as Netscape's Navigator, will displace user dependence on Windows, because without Passport, Navigator may end up being of little utility for e-commerce.
Nothing in the proposed settlement would prevent this chilling repetition of Microsoft's monopolizing conduct. By failing to adequately address the old tactics used (binding middleware to the operating system) and limiting the scope of the remedy in a manner which excludes new products and services, the proposed settlement fails in a critical way to end Microsoft's monopolizing conduct, let alone to deny Microsoft the fruits of its PC monopoly.14
As the Court of Appeals held, there is an economic disincentive for OEMs to offer, install and service a second middleware product such as a browser. CA at 66. However, nothing in the proposed settlement provides OEMs with an economic incentive to become a viable and effective means of distribution for alternative middleware products. Only by requiring Microsoft to provide OEMs with an economically-viable "stripped-down" version of Windows -- including the ability to completely remove Microsoft middleware from the operating system, see Final Judgment § 3(g)(i) -- will OEMs ever have an incentive to offer users products containing Non-Microsoft middleware alternatives.
Even if Microsoft were required to provide OEMs with an unbundled operating system, it would only be possible for OEMs to offer consumers a choice of an alternative middleware/software package for the PC if Microsoft's price to the OEM were reduced to reflect the lower value of a software package that does not include Microsoft middleware that the OEM wishes to replace with competing products. Put another way, a market for alternative middleware configurations will only arise if such alternatives can be priced competitively with the "fully loaded" version of Windows. If the cost of alternative middleware bundles is always higher than that of the Microsoft Windows bundle, the market for non-Microsoft middleware will be limited or nonexistent.
OEMs must have more than the Hobson's choice of either buying Windows XP fully bundled at $200, for example, or paying $199 for a stripped down version of Windows and then incurring the additional capital and labor costs of replacing a Microsoft middleware product with a competing product bought at an additional, separate cost.
Consequently, any remedial proposal that seeks to open the OEM distribution channel to competing middleware must address the pricing of the Windows operating system. The Final Judgment recognized the need for such a pricing mechanism. It required that the price of versions of Windows from which Microsoft middleware functions had been disabled or removed be reduced in proportion to the relative amounts of computer code bytes found in the operating system and middleware products in question. See Final Judgment § 3(g)(ii). Alternative formulations based on the relative product development costs are also available. See Litigating States' § 1.
Connected to the pricing issue is the failure of the proposed settlement to allow any party that is not an OEM (§ III.C) or end-user (§ III.H) to alter the configuration of the Windows platform. This omission has the effect of preventing third parties, who might fill a niche as customizers, to directly offer OEMs or end-users specific software/middleware packages that could be added to a stripped-down Windows operating system. For example, it is likely that absent Microsoft's illegal binding of its middleware to the Windows operating system, an industry of independent bundlers specializing in the sale of customized software packages would have developed. Using the operating system as a platform, these vendors could create customized software/middleware packages based on the need of particular consumer market segments, such as stock market buffs, antiques dealers or mathematicians.15
As it stands now, the proposed settlement creates no incentive for OEMs to pursue any of the objectives of section III.C. Yet, if the OEM distribution channel is not reopened, the decree will have no chance to succeed in its most important goal -- to restore competition in the monopolized market -- as no ISV will have equal access to consumers.
Although the government originally supported straightforward remedy provisions governing OEM flexibility as to what products could be offered with a PC operating system, it now retreats to complicated provisions whose limiting language undercuts the purported relief. Compare Final Judgment § 3(a)(iii) with RPFJ § III.C. When the government initially proposed provisions that would allow OEMs to reconfigure the products they offered to meet consumer demand free from Microsoft's restrictions, it stated:
Gov't D.Ct. Memo at 39. In response to Microsoft's objections, the government reiterated that the purpose of the provisions was to
Gov't D.Ct. Reply Memo at 45-46.
Notwithstanding the logic of the government's past proposals, the proposed settlement replaces clarity with ambiguity and loopholes. Section III.C.1 states that "Microsoft may restrict an OEM from displaying icons, shortcuts and menu entries for any product . . . to products that provide particular types of functionality," but nowhere defines "functionality." Without such definitions, Microsoft is free to decide what categories of middleware "functionality" qualify for display. Thus, nothing prevents Microsoft from excluding non-Microsoft middleware products for which no Microsoft counterpart exists -- an obvious deterrent to competing middleware products that are more innovative than Microsoft's own products.
Section III.C.2 ostensibly allows OEMs to distribute and promote non-Microsoft middleware through the display of shortcuts on the Windows desktop, but provides that the provision will apply only "so long as such shortcuts do not impair the functionality of the user interface." However, by never stating who determines when the "functionality" of Microsoft's operating system is impaired, the provision gives Microsoft free reign to decide which non-Microsoft products may be promoted by an OEM.
Section III.C.3 permits OEMs to configure competing middleware products to launch automatically at the conclusion of the initial boot sequence or upon connection or disconnection from the Internet. CIS at 31. It also appears to prohibit ISVs and OEMs from palming-off competing products by imitating Microsoft's trade dress. Nonetheless, the ambiguous wording of the provision would let Microsoft decide, in the first instance, which competing products may be displayed and what form the user interfaces (e.g., icons) may take. Moreover, as in § III.C.1, the provision's benefits are tied to a "functionality" determination made by Microsoft. The automatic launch of competing Middleware is only assured "if a Microsoft Middleware Product that provides similar functionality would otherwise be launched automatically at that time," which would again limit the settlement's reach to products with which Microsoft already competes.16
Subsection 5 allows OEMs to configure the Windows desktop to promote a non-Microsoft Internet access provider ("IAP") in the initial boot sequence. The provision is problematic for two reasons. First, it permits Microsoft to require that such offers meet "reasonable technical specifications established by Microsoft," which are never defined. Second, because it refers only to IAP offers, the proposed settlement prevents OEMs from offering any other type of product or service in the initial boot sequence. In striking contrast, the initial boot sequence for Windows XP offers a wide range of Microsoft products and services, including Passport, Hotmail, Instant Messenger and Internet telephony. Competition cannot be restored unless all competing middleware products, not just IAPs, are put on equal footing with Microsoft products. Because the proposed settlement allows Microsoft to retain the advantages of its operating system monopoly in the boot sequence by having an exclusive chance to promote its products and services, it fails to serve the public interest.
Finally, nothing in the proposed settlement discusses OEMs' ability to offer an alternative desktop. Prior to Microsoft's prohibiting the practice, OEMs would change the appearance of the desktop in ways they found beneficial. D.Ct. at ¶ 214. Some OEMs replaced the Windows desktop with a user interface of their own design or one that conformed with that of the OEM's selected browser. CA at 62-64. The government previously advocated a provision in the Final Judgment that assured OEMs the ability to offer an alternative to the Windows desktop, subject to the proviso that an OEM may not completely block access to the Windows desktop. See Final Judgment § 3(a)(iii)(3) (OEMs may "display any user interfaces, provided that an icon is also displayed that allows the user to access the Windows user interface").17 In the CIS, however, there is no explanation for the omission in the proposed settlement of this and other OEM configuration options that the government strongly advocated before the District Court and on appeal.
Section III.H of the proposed settlement purports to allow end-users the freedom to add and remove middleware as they see fit. In actuality, the provision fails to do so because: (i) Microsoft is never required to permit an end-user or OEM to remove a Microsoft Middleware product from the PC's memory, only to "disable" the functionality and "remove" the icon or other visual means of access; (ii) Microsoft continues to have full control over whether and when its products may override or launch in place of competing products; and (iii) the timetable for implementation renders the provision almost useless as a means of restoring competition.
The Court of Appeals held that Microsoft's removal of IE from the add/remove utility in Windows had the effect of reducing usage of rival browser products and violated Section 2. CA at 65. Loading the operating system with Microsoft middleware that cannot be removed imposes greater burdens on OEMs that choose to install competing middleware products.18 It also prevents consumers from receiving full access to the products and services of their choice. CA at 62, 65; D.Ct. at ¶ 174. Binding middleware products to the operating system also has a significant effect on the ability to remove Microsoft middleware, as it is difficult or impossible to remove the products without degrading Windows. CA at 66-67; D.Ct. at ¶ 159.
The government recognized that not being able to remove Microsoft middleware had the effect of "foreclosing customer choice and excluding competition," Gov't D.Ct. Memo at 6, and that Microsoft used this as a means of increasing the barriers to entry for middleware. Id. at 42; Shapiro Decl. at 25-26. Consequently, the relief initially requested by the government required that any Microsoft middleware product that was technologically bound to the operating system must be removable to create a "stripped down" version of Windows via the add/remove utility. See Final Judgment § 3(g)(i). No such requirement exists in the current settlement, however. Although the CIS states that section III.H "ensures that OEMs will be able to choose to offer and promote, and consumers will be able to choose to use, Non-Microsoft Middleware Products," CIS at 45, the government now discusses the provision in terms of "removing access" to the middleware product without explaining that "removing access" does not mean removing the product itself. Id.
Because section III.H of the proposed settlement fails to require Microsoft to enable OEMs and end-users to remove unwanted Microsoft middleware from Windows, it facilitates commingling of code, raises rivals' costs, and renders product substitution illusory.
On the subject of OEM/end-user control, the proposed settlement replaces a provision of less than fifty words in the Final Judgment with a series of interlocking provisions that run over six hundred words. Compare Final Judgment § 3(g)(i) with RPFJ § III.H(1-3). None of these limitations and exceptions were present in the interim relief the government advocated before the District Court previously, and the CIS is silent regarding the rationale for the avalanche of restrictions that it now proposes. Nor does the government suggest that the changes are needed in response to any holdings by the Court of Appeals.
What the proposed provisions do is create so many exceptions, limitations, and loopholes as to vitiate the broad pronouncements in the CIS. Two aspects of section III.H exemplify the manner in which the proposed settlement undermines its own efficacy: (a) permitting Windows to automatically ask an end-user if he or she wants to alter the computer's desktop configuration to restore Microsoft middleware that was previously removed by an OEM; and (b) permitting Microsoft virtually unbridled discretion as to when to override an end-user's selection of a default web browser or other middleware.
As discussed above, the proposed settlement does not allow OEMs or end-users to actually remove Microsoft middleware from their computers, instead limiting them to merely deleting icons and menu entries; the middleware itself remains physically in the computer, or in many cases, technologically bound to the Windows operating system. Even a conscious decision by an OEM or end-user to remove Microsoft icons and menu items is subject to interference by Microsoft under the proposed settlement. Section III.H.3(a) allows Microsoft to include in the Windows operating system a prompt that would ask the end-user, fourteen days after the initial boot up of the computer, for permission to automatically erase the OEM's or end-user's configuration of the system and reinstate the Microsoft middleware that was previously deleted. RPFJ § III.H.3(b).
This provision is troublesome for a variety of reasons, not least its Orwellian reminder of Microsoft's omnipresence. Most importantly, it allows Microsoft to undermine the configuration choices made by OEMs that may include significant promotion of competing middleware. It allows Microsoft to do this fourteen days after an end-user first boots up the computer, at a time when the end-user may not yet have gained a great deal of familiarity with the computer. Depending on how the question is asked and the user's level of sophistication, the user may not understand that he or she is removing the programs installed by the computer's manufacturer and replacing them with Microsoft products that may not work as well. Furthermore, the prompt is unnecessary, because if a user wanted a different configuration, she would be free to buy the computer from another OEM or purchase additional software on her own.
Nor is any limitation placed on the number of times Microsoft may "suggest" that the user alter the configuration. But regardless of how often Microsoft asks -- every day, every fourteen days, once a year, or only once -- the fact that it can raise the question at all not only undermines the OEM configuration, but also the goal of providing end-users with "a separate and unbiased choice with regard to each Microsoft Middleware Product or Non-Microsoft Middleware Product." RPFJ § III.H.1(b) (emphasis added); see also CIS at 48 (purpose of section III.H.3 is to prevent automatic alteration of OEM configuration, such as "sweeping the unused icons that the OEM has chosen to place on the Windows desktop"). There is no justification for permitting Microsoft to undercut this aspect of relief. Microsoft should be prohibited from ever prompting users to scuttle their OEM selections or desktop choices.
Although section III.H.2 of the proposed settlement ostensibly enables end-users and OEMs (and middleware producers themselves) to designate non-Microsoft middleware products (including web browsers) to be invoked automatically in lieu of a Microsoft product, loopholes and conditions destroy this provision's utility as a remedial device.
As an initial matter, the default election procedure is made reciprocal -- requiring that identical removal options be afforded Microsoft with respect to non-Microsoft middleware that would otherwise be the default. The government does not explain why such parity is being offered to an antitrust violator at the expense of those who have not violated the law.19
More troubling are the "Notwithstanding" clauses that follow subsection 3, which directly limit the benefits extended by section III.H.2. Part 1 of the first clause allows Microsoft to invoke a Microsoft Middleware Product if it is necessary for the computer to interoperate with a server maintained by Microsoft. RPFJ § III.H. Because so much middleware -- be it a web browser or a Java formulation -- now interacts with commercial web servers, which are to a large extent Microsoft web servers, the loophole created by this provision is enormous. As computing moves off the desktop onto Internet servers, communication with servers is becoming the norm. Moreover, because IE has captured over 90% of the market as a result of Microsoft's illegal conduct, Microsoft is now positioned to dominate the server operating system market by changing the protocols its browser uses to communicate with servers, from the current industry standard to its own proprietary protocols. This will leave those who host web servers with little choice but to use a Windows server operating system. See pp. 70-76 infra. Should this occur, the first "notwithstanding" clause of the proposed settlement's section III.H will allow Microsoft to override users' default browser selections in the vast majority of situations. The ultimate outcome will be that the illegal Windows monopoly will again be protected from the threat to its dominance posed by non-Microsoft web-based computing.
The second "Notwithstanding" clause in § III.H allows a Microsoft Middleware Product to launch if the designated Non-Microsoft Middleware Product fails to implement a "reasonable technical requirement (e.g., a requirement to be able to host a particular ActiveX control)." RPFJ § III.H. Because the proposed settlement leaves it to Microsoft to determine what a "reasonable technical requirement" would be, the loophole created by this provision is also enormous. To the extent the clause provides an example of such a failure to meet a technical requirement, the exception is overly broad. ActiveX is a programming environment that allows programs provided by servers to run locally on a PC inside the web browser. Its use replaces in part the cross-platform capabilities of Java and the open standard communication protocols used by most servers. Thus, by determining that the hosting of ActiveX is a "reasonable technical requirement," the proposed settlement ensures that anytime a Microsoft web product or service is launched or any product or service that relies on a Microsoft server is downloaded, Microsoft will be able to override a user's choice of browser. This provision grants Microsoft license to automatically override an end-user's browser choice when that user accesses a program or service that requires interaction with a Microsoft server. Far from restoring competition, this pernicious provision protects Microsoft's ill-gotten operating system monopoly from web-based competition.20
In addition to the foregoing serious deficiencies, the timetable in section III.H for implementation of the substantive provisions by itself renders the provision meaningless as a vehicle for restoring competition. Under the terms of the proposed settlement, section III.H will not be implemented until twelve months after submission of the settlement to the Court or at the release of the first service pack for Windows XP, whichever comes earlier. Because Microsoft is not bound by any of the provisions until that time, it has no incentive to release the first service pack prior to December 2002. The provision is thus rendered meaningless for a fifth of the lifespan of the decree.
Microsoft has no grounds to complain about burdens caused by making section III.H immediately effective. To the contrary, the year delay in implementation would reward Microsoft for its bad faith release of Windows XP, before a settlement was in place, with full knowledge that (notwithstanding the monopoly maintenance holding by the Court of Appeals) XP contains more bundled middleware, more commingled code, and more prompts for Microsoft-related products and services than any prior version of Windows. At the very least, the release of XP violates the spirit of the settlement by which Microsoft claims it is already abiding. No minimally adequate settlement would fail to provide relief to the marketplace as soon as practicable.
The proposed settlement also contains another glaring temporal loophole. The last paragraph of section III.H states that only the Microsoft Middleware Products that existed seven months prior to the last beta test on a new version of Windows will be subject to the requirements of the provision. This means that any new Microsoft product or service, developed six months or less prior to the date of the last beta test21 of a new Windows operating system release or major upgrade, would not be subject to the requirements that its icon or menu entry be removable from the operating system desktop or the requirement that the automatic launching of the product be disabled in favor of a competing middleware product.
The government offers no justification for a proposed settlement that guts the section III.H "removal" provision with myriad and, in some cases themselves anticompetitive, limitations and loopholes, and then delays their implementation for significant portions of settlement's proposed five-year duration.
In contrast to the current settlement's abundant accommodations to Microsoft, two years ago, the government categorically rejected Microsoft's complaints that it would be unable to comply with the "unbinding" provisions the government then advocated (i.e., requiring that OEMs and end-users have the ability to engage the Add/Remove utility to delete IE):
Gov't CA Brief at 132; Gov't D.Ct. Reply Memo at 63-64 (referring to its expert witness's ability to create a removal program that did not damage or degrade the operating system in a relatively short amount of time and the fact that Microsoft already provided a ready means of removing at least 80 components, many of which it considered "integrated" features of Windows). There is no reason, technical or otherwise, why the government should not insist upon timely and effective measures to prevent Microsoft from continuing to commingle its middleware with its operating system in blatant disregard of the Court of Appeals ruling.
In the earlier remedy proceedings, the government explained the indispensable competitive importance of "interoperability":22
Felton Decl. ¶¶ 51-52 (emphasis added). Indeed, full interoperability has long been recognized by Microsoft, quite correctly, to be a primary threat to its monopoly position in the PC operating system market because it would allow multiple, competing operating system platforms to perform essentially all the functions of a Windows PC. CA at 52-54; D.Ct. at ¶¶ 68-93; Henderson Decl. ¶¶ 12-18, 29-40; Shapiro Decl. at 20-21. Middleware, such as Netscape and Java, posed the initial competitive threat in using interoperability to shift computing away from the Windows PC (the "middleware threat").
But, as recognized in District Court findings cited with approval or undisturbed by the Court of Appeals, server-based computing, which would shift many computing tasks from the Windows PC to a server on the Internet, also poses a significant threat to the Microsoft operating system monopoly. D.Ct. at ¶¶ 24-27 (see CA at 52). By preventing full interoperability, however, Microsoft can neutralize the server threat.
Despite the government's claim that the proposed settlement achieved "seamless interoperability between Windows Operating System Products and non-Microsoft servers on a network" (CIS at 38), the proposed settlement would, in reality, enable Microsoft to withhold the disclosures necessary to achieve interoperability and thus defeat this goal. The proposed settlement would do nothing to achieve interoperability between a non-Microsoft server operating system and Windows or IE. Instead, the only disclosure requirements are both ineffectual and too narrow: They apply only to PC middleware and certain client protocols, disclosures which are insufficient, on their own, to create interoperability between a PC and a server or among servers. Thus, instead of preserving threats to the Microsoft monopoly from all sources, the settlement gives Microsoft a free shot at disabling server competition.
During the initial remedy proceedings before the District Court, the government recognized that new threats were emerging to Microsoft's monopoly in the PC operating system market. Gov't D.Ct. Memo at 29. The government recognized the likelihood that, in the future, most computing will be done through networks and on servers housed at remote locations, with personal computer use diminishing. The government acknowledged that, as a result, software for communicating with servers, operating systems for servers, and middleware designed to function on servers, had become a principal competitive threat to the Microsoft PC operating system monopoly. Id.; see also Henderson Decl. ¶¶ 13-16; Shapiro Decl. at 20-21.
The "server threat" arises from the following circumstances, as the government recognizes. Instead of using an expensive, "intelligent PC", which contains a Windows operating system, a substantial hard drive and a powerful microprocessor, consumers increasingly use simpler or smaller, more convenient devices, such as cell phones, PDAs (such as "Palm" or "Blackberry" devices), TV-set-top boxes, or "dumb PCs," all of which are typically equipped only with more basic (and often non-Microsoft) operating systems, a browser, smaller (if any) hard drives, and a microprocessor. D.Ct. at ¶¶ 22-27; Henderson Decl. ¶¶ 13-16, 91-92; Shapiro Decl. at 3-4. The consumer will then use this device's browser to connect to the network of servers on the Internet.
By accessing servers on the Internet, the consumer can perform most of the same computing functions (access/browse the Internet; word processing; e-mail; instant messaging, etc.) that are provided by a Windows PC, but at lower cost and much greater convenience. For example, under the PC computing model, to compose and spell-check a document, the PC's processor is used to process the relevant software program to perform the functions. Under the server-network computing model, however, the same function (compose and spell-check a document) is accomplished through the server, which processes the relevant software program and then transmits the document back to the PC. The PC operating system under this scenario does little more than transmit and receive the data. The actual computing functions are largely performed by the server's operating system and hardware. Similarly, far more complex applications can be offered on the Internet through high-powered servers effectively shared by thousands or millions of consumers.
The government recognized the overriding importance of server software and communications protocols in supporting the original Final Judgment:
Gov't D.Ct. Memo at 29 (emphasis added).
The government's expert in the remedy proceedings underlined the importance of the server-based computing model as a critical emerging threat to Microsoft's PC operating system monopoly. Rebecca Henderson, a professor from MIT with a doctorate in Business Economics from Harvard University, testified:
Henderson Decl. ¶¶ 14-15.
Microsoft recognizes the server threat to its PC operating system monopoly. D.Ct. at ¶ 60. Its strategy has been to use its monopoly control over the PC's operating system and IE to force websites on the Internet to use Microsoft server operating systems, even if they are otherwise not the most desirable choices. To do this, Microsoft can withhold disclosure of communications interfaces and protocols for IE, at the same time as it changes them from previously disclosed interoperable formulations. The objective is to make IE fully interoperable only with a Microsoft server operating system, and to restrict server-to-server communication only to Microsoft server operating systems. In addition, Microsoft can fail to disclose to competitors the server protocols that facilitate full interoperability between a Windows PC and a Microsoft server operating system, or between servers. As a result, Microsoft's server operating system will always interoperate better with a Windows PC operating system or Microsoft server operating system than any competitors' operating systems. These two actions, taken together, will enable Microsoft server operating systems to dominate the Internet, because website owners will need a server that interoperates with the more than 90% of all Internet users that use IE, while consumers will continue to buy the Windows PC operating system because of the applications barrier to entry. In this manner, Microsoft will easily defeat the threat that web-based computing poses to the PC operating systems monopoly by dominating server operating systems and server applications software. Any "dumb PC," cell phone or handheld device, which relies on a server on the Internet to perform the actual computing functions, will either have to use Microsoft operating system software or face elimination from the marketplace.
During the original remedy proceedings, the government's expert economist, Carl Shapiro, explained the importance of a powerful conduct remedy that would require Microsoft to provide timely disclosure of all APIs, protocols and other technical information necessary to allow all server operating systems to fully interoperate with a Windows PC operating system and Microsoft middleware, particularly IE:
Shapiro Decl. at 20-21 (emphasis added).
For this reason, the government proposed, and the District Court granted a remedy, requiring disclosure of "all APIs, technical information and Communications Interfaces" that enabled:
Final Judgment § 3(b)(iii).
The effect of this provision was to promote competition in the PC operating system market by using interoperability disclosure to support the server and middleware threats to Microsoft's monopoly. These crucial interoperability disclosure provisions required full disclosure of:
Microsoft's claim that remedies which affect the PC/server and server/server relationships are outside of the Sherman Act § 2 monopolization claims before the Court is insupportable. Defendant Microsoft Corporation's Remedial Proposal at 6-7 (Dec. 12, 2002) (brief filed in response to the Plaintiff Litigating States' Remedial Proposals). To the contrary, both sides presented evidence on this issue in the prior proceedings, and both this Court and the Court of Appeals were particularly concerned to ensure that the nascent middleware threats to Microsoft's PC operating system monopoly be protected from further anticompetitive conduct. See, e.g., CA at 79 ("it would be inimical to the purpose of the Sherman Act to allow monopolists free reign to squash nascent, albeit unproven, competitors at will -- particularly in industries marked by rapid technological advance and frequent paradigm shifts"); D.Ct. at ¶¶ 24-27, 56, 60.
Indeed, when Microsoft made the same argument during the original remedy proceedings, the government tersely exposed its fallacy:
Gov't D.Ct. Reply Memo at 49 (internal citations omitted).
The interoperability disclosure provision in the proposed settlement is seriously deficient in the following ways: (1) no interoperability disclosure protection is afforded to important competitive threats to Microsoft's PC operating system monopoly, including non-Microsoft operating systems for servers and embedded devices (i.e., cell phones, PDAs, set-top boxes)23; (2) the technical information that is required to be disclosed is too limited to be effective; (3) the timing of required disclosures is either too late or too vague; and (4) the definitions of major terms (API, operating system, middleware) would enable Microsoft to avoid disclosure to competitors, by claiming certain middleware or application products are part of the operating system.
The proposed settlement fails to provide essential disclosure of technical information necessary to ensure interoperability in at least four critical areas: (a) between Windows PC operating systems and non-Microsoft server operating systems; (b) between Microsoft middleware, particularly IE, and non-Microsoft server operating systems; (c) between Microsoft and non-Microsoft server operating systems; and (d) between Microsoft PC or server operating systems and non-Microsoft embedded devices. The absence of such protection effectively encourages Microsoft to dominate server operating systems and software in order to protect its PC operating system monopoly.
Although the government continues to espouse the public interest goal of "seamless interoperability" for servers, CIS at 38, the proposed settlement does not if fact achieve that result. The failure to ensure this essential remedial goal contrasts sharply with the District Court's findings of fact and conclusion of law, which were entirely affirmed or undisturbed by the Court of Appeals, establishing that Microsoft's conduct, in selectively disclosing or entirely withholding such technical information, plainly violated Sherman Act § 2. CA at 71-73; D.Ct. at ¶¶ 90-92, 338-40.
As late as November 2, 2001 -- four days before it reached the present settlement agreement with Microsoft -- the government still insisted that server interoperability was essential to any settlement. The government's settlement proposal on that date expressly required server interoperability disclosure:
Department of Justice, Proposed Final Judgment, Draft of November 2, 2001 at § III.E. (emphasis supplied). Pursuant to this provision, Microsoft would have been required to disclose both its client protocols and the server protocols which enable a PC and a server operating system to accept and transmit data to each other.24
The proposed settlement, however, deleted the requirement that server protocols be disclosed. This was accomplished by removing the words "or server" from the provision quoted above. RPFJ § III (E), November 6, 2001. This directly contradicts the view of the government's technical expert, who testified that if Microsoft were able to withhold from disclosure the server protocols:
Felton Decl. ¶¶ 53-57.
By removing the server protocol disclosure requirement, the proposed settlement virtually ensures that non-Microsoft server operating systems will never be viable, competitive alternatives to the Windows PC operating system monopoly. The client protocols that Microsoft is required to disclose will only allow the server to receive data or services from the PC. The other half of the transaction, whereby the server responds and sends data to the PC, cannot be accomplished without the server protocols. As a result, ISVs will not be able, on their own, to develop server operating systems that can fully interoperate with Windows PC. The government's expert has admitted that this eliminates the possibility that non-Microsoft servers would ever become a competitive threat to Microsoft's PC operating system monopoly:
Henderson Decl. ¶¶ 115-121 (emphasis added).
Quite simply, because Microsoft will not have to disclose any server protocols, this disclosure provision will not achieve "seamless interoperability" between a Windows PC and a non-Microsoft server operating system or aid in restoring even a vestige of competition to the PC operating system market.
During the original remedy proceedings the government acknowledged the crucial importance of requiring full disclosure of all technical information relevant to the interoperability between Microsoft's middleware products, particularly IE, and server operating systems.
As explained by a government expert, if Microsoft maintains control over the browser-server interaction (as it would under the proposed settlement), it can maintain its PC operating system monopoly by foreclosing the ability of a web server to interoperate with IE:
Henderson Decl. ¶¶ 81-86 (internal citations to trial record). The proposed settlement, however, completely abandons the disclosure provision necessary to prevent Microsoft from using its control of IE to eliminate demand for non-Microsoft server operating systems as a competitive alternative to Microsoft's PC operating system monopoly.
Under the proposed settlement, Microsoft has no obligation whatever to disclose any technical information -- APIs, communications interfaces or otherwise -- that would permit a non-Microsoft server operating system to interoperate with Microsoft's Middleware, including IE. The only disclosure obligation under the proposed settlement involving Microsoft Middleware requires disclosure of APIs relevant to interoperability with a "Windows Operating System Product," a term which is defined to include only "the software code...distributed commercially by Microsoft for use with a personal computer." RPFJ §§ III.D, VI.U (emphasis added). For example, this provision would not require any disclosure for the purpose of allowing competing server operating systems to interoperate with IE, the very product that the Court of Appeals held was used by Microsoft to illegally maintain its monopoly power. CA at 64-68.
As a result, under the proposed settlement, only Microsoft server operating systems will be guaranteed access to the proprietary APIs and communications interfaces necessary for a server to interoperate with IE. If a website owner purchases a non-Microsoft server operating system, the more than 90% of consumers who use IE on their Windows PC would be unable to access that website unless Microsoft had agreed to separately license the technical information required for interoperability.
Of equal significance is Microsoft's recent decision not to distribute Java as part of Windows XP. Java had been included in prior versions of Windows. Java is a software program that is an open industry standard; it allows websites both to operate on numerous non-Microsoft operating systems and to display rich colors and graphics to enhance the website's appearance. CA at 74-75; D.Ct. at ¶¶ 386-405. Approximately 50% of all websites currently on the Internet, including SBC's website, are Java-compatible. Microsoft dropped distribution of Java in favor of promoting ActiveX, which is Microsoft's proprietary software that competes with Java by allowing a web server to process an Internet-based application in a fashion similar to Java. ActiveX is a proprietary browser interface that is installed as part of the software code for IE. As a result, the only way a non-Microsoft server operating system can obtain the proprietary interfaces for ActiveX (or for Internet Explorer generally) is through a license from Microsoft.
If Microsoft chooses to make IE's protocols a completely undisclosed, proprietary standard, which it is free to do under the proposed settlement, Microsoft's PC operating system monopoly will be perpetuated, because the already formidable applications barrier to entry will be increased, and the server threat will be further diminished. This will occur as a result of two interrelated effects. First, website owners will be forced to purchase Microsoft server operating systems to ensure that their website remains fully accessible by the more than 90% of consumers who use IE. Second, for this reason among others, the vast majority of applications that are already written exclusively to interoperate with Windows will be increased, as ISVs' commercial need to write their Internet-based applications to be compatible with IE and ActiveX will increase. As time goes on, the number of servers which interoperate with Java and other browsers will continue to fall. Moreover, consumers who want to browse the Internet, that is, to access what will become the overwhelming majority of websites run on a Microsoft server operating system, will have to use IE and to get it, they will need a Windows PC operating system or another device that runs on Microsoft software. In the end, the prospect of "dumb PCs," cell phones and handheld devices equipped with non-Microsoft operating systems and browsers also will be eliminated.
SBC's own website provides an example of the exclusionary effect this lack of disclosure will have on non-Microsoft server operating system. SBC uses non-Microsoft operating systems on its website servers, and the website is designed to be compatible with Java. Because Windows XP (unlike earlier Windows versions) does not contain Java, when Windows XP users attempt to access SBC's website, they receive a message that "to display this page correctly, you need to download and install the following components: Java Virtual Machine." However, to make this (free) download on a normal dial-up connection will take the consumer over 30 minutes in normal conditions (i.e., low network congestion and latency), and an hour or more during times of peak usage. Thus, by dropping Java from Windows XP and failing to disclose its browser interfaces (which would enable SBC to obtain programs that could achieve "seamless interoperability" with Windows XP), Microsoft has compelled the consumer to undertake a confusing and lengthy download process. This creates an anticompetitive barrier to consumers' use of SBC's website and entrenches the Windows monopoly.25
Another corollary to the potential transition from the PC to a server-based computing environment is the need for a vastly increased volume of server-to-server communication transmissions. The proposed settlement contains no provision requiring disclosure of any technical information whatever to facilitate such communications -- "interoperability" -- between Microsoft and non-Microsoft server operating systems. The government offers no explanation for its absence, which will have the predictable effect of further diminishing the server threat to Microsoft's PC operating system monopoly, while also restraining competition in the server operating system market itself. All of the same deficiencies in interoperability discussed above with respect to PCs and servers also apply to server-to-server communications as well. For this reason, the same outcome is certain to occur: an overwhelming percentage of servers on the Internet will be forced to use a Microsoft operating system. If Microsoft is not required to disclose any of the technical information necessary interoperate with Microsoft's server operating system, the demand for non-Microsoft server operating systems will be significantly reduced. As a result, the Internet-based threat to Microsoft's PC operating system monopoly will be neutralized.26
Like a "dumb PC," an "embedded device" (such as a cell phone, PDA or set-top box) also can provide a viable, competitive alternative to a Windows PC. The government once again admitted this in prior remedy proceedings when it included embedded devices in the interoperability disclosure provisions:
Henderson Decl. ¶ 16. See Final Judgment § (3)(b)(iii) (requiring full disclosure of all technical information relevant to interoperability between operating systems of handheld devices and Windows PCs).
The government's position at the time rested squarely on District Court findings that such devices could present an alternative to a Windows PC in the future. D.Ct. at ¶¶ 22-23. Those findings were not disturbed on appeal and remain binding today. CA at 52 (handheld devices could, but do not yet, perform enough functions to be an alternative to a Windows PC.)
In the proposed settlement, this essential salutary provision has been removed without explanation by the government. The result is to eliminate another potential threat to Microsoft's PC operating system monopoly, while also giving Microsoft a significant advantage in the closely connected market for operating systems for such embedded devices.27
Even in those situations where the proposed settlement does require Microsoft to disclose certain technical information (interoperability between Microsoft middleware and Windows PC operating system; client protocols), the type and extent of the disclosure is inadequate to promote competition, because it fails to achieve the "seamless interoperability" that the government admits is essential to provide an effective remedy for Microsoft's antitrust violations.
While the government now claims to have achieved equal access to the "same interfaces and related information" for non-Microsoft and Microsoft middleware developers (CIS at 33), this is not correct. Specifically, the proposed settlement only requires disclosure of "the APIs and related documentation" used by Microsoft Middleware to interoperate with a "Windows Operating System Product." RPFJ § III.D. This limited disclosure is in stark contrast to the disclosure the government sought, and obtained, in the Final Judgment. Final Judgment § 3(b). At that time, the government required disclosure of "all APIs, technical information and communications interfaces" to achieve interoperability. Id. (emphasis added).
For example, by limiting disclosure to APIs, the government has left out important additional technical information that is indispensable for a middleware product to achieve "seamless interoperability" with a Windows PC operating system. In addition, the definition of what constitutes an "API" is drawn too narrowly in the proposed settlement; it does not include items like registry keys, file formats, communications protocols and other necessary technical information that is critical for an ISV to develop a middleware product that is fully interoperable. See RPFJ § VI.A. This stands in sharp contrast, once again, to what the government advocated in the Final Judgment, when its own expert explained that disclosure of all APIs, communications interfaces and other related technical information was essential to promote full interoperability. Felton Decl. ¶¶ 15-28. The government offers no explanation for its change of position. See Final Judgment § 7(b); Litigating States' § 22(c).
The same deficient definitions apply to the type of technical information that Microsoft must disclose to promote interoperability between a non-Microsoft server operating system and a Windows PC operating system. RPFJ § III.E. Such disclosure is now limited only to "client protocols," while the Final Judgment required disclosure of "all APIs, Technical Information and Communications Interfaces" necessary to achieve full interoperability. Final Judgment § 3(b); Litigating States' § 4. The proposed settlement's limited disclosure obligation has grave negative ramifications for ISVs seeking to achieve full interoperability. First, as stated above, both the client and server protocols are necessary to achieve interoperability between a PC operating system and a server operating system. Moreover, there is substantial additional technical information that far exceeds a "communications protocol." Such additional information includes APIs, software tools, file formats and other technical information without which a non-Microsoft server operating system will never achieve "seamless interoperability" with a Windows PC, let alone operate as well as a Microsoft server operating system.
The possibility that Microsoft will maintain its PC monopoly in this manner is not hypothetical. In fact, in the absence of strong remedial provisions, not only is Microsoft certain to use disclosure, or the lack thereof, to create and maintain an advantage over its competitors, but it has already used the very same control over communications protocols and the like to disrupt competing browsers' ability to communicate with its own servers over the Internet. At the time of the release of Windows XP just last October, Microsoft secretly changed the MSN web server program codes to specifically prevent the competing browsers Netscape Navigator and Opera from interoperating with the MSN web server.28Browser Bruiser, Chicago Sun Times, October 27, 2001, at 36 ("Microsoft's premiere web portal, MSN.com, denied entry to millions of people who use alternative browser software such as Opera and told them to get Microsoft's products instead."); MSN Shuts Out Other Browsers, Associated Press, October 28, 2001 ("Microsoft's premiere web portal, MSN.com, denied entry to millions of people who use alternative browser software such as Opera...The blockage coincided with Microsoft's showcase launch of its Windows XP operating system. Instead of getting MSN's news, games and shopping features, Opera users were given links to download Microsoft's browsers.").
To restore competition in the PC operating system market, proper timing is no less important than the substance of the required disclosures. The District Court made multiple findings of fact, affirmed by the Court of Appeals, which established that delayed disclosure of technical information to achieve interoperability effectively nullifies its value. D.Ct. at ¶¶ 338-40; CA at 71-73. The "time to market" in developing software is of the utmost importance. Id. The ability of an ISV even to attempt to compete with Microsoft is "highly dependent" on Microsoft's release of its technical information relevant to interoperability. Id. Netscape learned this lesson in 1995 when Microsoft, in the face of repeated demands from Netscape for technical information regarding interoperability with Windows 95, withheld this technical information from Netscape for approximately three months. D.Ct. at ¶¶ 90-92. While Netscape was waiting for this information, Microsoft brought to market its competing product -- Internet Explorer. Id. The result of this delay was to destroy any fair competitive challenge Netscape might mount against Internet Explorer. Id.
The government has previously acknowledged that the proper timing of disclosure of technical information related to interoperability is critical to restore competition to the PC operating system market. Gov't D.Ct. Reply Memo at 21-23. As a result, the government sought provisions in the Final Judgment that required the timing of all disclosures to be made when Microsoft disclosed the information to its own developers, and well in advance of when any new Windows product is brought to market. Final Judgment § 3(b).
Having manipulated for its own competitive advantage the timing of interoperability disclosures in the past, it is not surprising that Microsoft demanded very liberal and vague timing in the proposed settlement. The government, however, having litigated and prevailed on the timing issue, now largely gives up. The proposed settlement does not require the disclosure of technical information related to interoperability of Microsoft Middleware products to begin until the "earlier of the release of Service Pack I for Windows XP or twelve months after the submission of this Final Judgment." RPFJ § III.D. If Microsoft intends to introduce a new Middleware Product, it does not have to disclose any technical information related to interoperability until the product's "last major beta test." Id. All other disclosures must be made in a "Timely Manner," which is defined as the "time Microsoft first releases a beta test ... that is distributed to 150,000 or more beta testers." Id. § VI.R.
Similarly, the disclosure of client protocols contained in the Windows PC operating system are not required to begin until nine months after the submission of the final judgment, and all subsequent disclosures are not regulated as to time, and thus left solely within the discretion of Microsoft. RPFJ § III.E.
The inadequacy of these timing requirements is patent. There is no conceivable justification, and none has been offered, for delaying disclosure with respect to Microsoft's current products for nine to twelve months from the date of the settlement. By definition, Microsoft knows, and its programmers have access to, current product information today. The delay built into the settlement simply allows Microsoft time to exploit its proven monopoly for another year so that Microsoft's products will have an even greater advantage when disclosure finally begins.
With respect to new middleware products and others, the timing of disclosure also fails to serve the public interest. Disclosure for a new middleware product is not required until the new product's "last major beta test", which is also an undefined term. RPFJ § III.D. As to all other middleware disclosures, Microsoft is free to decide when to conduct the required beta test to 150,000 or more beta testers. RPFJ § III.D. With respect to the disclosure of client protocols, Microsoft is not subject to any time limit whatever. RPFJ § III.E. These provisions effectively immunize continued anticompetitive conduct. Microsoft is essentially given free rein to choose when it will be most advantageous in terms of marketing its products to make the required disclosures; and prior to disclosure it is free to develop and position its products for maximum competitive advantage.
Once again, the CIS provides no explanation why Microsoft's disclosure obligations should not commence immediately, and why at all times thereafter, should not be made as soon as information is available to Microsoft's own programmers. CIS at 34-35. For example, the "beta test" standard in the proposed settlement is far too late to be competitively meaningful. A beta test is one of the last steps taken by a software developer before placing a new product on the market. It is often viewed as more of a marketing tool (to create a "buzz" among technology writers and other cognoscenti), rather than a true development step. Thus, if Microsoft is allowed to wait until this point, it will be able to do to other software developers exactly what it did to Netscape in 1995 -- ensure that a competing product is so late to market that it faces an insurmountable barrier to overtake Microsoft's lead. D.Ct. at ¶¶ 90-92.
The Final Judgment and Litigating States' proposal are much more rational, and likely to lead to meaningful disclosure that would promote interoperability and competition. All disclosures under the Final Judgment and Litigating States' proposal are required to be made in a "timely manner," which is precisely defined as, at a minimum, the earliest of the following times:
(i) When the information is disclosed to Microsoft's own application developers;
(ii) When the information is used by Microsoft's "own platform software developers in software released by Microsoft in alpha, beta release candidate, final or other form";
(iii) When the information is disclosed to any third party; or
(iv) Within ninety days of the final release of the Windows operating system product, but "no less than five days after a material change is made between the most recent beta for release candidate version and the final release."
Final Judgment § 7(ee); Litigating States' § 22(pp).
The definitions of important terms relevant to interoperability are so vague that Microsoft can largely avoid its disclosure obligations. Among other things, the proposed settlement's definitions of "Microsoft Middleware" and "Windows Operating System Product" give Microsoft the ability to completely circumvent even the otherwise paltry disclosure requirements. See pp. 36-40, 42-43 supra. Just as important, the critical term "interoperability" is not defined in the proposed settlement. It should be defined as the ability of a system or product to work with other systems or products in such a way as to effectively access, utilize and support the full features and functions of one another. See Litigating States' § 22(q).
In addition, the definition of "Windows Operating System Product" provides that "[t]he software code that comprises a Windows Operating System Product shall be determined by Microsoft in its sole discretion." RPFJ § VI.U (emphasis supplied). Essentially, this provision grants Microsoft the ability to avoid disclosure to competitors of technical information -- even that necessary to achieve the government's promised "seamless interoperability" -- merely by embedding a middleware product in the Windows PC operating system code. Microsoft can then argue that the product at issue is not middleware, but rather part of Windows, and thus outside all disclosure obligations.
The definition of "Microsoft Middleware" demonstrates the intentional nature of the government's concession on this point and serves no function other than to dilute the effectiveness of the proposed settlement. That definition, which is essentially the flip-side of the "Operating System" coin, also allows Microsoft the flexibility to define whatever it wants to as middleware, by either not obtaining a trademark for the product or by simply bundling it with the Windows PC operating system. RPFJ § VI.J. In either instance, the effect is the same: Notwithstanding the conclusion of the District Court and the Court of Appeals that Microsoft bound middleware to its operating system for the purpose of defending its operating system monopoly in violation of § 2, the company will remain free to continue the same conduct.
The proposed settlement provides that Microsoft must license to its competitors the intellectual property rights for any technical information it is required to disclose. The CIS explains that "[t]he overarching goal of this section is to ensure that Microsoft cannot use its intellectual property rights in such a way that undermines the competitive value of its disclosure obligations." CIS at 49. Limitations on Microsoft's licensing obligations, however, make the provision's impact largely illusory. Indeed, they may well benefit Microsoft to the exclusion of its competitors and competition generally in the PC operating system market.
First, Microsoft is permitted to charge a "reasonable royalty" to any competitor who requests disclosure related to interoperability. RPFJ § III.I.1. This provision is anticompetitive. On its face, it allows Microsoft to enjoy the fruits (i.e., licensing royalties) of its proven illegal monopoly. Moreover, it gives Microsoft the opportunity to use a "royalty" charge to control crucial technical information in a way that restrains its competitors -- a practice Microsoft has already shown a willingness to undertake.29 The government earlier acknowledged that these dangers can be avoided only by requiring any license to be royalty-free:
Gov't D.Ct. Sum. Resp. at 14.
The relevant case law also supports the position that royalty-free licenses are necessary to prospectively remedy Microsoft's illegal monopolization of the PC operating system market. See United States v. General Elec. Co., 115 F. Supp. 835, 844 (D.N.J. 1953) (royalty-free licenses an essential remedy to prevent a continuance of monopoly). The government made this same legal point when it recognized that the instant case is analogous to United States v. Western Elec. Co., 569 F. Supp. 1057, 1082-1091 (D.D.C. 1983), aff'd sub nom. California v. United States, 464 U.S. 1013 (1983), where compulsory, royalty-free, sublicensable licenses were required to remedy past anti-competitive conduct. Gov't D.Ct. Reply Memo at 27-29. As the government pointed out, this provision was one of several that insured that "telecommunications [would] continue to operate in an engineering sense as one national network." Id. at 29 (internal quotation marks omitted). The same functional interoperability is necessary to ensure maximum innovation and competition in all aspects of the computer industry.
Moreover, as in the AT&T case, whether royalty-free licenses are necessary is not an issue of remedying a monopolist's past anticompetitive use of its intellectual property per se, but rather a matter of making sure that the relief granted (in this case, disclosure of certain APIs and communications protocols) is not impeded by onerous license terms. Thus, requiring royalty-free licensing is merely in aid of a remedy for antitrust violations that are not directly related to Microsoft's licensing of its intellectual property.
The public interest also is not served by giving Microsoft the right to condition the grant of any license of its own software upon the licensee's "cross-licensing" any intellectual property rights it may have "that are related to the licensee's exercise of its rights" under the settlement. RPFJ § III.I.5. It was established long ago that cross-license requirements are inconsistent with restoring competition to a monopolized market:
General Elec. Co., 115 F. Supp at 847.
The government's present justification, that this provision is necessary to ensure that Microsoft would not be exposed to "infringement liability" as a result of the interoperability disclosures, is difficult to understand. CIS at 50. No competitor should be forced to disclose its own proprietary information in order to exercise rights put in place to restore competition in a monopolized market. Again, the government does not explain why it was wrong when it earlier concluded that such a provision has no pro-competitive benefit:
Gov't D.Ct. Sum. Resp. at 14-15.
In addition, the disclosures required by cross-licensing would enable Microsoft to get a jump on developing its own product to compete with that which the licensee was forced to disclose. The settlement would thereby further increase Microsoft's timing advantage over its licensee in selling a new product. Nothing requires giving a proven monopolist such a benefit.
The provision in the proposed settlement that limits Microsoft's interoperability disclosure obligations based on security considerations is another loophole that Microsoft can use to justify withholding crucial technical information. Under this provision, if technical information would "compromise the security of a particular installation or group of installations of anti-piracy, anti-virus, software licensing, digital rights management, encryption or authentication systems, including without limitation, keys, authorization tokens or enforcement criteria," it is exempt from disclosure. RPFJ § III.J.1. In addition, Microsoft is also given the right to require that the following conditions be met prior to licensing APIs or communications protocols related to the foregoing subjects:
(i) that any licensee never have participated in software counterfeiting or piracy, and never have willfully violated intellectual property rights;
(ii) that the licensee have a "reasonable business need" for the technical information, and the need must be related to a product that is currently being planned or shipped;
(iii) that the licensee meet "reasonable, objective standards established by Microsoft for certifying the authenticity and viability of its business"; and
(iv) that the licensee agree to submit any program using the disclosed technical information to a third party, approved by Microsoft, to "ensure verification and compliance with Microsoft specifications."
RPFJ § III.J.2.
The government did not permit any such "security" exception in its proposals for the Final Judgment. The government now attempts to justify this provision as a "narrow exception" that is necessary to maintain the integrity of the security-related features of Windows. CIS at 10. To the contrary, the broad language and significant discretion given to Microsoft create loopholes for Microsoft to withhold information essential to interoperability from disclosure generally or from specific rivals it wishes to prejudice.
First, virtually all APIs, communications interfaces or other technical information that are relevant to interoperability, on some level will perform "authentication" or "encryption" functions related to the security of an operating system. Accordingly, the allegedly "narrow" security exception in reality gives Microsoft a virtual carte blanche to withhold information necessary for interoperability, simply by citing this section and claiming the code at issue provides "authentication" or "encryption" functions.
The language requiring that a person seeking disclosure must meet "reasonable, objective standards established by Microsoft for certifying the authenticity and viability of its business" invites abuse by Microsoft. RPFJ § III.J.2.c (emphasis added). For example, Microsoft could exercise its veto power over a disclosure request from an open source developer, on the ground that Microsoft does not consider "its business" to be "authentic" or "viable". The open source community typically operates on a not-for-profit basis, and has long been a competitive adversary of Microsoft. Rebecca Buckman, Microsoft is Suing Linux Start-up Over Lindows Name, Wall Street Journal, December 24, 2001 (Microsoft brings trademark action against open source operating system developer); Lee Gomes, Linux Campaign Is An Uphill Battle For Microsoft, Wall Street Journal, June 14, 2001 ("A Microsoft Corp. effort to vilify Linux and other open-source software appears to be backfiring ... Microsoft Chief Executive Steve Ballmer ...[called] Linux `a cancer that attaches itself in an intellectual property sense to everything it touches.'"); Byron Acohido, Microsoft Memo to Staff: Clobber Linux, USA Today, Jan. 4, 2002 ("Microsoft is escalating its war against Linux.").
Similarly, the requirement that the prospective licensee have a "reasonable business need" for the information also gives Microsoft anticompetitive powers. Microsoft could dispute the asserted "need" or use to its own competitive advantage the information that licensees would presumably submit to it to demonstrate their "need," since the provision effectively gives Microsoft advance notice of its competitors' new products. For this reason, the provision will discourage competitors from even exercising disclosure rights. Finally, Microsoft is given the gratuitous right to require all competitors' programs that use Microsoft's APIs to be verified by a third party, who is "approved by Microsoft," to ensure compliance with "Microsoft's specifications for use of the API or interface." RPFJ § III.J.2.d.
Microsoft's alleged concern for security is a pretext to create a loophole, as well as to allow Microsoft to obtain an unwarranted advantage by having early access to its competitors' trade secrets. The fact that certain code may provide a security function is not a legitimate reason to withhold disclosure. Although Microsoft has not been required to license this information in the past, its software security record is arguably one of the worst in the industry. See, e.g., Wayne Epperson, NT Insurance at a Premium, HostingTech (August 2001), at www.hostingtech.com/security/01_08_nt.html (reporting that insurer J.S. Wurzler Underwriting Managers had discovered "that clients who used Microsoft Windows NT software in their Internet operations were at a greater risk of loss to computer hackers than were the insured Unix or Linux users .... After 5 months of analysis, Wurzler Underwriting Managers made the decision to charge its NT clients an extra premium for insurance coverage."). On the other hand, several of the most effective security programs, such as Kerberos and "Pretty Good Privacy," are available on an open-source basis and freely accessible by the public. These examples prove that even full public disclosure is not inimical to security. What disclosure does prevent, however, is the exercise of monopoly power. For purposes of this proceeding, therefore, the "security" loopholes are not in the public interest.
As determined by the District Court in findings upheld by the Court of Appeals, Microsoft threatened to and did withhold critical technical information from software developers; Microsoft provided or withheld financial benefits depending on a party's willingness to aid in its anticompetitive campaign; and Microsoft contractually prohibited third parties from distributing competing software. CA at 71-73, 76-77. Yet, the proposed settlement fails in many respects to protect third parties from new versions of such past conduct which may be used by Microsoft to protect its operating system monopoly.
In earlier proceedings, the government recognized that any remedy must be "directed towards future competition and innovation," and that while the remedy was based on historical experience, the analysis was "done on a forward-looking basis." See Shapiro Decl. at 2. The government further acknowledged that in this fast-moving industry, any remedial conduct provisions must "be broad enough to prevent Microsoft from engaging in a number of categories of anticompetitive tactics in the future, precisely because the specific tactics that Microsoft might employ in the future are hard to predict today in the face of changing products and technology." Id. (emphasis in original)
Having recognized these elements as essential to any remedy that would serve the public interest, the government now proposes to settle on terms that do virtually nothing to anticipate and prohibit new forms of exclusion. Indeed, in many instances, the government has given up or severely limited provisions in the original Final Judgment that were fully justified by the Court of Appeals' affirmance. On a number of such issues, the proposed settlement accepts positions that Microsoft sought to include in the Final Judgment, but the government then specifically rejected. The government rejected Microsoft's proposals precisely because they "consist[ed] largely of changes that would create loopholes and permit Microsoft to continue to engage in anticompetitive practices like those found by the Court or otherwise to frustrate or undermine the purposes of the Final Judgment." Gov't D.Ct. Sum. Resp. at 6. However, the recent CIS does not attempt to justify the government's acquiescence in what were once viewed as "loopholes" and a license to resume "anticompetitive practices."
The Court of Appeals held that Microsoft had engaged in exclusionary acts and threats of retaliation in violation of § 2. CA at 73, 77-78. In its efforts to promote IE and restrict the distribution of Navigator, Microsoft successfully made threats against Apple. To halt the development of cross-platform interfaces for Java, Microsoft threatened to retaliate against Intel. Microsoft also made threats and retaliated against others who posed potential threats to its monopoly. D.Ct. at ¶¶ 83-84, 91, 101, 102.
The government has recognized the established and urgent need to prevent Microsoft from engaging in acts or threats of retaliation. The original Final Judgment prohibited Microsoft from "tak[ing] or threaten[ing] any action [that] adversely affect[s] any OEM. . . based directly or indirectly, in whole or in part, on any actual or contemplated action by that OEM to use, distribute, promote, license, develop, produce or sell any product or service that competes with any Microsoft product or service." Final Judgment § 3(a)(i). The Final Judgment further prohibited Microsoft from "tak[ing] or threaten[ing] any action affecting any ISV or IHV. . . based directly or indirectly, in whole or in part, on any actual or contemplated action by the ISV or IHV to use, distribute, promote or support any Microsoft product or service, or develop, use, distribute, promote, or support any software that runs on non-Microsoft Middleware or a non-Microsoft Operating System or that competes with any Microsoft product or service." Final Judgment §§ 3(d)(i) and 3(d)(ii).
With respect to OEMs, the government recognized that Microsoft's retaliatory actions "highlight the potential for misuse of monopoly power that must be prevented if potential rivals to Windows and new innovations in software can be expected to emerge." Gov't D.Ct. Memo at 38. The government specifically admitted that the broad anti-retaliation provision was needed "both to prevent subtle or varied forms of coercion and to avoid difficulties in determining the scope of the restriction in an enforcement proceeding." Id. at 38-39.
With respect to banning retaliation against ISVs and IHVs, the government said it was necessary to "ensure that Microsoft does not use its operating system monopoly to nip new competitive threats in the bud." Gov't D.Ct. Memo at 41. The provision was a "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." Gov't D.Ct. Reply Memo at 55.
Moreover, because Microsoft repeatedly sought anticompetitive agreements, such as its attempted market allocation agreements with Netscape and Intel, the Final Judgment included a provision that flatly prohibited agreements that limit competition (Final Judgment § 3(h)) "to ensure that the defendant will be unable to repeat its unlawful conduct." Gov't D.Ct. Sum. Resp. at 17. "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." Gov't D.Ct. Reply Memo at 65.
Even today, the government recognizes the need to prohibit retaliation based on the specific findings of illegal conduct upheld by the Court of Appeals. In its attempts "to protect the applications barrier to entry, Microsoft embarked on a multifaceted campaign to maximize [IE]'s share of usage and to minimize Navigator's." CIS at 13. Not content to merely develop its own browser, "Microsoft decided to constrict Netscape's access to the two distribution channels that led most efficiently to browser usage; installation by OEMs on new personal computers and distribution by [IAPs]." Id. "To ensure that developers would not view Navigator as truly cross-platform middleware," Microsoft also pressured Apple "to make Navigator less readily accessible on Apple personal computers." CIS at 14.
Additionally, as part of "its effort to hamper distribution of Navigator and to discourage the development of software that used non-Microsoft technology, Microsoft also targeted [ISVs] by contractually requiring ISVs to use [IE]-specific technologies in return for timely and commercially necessary technical information about Windows, and precluded important ISVs from distributing Navigator with their products." CIS at 14-15. Ultimately, "Microsoft's actions succeeded in eliminating the threat that the Navigator browser posed to Microsoft's operating system monopoly. . . Navigator lost its ability to become the standard software for browsing the Web because Microsoft had successfully -- and illegally -- excluded Navigator from that status." CIS at 15.
The proposed settlement (RPFJ §§ III.A, F) fails to implement the procompetitive goals that the government has repeatedly expressed, and substitutes weak and narrow protections for broad prohibitions to interdict new forms of exclusionary conduct adopted by Microsoft in response to new forms of competition.
The settlement would only protect OEMs, ISVs and IHVs against "retaliation." Microsoft's record of retaliatory conduct, however, demonstrates that the ban against retaliation must apply to all third parties.30 Microsoft has demonstrated that it will take any action necessary against any entity that poses a threat to its monopoly, by making threats, offering inducements, coercing or contractually restricting others. The government offers no justification whatever, let alone any persuasive reason, to limit the types of third-parties against which Microsoft cannot engage in unlawful retaliation. The settlement's failure to ban retaliation broadly is all the more troubling because this provision does not impose any affirmative duties on Microsoft. The only "burden" is that Microsoft must refrain from punishing those who might challenge Microsoft's illegal monopoly.
The proposed settlement states that Microsoft "shall not retaliate," but never defines "retaliation." Although the government stated that its intention was to "prevent subtle or varied forms of coercion and to avoid difficulties in determining the scope of the restriction in an enforcement proceeding" (Gov't D.Ct. Memo at 38-39), the vague language of the settlement fails to meet that goal. Without a definition of "retaliate," such as a prohibition against "taking or threatening adverse actions" (see Final Judgment §§ 3(a)(i), 3(d)), Microsoft will be free to argue that no violation has been established on a particular set of facts. Given the extraordinary record of Microsoft's ingenuity in abusing its monopoly power, any definitional doubt must be resolved against the wrongdoer by imposing a broad definition of "retaliate."
Moreover, with respect to OEMs, Microsoft is only prohibited from "retaliating against an OEM by altering Microsoft's commercial relations with that OEM or by withholding newly introduced forms of non-monetary Consideration." See RPFJ § III.A (emphasis added). The use of the words "newly introduced" is ambiguous in that it suggests that Microsoft is permitted to withhold existing forms of non-monetary compensation. The CIS offers no explanation for this ambiguity concerning whether Microsoft may continue to engage in conduct previously adjudged illegal. CIS at 25-26.
The proposed settlement too narrowly limits the type of conduct by third parties for which Microsoft may not "retaliate." It only prohibits "retaliation" that occurs "because" of conduct by the OEM, ISV or IHV. By imposing a causation requirement on the injured party, the government again gives a proven wrongdoer the benefit of a doubt to which it is not entitled. The prior Final Judgment struck the correct balance by prohibiting any adverse action by Microsoft based directly or indirectly on any actual or contemplated action by the protected party. Final Judgment §§ 3(a)(i), 3(d); see also Litigating States' § 8.
The proposed settlement's ban on "retaliation" currently only applies to a protected party where it is developing, distributing, promoting or using products that compete with Microsoft Platform Software or middleware.31 Given the multitude of ways in which new threats can (and did) develop to contest the Windows operating system, the ban on retaliation will not be truly effective unless it protects any action or contemplated action involving products or services that compete with any Microsoft products or services.
The settlement contains broad savings clauses and exceptions that give Microsoft loopholes for abuse and are not justified by the Court of Appeals' findings. The OEM retaliation provision permits Microsoft to provide "Consideration" to any OEM with respect to any Microsoft product or service if the "Consideration is commensurate with the absolute level or amount of that OEM's development, distribution, promotion, or licensing of that Microsoft product or service." RPFJ § III.A. Similarly, Microsoft may enter agreements with ISVs limiting their ability to develop, promote or distribute competing software, if the limitations are "reasonably necessary to and of reasonable scope and duration in relation to a bona fide contractual obligation of the ISV to use, distribute or promote any Microsoft software or develop software for, or in conjunction with, Microsoft." RPFJ § III F.2.
These vague provisions are an invitation for abuse. Microsoft has repeatedly used these very practices to maintain an illegal monopoly for over a decade. Under these circumstances, a broad prohibition that puts the burden entirely on Microsoft to prove the bona fides of any "consideration" or exclusivity is not only appropriate, but essential to revive competition.
The retaliation provisions also include broad savings clauses that provide that nothing in the provisions "prohibit Microsoft from enforcing any provision of any agreement with any [OEM, ISV or IHV] or any intellectual property right, that is not inconsistent" with the settlement. RPFJ §§ III. A and III. F. 3. Given the ambiguities in the settlement, this loophole, too, will invite aggressive interpretation by Microsoft, and further litigation.
The provision in the Final Judgment banning agreements to limit competition (Section 3(h)) has been completely eliminated, leaving Microsoft free to seek to enter market allocation agreements such as the ones it proposed to enter with Netscape and Intel. D.Ct. at ¶¶ 83-84, 97, 101. The proposed settlement (§ III.G) only prohibits Microsoft from entering into agreements with certain entities to use or distribute Microsoft Platform Software exclusively or in fixed percentages. There is no provision that limits Microsoft's ability to enter into agreements with competitors providing that they refrain from developing or distributing products that compete with the Windows operating system or a Microsoft middleware product. This CIS does not explain this glaring omission.
Nothing in the proposed settlement protects individuals or entities from retaliation by Microsoft for participating or cooperating in this litigation. In a letter to the Senate Judiciary Committee, former Solicitor General Robert H. Bork expressed the realistic concern that "any potential witness with knowledge of anticompetitive conduct in a monopolized market has to weigh the potential benefit of his or her testimony against the likely response of the defendant monopolist. The [government's] proposed meaningless remedy would insure that no witness would ever testify against Microsoft in any future enforcement action." See Letter from Robert H. Bork to Senate Judiciary Committee of 12/11/01, at 4. Again, the CIS does not address this issue of obvious concern, given Microsoft's track record of anticompetitive abuse.
The District Court made findings of fact, which were not questioned by the Court of Appeals, that provide examples of incentives and threats used by Microsoft to induce OEMs to promote IE and not pre-install or promote Navigator. Thus, Microsoft gave reductions in price to OEMs who set IE as the default browser on their PC systems; Microsoft gave further reductions to OEMs who displayed IE logos and links on their home page; and Microsoft gave OEMs millions of dollars in co-marketing funds in exchange for carrying out other promotional activities for IE. D.Ct. at ¶ 231.
The original Final Judgment contained a strong ban on discriminatory license terms. It compelled Microsoft to license Windows to all covered OEMs on uniform terms and prohibited Microsoft from offering market development allowances or discounts. Final Judgment § 3(a)(ii). Microsoft was further required to give OEMs equal access to, inter alia, licensing terms, discounts, technical and marketing support and product and technical information; and to give written notice and an opportunity to cure before terminating an OEM's license. Id.
The government recognized the necessity for this provision requiring "transparent and uniform pricing to the largest OEMs . . . so that Microsoft cannot retaliate against an OEM for supporting non-Microsoft software." Gov't D.Ct. Memo at 39. Uniform terms and pricing were also seen as necessary to "terminate Microsoft's practice of charging substantially different prices for Windows licenses to reward cooperative OEMs, effected in part by its market development allowances, and will thus make it easier for OEMs to promote non-Microsoft products in response to consumer demand." Id. The government found that this uniformity was "necessary to prevent Microsoft from employing the myriad forms of coercion and reward that" have been held to injure competition. Gov't D.Ct. Reply Memo at 43-44. "Such coercion is difficult to detect, and the mere threat of its use may be sufficient to accomplish the desired, anticompetitive result." Id. at 44.
In the context of the present settlement, the government continues to define as a "critical" objective ensuring that OEMs are truly "free to choose to distribute and promote middleware without interference from Microsoft." CIS at 25. It recognizes that Windows' license royalties and terms are "inherently complex and easy for Microsoft to use to affect OEMs' behavior, including what software the OEMs will offer to their customers." CIS at 28. By purportedly requiring uniform licensing, the government says that the proposed settlement eliminates "any opportunity for Microsoft to set a particular OEM's royalty or license terms as a way of inducing that OEM to decline to promote non-Microsoft software or retaliating against that OEM for its choices to promote non-Microsoft software." CIS at 28. The government's stated goal is ensuring "that OEMs can make their own independent choices." CIS at 28.
Here, as in many other respects, the proposed settlement fails to fulfill the government's stated objectives.
Unlike the Final Judgment (§ 3(a)(ii)), which prohibited market development allowances ("MDAs") outright, the proposed settlement would permit MDAs (RPFJ § III.3.B.3), if certain restrictions are met, despite the fact that MDAs have been repeatedly used by Microsoft to induce OEMs to take actions that protect Microsoft's monopoly. As the government earlier acknowledged, making any MDAs permissible creates a loophole that will allow the very discrimination against OEMs that the provision is intended to prevent. See Gov't D.Ct. Sum. Resp. at 10 (the use of undefined and unbounded "objective" pricing criteria will permit Microsoft to reward or punish OEMs by charging them different prices).
The proposed settlement does not require that Microsoft provide equal access to licensing terms, discounts, technical support, etc. Without this information, OEMs cannot fairly negotiate license terms. In the current market, Microsoft offers significant discounts to OEMs that take the entire Windows package; those discounts enable the OEMs to be competitive with other PC manufacturers. However, if an OEM tries to negotiate anything out of the package, Microsoft significantly increases the price, making the OEM non-competitive. OEMs do not know what terms are negotiable and are afraid to negotiate aggressively out of fear they will be punished by Microsoft.
For this reason, when Microsoft previously requested deletion of the equal access provision, the government rejected the idea because it "would allow [Microsoft] to reward or punish Covered OEMs with different Windows prices and non-price licensing terms and conditions and thus to evade the purpose of the Final Judgment." See Gov't D.Ct. Sum. Resp. at 10. Moreover, the government's view then was that "the burden should not be on OEMs to know of and affirmatively seek out equal treatment; Microsoft could take advantage of a Covered OEM's ignorance of what has been provided to other Covered OEMs to reward or punish that OEM and thus to evade the purposes of the Final Judgment." Id. at 10-11. The government provides no satisfactory rationale for changing that view.
The proposed settlement allows Microsoft to provide reasonable volume discounts based upon the actual volume of licenses. (RPFJ § III.B.2) Unless the provision requires that the volume discounts be based on the independently determined actual number of shipments, however, Microsoft will continue to have the power that it exercised in the past to manipulate discounts.
The proposed settlement creates an unnecessary exception to the written notice requirement for termination of OEM licenses. It provides that Microsoft need not provide notice or opportunity to cure if Microsoft has given two prior written notices. This exception invites Microsoft to abuse the notice provision and then arbitrarily revoke an OEM's license. Moreover, Microsoft is not even required to show "good cause" for termination. Again, the government provides no rationale why a proven monopolist should be given any such advantages.
The Court of Appeals held that Microsoft's exclusive contracts with IAPs are exclusionary devices in violation of § 2. CA at 71. By ensuring that the vast majority of IAP subscribers were offered IE as their default browser or as the only browser, Microsoft's deals with IAPs had a significant effect in maintaining its monopoly by keeping usage of Navigator below the critical level necessary for it or any other rival to threaten Microsoft's monopoly. CA at 71.
In addition to the evidence specifically relied on by the Court of Appeals, the District Court made findings, not reversed, of other unlawful exclusive agreements. For example, in exchange for an agreement by IAPs to promote and distribute IE preferentially over Navigator and to convert existing subscribers from Navigator to IE, Microsoft gave fourteen IAPs placement in its Windows Referral Server. D.Ct. at ¶ 255, 256. Microsoft also entered into agreements with AT&T, WorldNet, Prodigy and CompuServe limiting their ability to promote non-Microsoft browsers. D.Ct. at ¶ 305.
The Court of Appeals also held that Microsoft's illegal agreements with ISVs further foreclosed rival browsers from the market. CA at 72, 76. Microsoft entered dozens of "First Wave" agreements with ISVs, promising to give them preferential support in using Windows in exchange for ISVs agreeing to use IE as the default browser in any software they developed. The "First Wave" agreements with ISVs also required the ISVs to use Microsoft's JVM rather than Sun's JVM. This directly protected Microsoft's monopoly from the middleware threat. CA at 76.
To redress these exclusionary agreements, the government, earlier in this case, advocated a general and broad prohibition against any and all manner of exclusive dealing by Microsoft. In the Final Judgment, Microsoft was generally prohibited from offering a third party any consideration in exchange for that party's agreement to restrict development, production, distribution, promotion or use of, or payment for, any non-Microsoft Platform Software; distributing, promoting or using any Microsoft Platform Software exclusively; degrading the performance of any non-Microsoft Platform Software; and with respect to IAPs or ICPs, distributing, promoting or using Microsoft software in exchange for placement with respect to any aspect of a Windows Operating System. Final Judgment § 3(e).
The government recognized that such a ban was necessary because Microsoft had "coerced and bribed" third parties into becoming, willingly or unwillingly, participants in strengthening the applications barrier to entry. Gov't D.Ct. Memo at 41. The government stated that to prevent recurrence, Microsoft had to be barred from any exclusive dealing or percentage contracts that require a third party to limit its dealings in, or to degrade the performance of, non-Microsoft products, to deal solely in Microsoft software, or, in the case of IAPs or ICPs, to exchange promotion of Microsoft products for placement in the Windows operating system. Id. Significantly, the government advocated a general ban because Microsoft had dealings with a wide range of companies and because "it is difficult to predict precisely which trading partners Microsoft might otherwise seek to tie up under exclusive arrangements in the next several years." Shapiro Decl. at 19. Only a general ban on exclusionary contracts would "serve to lower entry barriers more effectively than would more limited provisions directed at specific categories of trading partners." Id.
In the CIS, the government continues to recognize the necessity of preventing "Microsoft from using either money or the wide range of commercial blandishments at its disposal . . . to hinder the development and adoption of products that, over time, could emerge as potential platform threats to the Windows monopoly." CIS at 42. However, the exclusive dealing provision in the proposed settlement (RPFJ § III. G) fails to meet the goals that the government recognizes are essential. Nor will it prevent Microsoft from engaging in the same types of conduct that were found to be unlawful.
The provision is limited to the listed entities (IAP, ICP, ISV, IHV or OEM), but should be extended to all third parties. The government has specifically acknowledged that a "general ban" is necessary precisely because it is too difficult to predict which entities Microsoft might seek to tie up under exclusive arrangements in the next several years. Only a general ban may effectively lower entry barriers, as compared to "more limited provisions directed at specific categories of trading partners." Shapiro Decl. at 19.
The proposed settlement only prohibits Microsoft from entering into agreements with certain third parties that grant consideration on the condition that the entity "distributes, promotes, uses, or supports, exclusively or in a fixed percentage, any Microsoft Platform Software." RPFJ § III.G.1. The Final Judgment prohibited Microsoft from entering into an agreement with any third party that grants consideration to "distribute, promote or use any Microsoft Platform Software exclusively" (Final Judgment § 3(e)(ii) and "to restrict its development, production, distribution, promotion or use of, or payment for, any non-Microsoft Platform Software" (Final Judgment § 3(e)(i)). Microsoft attempted to delete the provision that prohibited agreements with a third party to restrict the development, production, distribution, promotion or use of non-Microsoft Platform software, but the government rejected the proposal. See Microsoft D.Ct. Com. at 12.
The proposed settlement does not prohibit Microsoft from granting consideration to a party that agrees to refrain from using or distributing products or services that compete with Microsoft.32 RPFJ § III.G.1. Microsoft would thus be allowed to grant consideration (in the form of money, technical information or support, or otherwise) in exchange for the party's agreement not to use or distribute a competing product. Such an agreement would be the functional equivalent of the "First Wave" agreements with ISVs found to be illegally exclusionary by the Court of Appeals. The CIS does not articulate a satisfactory basis for omitting from the settlement a prohibition on the types of actions that were adjudged illegal.
Under the proposed settlement, Microsoft is permitted to enter into agreements with IAPs and ICPs that condition their placement in Windows on their agreement to distribute or promote Microsoft Platform Software. RPFJ § III.G.2. When Microsoft argued against this provision in the original Final Judgment, the government rejected its proposal because phrasing it the way Microsoft proposed (and as it is now phrased in the proposed settlement) would allow Microsoft to achieve the same anticompetitive purpose, by simply amending its agreements to require distribution of Microsoft's browser instead of limiting distribution of competing browsers. See Gov't D.Ct. Sum. Resp. at 16; Microsoft D.Ct. Com. at 12. The government has not explained why it has now completely reversed its position.
The various exceptions built into the exclusive dealing ban render it potentially meaningless. While purportedly prohibiting exclusive or fixed percentage arrangements, such agreements are actually permitted when Microsoft: obtains a representation that it is "commercially practicable" for the entity to provide equal or greater distribution of a competing product; if Microsoft enters into any type of loosely defined "joint venture" agreement; or if Microsoft licenses in intellectual property from a third party. (RPFJ § III.G). When Microsoft proposed to include a similar joint venture exception in a related provision of the Final Judgment (Final Judgment § 3(h), Ban on Agreements Limiting Competition), the government rejected the proposal as "unnecessary" and because it "would enable Microsoft to enter into anticompetitive market division agreements regarding such products by labeling them, as it attempted to label at trial the June 1995 Netscape meeting, `joint development agreements' or `agreements ancillary to lawful joint ventures.'" Gov't D.Ct. Sum. Resp. at 18. Once again, the government has agreed to a loophole that can only benefit Microsoft by inviting abuse and further litigation.
As demonstrated above, without correction of numerous deficiencies, the proposed settlement will not restore competitive conditions because it largely permits Microsoft to conduct business as usual and it effectively rubberstamps further anticompetitive conduct. Even if all of the other deficiencies were corrected, however, the term of the proposed settlement is too short to restore meaningful competition with the Windows monopoly. Although this Court's finding that Microsoft illegally maintained its decade-long monopoly has now been affirmed by the Court of Appeals, Microsoft has availed itself -- up to and including the present -- of every opportunity to maintain and extend its monopoly through anticompetitive actions. For example, Windows XP, designed during the height of Microsoft's litigation with the government and released just before the settlement was announced, commingles code in the exact manner found unlawful by the Court of Appeals.
The government now claims that a five-year decree will be sufficient to restore competition. CIS at 60. This assertion is inconsistent with the Department of Justice's own Antitrust Manual, which states that "the Division's standard decree language requires that the consent decree expire on the tenth anniversary of its entry by the court. [T]he staff should not negotiate any decree of less than 10 years' duration, although decrees of longer than 10 years may be appropriate in certain circumstances." Department of Justice, Antitrust Division Manual, ch. IV at 54 (3d ed. 1998) (emphasis added). As the government argued in earlier urging entry of a ten-year decree:
Gov't D.Ct. Sum. Resp. at 20 (emphasis added).
The government offers no "sound justification" for its acceptance of a settlement that would last only five years. RPFJ § V.A. The CIS only states that five years "provides sufficient time for the conduct remedies contained in the Proposed Final Judgment to take effect in this evolving market and to restore competitive conditions to the greatest extent possible." CIS at 60. There is no factual support cited for the ipse dixit that this is "sufficient time," while it is certain that the standard ten-year decree would restore competition to a greater extent. The government's present position also conflicts with other assertions that it previously made in this case. See Gov't D.Ct. Memo at 27 ("Looking forward, the Court must anticipate that Microsoft, unless restrained by appropriate equitable relief, likely will continue to perpetuate its monopoly by the same anticompetitive methods revealed at trial, although directed at whatever new competitive threat arises.").
Moreover, the government's only recourse under the proposed settlement -- the possibility that the proposed decree could be extended "one time" for a "maximum of two years" -- is so short as to be virtually meaningless. RPFJ § V.B. And, even to obtain the "one-time" two year extension, the government would first have to demonstrate through a complex, lengthy and burdensome enforcement procedure, that Microsoft engaged in "a pattern of willful and systematic violations" of the decree. Id. The government should be seeking -- as it originally sought -- more than just conduct remedies "tak[ing] effect." CIS at 60. It should instead be trying to ensure that, once the remedies "take effect," they remain in effect for a period sufficient to restore competition to the greatest extent possible. As the government told the District Court, "[t]en years is customary in antitrust cases and in any event four years is too short." Gov't D.Ct. Sum. Resp. at 20. There is nothing in the Court of Appeals' decision that justifies the government's decision to depart from its own formal policy.
As a matter of law, the government's previously stated position was correct. The case law demonstrates that in cases where a monopolist has committed a Section 2 violation, it has been "customary" for courts to impose remedial decrees lasting ten years. In over 70 cases since the Department of Justice's Antitrust Manual was adopted in 1978 to change the prior policy of seeking decrees of unlimited length, the government has required consent decrees having a minimum ten year duration. See e.g., United States v. Greyhound, Civ. No. 95-1852 (RCL), 1996 WL 179570 (D.D.C. Feb. 27, 1996) (bus companies); United States v. Playmobil USA, Inc., Civ. No. 95-0214, 1995 WL 366524 (D.D.C. May 22, 1995) (toy companies); United States v. Republic Services, Inc., Civ No. 00-2311, 2001 WL 77103 (D.D.C. Jan. 18, 2001) (waste collection companies).33 The government also imposed restrictions on broadcasters' purchase of television program rights for a period of 15 years. See United States v. American Broadcasting Co., Inc., Civ. No. 74-3600 (RJK), 1980 WL 2013 (C.D. Cal. Nov. 14 1980).
In contrast to these settlements, the proposed decree here is to last only five years, although this case has significantly more importance to the national economy. In addition, Microsoft's prior conduct, the importance of this case to the national economy, and the explicit findings, upheld on appeal, of Microsoft's illegal monopolization activities mandate, if anything, that Microsoft's conduct be supervised for a period longer than the standard ten-year term.
At bottom, a five-year injunction is too short to allow meaningful competition to develop in the operating system market. It has been over ten years since the government first began to investigate Microsoft's practices, and it took six years from Microsoft's first anticompetitive act cited by the District Court for this case to reach the appellate level. The effects of Microsoft's past and present anticompetitive actions, which have already continued for over a decade, will likely last much longer. The government itself concedes that Netscape and Java are likely dead and no longer pose credible threats to Microsoft's operating system monopoly. CIS as 16-17. Even if the proposed decree's numerous loopholes were plugged, it will take considerably longer than five years for strong new competitors to emerge. Most important, the ability of those competitors to become viable depends upon the existence of judicial protection. See United States v. GTE Corp., 603 F. Supp. 730, 742-43 (D.D.C. 1984) (rejecting as too short five-year expiration date for decree provisions in section 2 case).
The government claims that the various obligations imposed upon Microsoft in the proposed settlement are supported by "strong enforcement provisions," "including the power to seek criminal and civil contempt sanctions and other relief in the event of a violation." CIS at 5. It also states that Plaintiffs' right, under certain circumstances, to request a one-time extension of the final judgment of an additional two years "is designed to supplement the government's traditional authority to bring contempt actions." CIS at 60.
The reality, however, is to the contrary. The compliance and enforcement provisions in the proposed settlement are entirely inadequate to prevent Microsoft from engaging in future anticompetitive conduct. The provisions in the proposed settlement will result in time delays, inefficient administration of the decree, and ultimately give Microsoft the opportunity to continue its anticompetitive acts unabated. The most critical deficiencies include:
By agreeing that a "Technical Committee" comprised of computer programming and software experts should perform a monitoring role (RPFJ §§ IV.B.1 and IV.B.2), the government seemingly recognizes the difficulty of monitoring and enforcing Microsoft's compliance with the decree. The government also appears to recognize the obvious -- that "Internal Antitrust Compliance" by Microsoft, though necessary, is insufficient. Unfortunately, the government fails to recognize that its own concessions make the compliance task vastly more difficult.
The Technical Committee contemplated by the settlement is simply not an adequate answer, much less a substitute for self-enforcing prohibitions. The person charged with responsibility for monitoring and enforcing Microsoft's compliance must be an experienced antitrust lawyer or former federal judge. He or she can then hire software and programming experts to render assistance, but the responsibility for determining whether the specific provisions of a complex court order have been violated must be made by an individual with impeccable legal credentials and long experience in antitrust law and decree interpretation. No novel device such as a "Technical Committee" is required. The mechanism of a special master under Rule 53, Fed.R.Civ.P., is readily available and entirely appropriate. See United States v. Microsoft Corp., 147 F.3d 935, 954 (D.C. Cir. 1998) (recognizing "well-established tradition allowing use of special masters to oversee compliance").
Aside from its dual, repetitive investigative and reporting procedures (Technical Committee to the government, then the government to the Court, see pp. 114-115 infra.) (RPFJ § III.D.4), the proposed settlement is flawed because it imposes substantial constraints upon how the Technical Committee's findings may be used to assure compliance. The settlement prohibits the admission into evidence of the Technical Committee's findings "in any enforcement proceeding before the Court for any purpose;" and prohibits any technical committee member from testifying in any proceeding or before any tribunal regarding any matter relating to the Final Judgment. RPFJ § IV.D.4.d. Each of those prohibitions denies the Court information from the independent technical personnel who are uniquely knowledgeable about the nature of the violation. Indeed, although the decree proposal allows Microsoft to offer any evidence it wants, it shuts off from the Court the evidence in the possession of the technical committee members who rejected Microsoft's explanations.
The proposed settlement only requires that the officers and directors of Microsoft receive copies of the decree and be "annually briefed on [its] meaning." RPFJ § IV.C.3. In order to be effective, however, all managers (not just corporate "officers") and all employees who have positions that enable them to initiate or implement anticompetitive conduct must be required to read, understand and comply with the decree. Of the published consent decrees that require employees of the company to certify that they have read, understood, and will comply with the decree, most extend compliance certification beyond officers and directors of the company, to also include other managers and employees who have responsibility for overseeing the business activities of the antitrust violator. See, e.g., United States v. Western Elec., Civ. No. 82-0192 (HHG), 1991 WL 33559, at *5 (D.D.C. Feb. 15, 1991) (requiring certification of compliance from each officer and management employee); United States v. Delta Dental of R.I., Civ. No. Civ. A. 96-113P, 1997 WL 527669, at *2 (D.R.I. July 2, 1997) (requiring certification of compliance from all officers, directors, and employees who had responsibility for approving, disapproving, monitoring, recommending, or implementing any provisions in agreements with participating dentists); United States v. Business Inv. & Dev. Corp., No. MO-81-CA-20, 1982 WL 1866, at *2 (W.D. Tx. July 16, 1982) (requiring certification of compliance from all officers, directors, employees and franchisees).
Moreover, the Chairman, CEO or other responsible senior officer of Microsoft should be required to certify periodically to the Court that Microsoft is in compliance with its obligations. The record evidence that Microsoft's highest officials were not only aware of, but actively encouraged, initiated or directed Microsoft's anticompetitive practices, see, e.g., CA at 73, 77; D.Ct. at ¶¶ 80-87, 100, 108, 112-13, 124-29, 340-349, 396, 406-07, makes it all the more necessary to include such certification provisions to ensure that Microsoft takes seriously its responsibilities under any decree to abide by the antitrust laws.
Because of the extraordinarily rapid pace of technological and business developments in the computer industry, avoiding delays in compliance is a critical element in effectively eliminating Microsoft's unlawful behavior and restoring competition. Whether the monitoring function is performed by a Technical Committee or Special Master, the monitor should simultaneously report to both the Court and the Plaintiffs. In defending its decisions to make numerous substantive concessions to Microsoft during the settlement process, the government has cited the substantial time it might take to litigate this case to conclusion if it held out for stronger relief than Microsoft would accept. CIS at 61. Those same time considerations militate against the time-consuming enforcement process contained in the proposed settlement.
The proposed decree contemplates an elaborate procedure whereby the Technical Committee, after receiving a complaint about Microsoft's conduct, would be required to meet with Microsoft's internal compliance officer, and allow Microsoft to respond to the complaint, before it determines whether the complaint can be resolved informally. RPFJ § III.D.4.b. There are no time limits on most of these procedures. If, after completing that procedure, the Technical Committee believes the dispute cannot be resolved and that Microsoft's conduct violated the decree, the Technical Committee would then report the violation to the government in the first instance. RPFJ § III.D.4.c. It then would be up to the government, in turn, to evaluate Microsoft's conduct and determine whether the violation should be reported to the Court.
This process guarantees that considerable time will lapse between a violation of the decree by Microsoft and the Court's eventual review of the problem. First, the process for "Voluntary Dispute Resolution" contemplated by the proposed decree will substantially delay, and, in some circumstances, entirely eliminate, the reporting of violations to the Court. However, effective enforcement requires that any violation of the decree should be reported by the Technical Committee or Special Master immediately and directly to the Court. Action by Microsoft to "voluntarily" cease the unlawful conduct may then, along with other factors, be considered by the Court in determining the severity of any sanction imposed.
In sum, for any remedy to be effective in this case, it must be imposed quickly -- not after months or years of further "dispute resolution." Under the enforcement scheme contemplated by the proposed settlement, however, that simply cannot happen.
SBC is one of the world's leading businesses in the provision of data and voice communications and Internet access. SBC's affiliates serve nearly 60 million telephone access lines nationwide and 21 million wireless customers. SBC and its affiliates are major providers of DSL high-speed and dial-up Internet service, voice messaging services, and directory advertising and publishing products. SBC, through its affiliates, has committed substantial resources to the development of a host of computer- and Internet-related businesses. These businesses are designed to provide consumers with flexibility, convenience and, most importantly, more choice.
With these initiatives, together with its expanding telephone, wireless and Internet operations, SBC is prepared to compete vigorously during the coming decade as the "convergence" of communications and computing technologies continues to accelerate. That highly competitive environment, however, is threatened by Microsoft's ability -- unless restrained by a strong and effective decree in this case -- to use its Windows operating system monopoly to control the electronic "gateways" that link the Internet and its myriad service and content providers to consumers' homes and offices. That control of the gateways, in turn, will enable Microsoft to entrench the Windows monopoly even more firmly.
SBC currently has, or is developing, several businesses in competition with Microsoft, which, together with other similar businesses, directly or indirectly, threaten or are threatened by the Windows operating system monopoly.
SBC affiliates Southwestern Bell, Ameritech, Pacific Bell, Nevada Bell, and SNET are the Incumbent Local Exchange Carriers that provide telephone service in thirteen states.34 In addition, SBC owns a sixty percent interest in Cingular Wireless, which provides nationwide cellular telephone and Internet-related services.
Various SBC affiliates, including Prodigy, provide ISP services, as well as dial-up and broadband (via DSL) Internet access services nationwide. SBC recently finalized a joint venture with Yahoo, whereby Yahoo will provide SBC's Internet portal (home page). In addition, SBC owns an "Internet Data Center" which rents server usage to businesses and e-commerce clients.
Microsoft is an actual competitor of SBC in all of these businesses, including voice telephony. Microsoft is actively developing its Voice over Internet Protocol ("VoIP") through its "Net2Phone" business. This service is being embedded in Windows XP, with the aim of convincing customers to use the Internet for long distance and local calls. Microsoft also provides an Internet access service, MSN, which takes advantage of Microsoft's operating system monopoly by virtue of its being bundled with Windows. By bundling additional products and services such as its ".NET" initiative and its Passport service with Windows XP, Microsoft is also using its monopoly power to give itself an unfair advantage in new markets for Internet and e-commerce business solutions.
Even in businesses in which Microsoft is not now a direct competitor of SBC, however, such as local, long distance and cellular telephone service, Microsoft's operating system monopoly poses grave risks to the competitive marketplace that Congress sought to ensure in the Telecommunications Act of 1996. If Microsoft is permitted by the proposed settlement to maintain and expand its PC operating system monopoly, it will become the gatekeeper for competitors to offer and for consumers to access key communications and entertainment products and services, including telephone, Internet access, voice messaging, instant messaging, music, video services, e-commerce, and interactive games. Without strong remedial measures designed to break its operating system monopoly (which the proposed settlement does not contain), Microsoft will be in a position to favor its own and its partners' communications, entertainment and related services, to exclude competitors' services from access to consumers, to impose costs on rivals and to degrade their services (whether through a toll, a charge for being listed as an available service or an interoperability obstacle), all with the effect of squelching competition and harming consumers. No provisions in the proposed settlement even address, let alone bar, such anticompetitive conduct.
Just as Microsoft has for years successfully imposed on consumers its own software products and other services, irrespective of the comparative merits of competing products it excluded from the market, Microsoft will be able to repeat its anticompetitive strategy in collateral markets including the key communications and entertainment markets described above. This is not speculation; Microsoft has already announced that it is developing an extension to Windows XP, named Freestyle, that will make its Windows PC the communications gateway to the home. See Byron Acohido, Challenge Microsoft? It Could Take Moxi, USA Today, Jan. 16, 2002, at B-3; Microsoft Unveils New Home PC Experiences with "Freestyle" and "Mira" (Jan. 7, 2002), at www.Microsoft.com/presspass/Press/2002/Jan02/01.
SBC's Unified Messaging Service ("UMS") is a good example of a new business placed at serious risk because of the anticompetitive actions that the proposed settlement would allow. UMS is currently in the final development stages, with a projected market introduction in late 2002. UMS will allow customers to retrieve their voice, e-mail and fax messages from a single "mailbox" that will be accessible by phone (wire or wireless) or via the Internet. In the future, SBC plans to add other services to UMS, including instant messaging and video messaging.
UMS is the first in a new generation of services that will create a convergence of all voice, video and data services into one application. Central to UMS is the principle of "any device, anywhere." SBC has designed UMS to be fully accessible through any type of telephone, personal computer, or handheld device. UMS will operate on any software operating system and with any combination of other software applications and services. UMS is also "agnostic as to provider," meaning that it will function regardless of the provider of Internet access, phone or wireless service. For example, consumers can access UMS as easily and effectively through an inexpensive cellular phone that makes no use of Microsoft technology as they can through an expensive Windows PC using Internet Explorer as its browser.
UMS and similar services offered by other companies will compete by giving consumers the ultimate in choice and convenience, enabling them to access all UMS services from virtually any phone or computing device anywhere in the world. UMS will be a direct competitor of Microsoft's e-mail service (Hotmail), and with MS Messenger, a video and instant messaging service, both of which are promoted through integration with Windows.
Because UMS will function with any operating system or Internet browser, and will provide a number of the same functions (voice/data messaging, e-mail, instant messaging) that are or will be provided by a Windows PC, UMS presents a significant competitive threat to Microsoft's PC operating system monopoly. UMS is part of the "movement off the desktop" taking place in the computer industry, which offers increased flexibility and choice to the consumer.
Central to this innovation is that the vast majority of actual computing functions will be performed away from the consumer's computing devices, on servers connected through the Internet. A consumer will no longer need a Windows PC, containing a large hard drive and powerful microprocessor. Instead a simpler, inexpensive device, such as a cell phone or PDA, with a very basic operating system and an Internet browser, when coupled with products like UMS, will allow the consumer to perform many of the functions of a Windows PC at a significantly reduced price and with much greater flexibility and convenience. D.Ct. at ¶¶ 22-27, 56, 59-60 (cited with approval in CA at 52, 79).
In order for UMS to function, however, and to present a competitive alternative to the Windows PC operating system monopoly, SBC must have the ability to effectively process voice and data transmissions through a complex network of different servers, each of which performs separate functions and employs different technologies. At a minimum, when a UMS customer seeks to retrieve a message, either by phone or through the Internet, the voice or data transmission will travel between and through several separate SBC servers (gate, mail, LDAP (lightweight directory access protocol), directory and web server). Each of these servers performs independent functions. As a result, it is critical to UMS that any type of PC, Internet browser, cell phone or handheld device be fully interoperable with all of SBC's servers. UMS also has the potential to be used in a home network, thus requiring full interoperability to extend to set-top boxes.
In short, for SBC and other companies to deliver Internet-based services like UMS, they absolutely must have a "protected chain of interoperability" extending throughout all computers, servers, and other devices which participate on the Internet -- including the Internet browser and PC. If only one link on the chain is not fully interoperable with the entire network, UMS will not be able to process its voice and data transmissions, and thus the convenience and vast array of choices UMS could bring to consumers as an alternative to the Windows PC operating system monopoly will be eliminated.
There is little doubt that Microsoft will continue to recognize the danger that server-based computing, and multi-platform, multi-device products like UMS, pose to the applications barrier to entry. D.Ct. at ¶ 60 (cited with approval in CA at 79). Such alternative server-based computing pathways on the Internet, which rely on open operation and architecture, like Java, will attract applications used for Internet communications. In the past when such threats to the applications barrier to entry that protects the Microsoft monopoly have emerged, Microsoft has responded with anticompetitive conduct. Indeed, the actions taken by Microsoft to eliminate Netscape and Java, found to be illegally exclusionary by the District Court and the Court of Appeals, had the sole purpose of protecting the applications barrier to entry. As shown below, however, the deficiencies in the proposed settlement are so pervasive that SBC's competitive, Internet-based offering and similar products from other companies are threatened with the same fate as Netscape and Java.
Under the proposed settlement, Microsoft could use its monopoly power to (i) prevent UMS and related products from being accessed by anyone using a Windows PC and IE; (ii) degrade or impede the ability of UMS to function on a Windows PC; and/or (iii) deny UMS access to the Windows desktop. Moreover, Microsoft could avoid the requirements of the proposed settlement by simply claiming UMS was a "service." See pp. 46-48 supra. Because the proposed settlement does not require Microsoft to ensure that UMS will function smoothly on Windows, or even have proper access to the Windows desktop, it could significantly harm the ability of UMS to compete with Microsoft products providing similar services. The danger, of course, is not limited just to UMS, but applies equally to related communications and entertainment products and services that are being developed and offered by other companies.
The inadequate interoperability/disclosure provisions in the proposed settlement would allow Microsoft to completely block all access to a competing product, like UMS, for all users of a Windows PC and IE. Because the proposed settlement does not require any disclosure to ensure interoperability between IE and a non-Microsoft server operating system, Microsoft is able, and indeed encouraged, to change, and then withhold disclosure of, IE's protocols in order to prevent interoperability with those SBC servers that run on non-Microsoft operating systems. In that case, all UMS customers would be unable to access their UMS account from a Windows PC operating system equipped with IE.
Should SBC decide to convert the entire UMS network of servers to Microsoft operating systems, Microsoft would still, under the proposed settlement, be able to block access to UMS for some users. In this event, Microsoft could merely program its server operating system so that it could not interoperate with a non-Microsoft browser. In fact, Microsoft employed this very strategy recently, when it reprogrammed its web servers for the MSN website to block all access by consumers using the competing Netscape and Opera browsers. See p. 80 supra.
Should Microsoft choose not to completely block access to UMS, the proposed settlement permits Microsoft to substantially degrade UMS' functionality on a Windows PC operating system. The degradation can be accomplished by programming Microsoft's PC operating system in such a way that UMS' functions are purposefully disadvantaged. For instance, by altering the program codes for IE or Microsoft's version of Java, Microsoft can hinder the performance of UMS on a Windows PC operating system, including the speed at which UMS processes requests, its efficiency and the graphical presentation the user sees.
SBC's strategy for UMS is largely dependent on having access to and visibility on the desktop, as well as on OEMs and end-users being able to change default settings in the Windows operating system to select the SBC home page or set the pre-login screen to allow for message notification. Without provisions in the settlement guaranteeing these rights, Microsoft can prevent UMS from having its own icon on the Windows desktop, or being on the Windows start menu. Furthermore, the proposed settlement would allow Microsoft to prevent UMS customers from choosing to set the SBC-Yahoo homepage as their default homepage on IE. Likewise, nothing in the proposed settlement would stop Microsoft from denying SBC equal access to the pre-login indicators on the desktop for message notification. Only through such notification could UMS compete with Microsoft's Hotmail or Messenger, which have such indicators on both the desktop and the pre-login screen for the purpose of notifying the subscriber that messages are waiting.
By not requiring "seamless interoperability," the proposed settlement would allow Microsoft to raise substantially the fixed costs associated with a competing product or service, as well as the ultimate cost to the consumer, to the point that such products are unable to compete. The anticompetitive initiatives Microsoft can pursue under the proposed settlement will force Internet-based businesses to move toward using Microsoft server operating systems and software exclusively. At the same time, the settlement will channel consumer access to the Internet through Windows PC operating systems, which consumers will have to purchase in order to obtain IE.
For example, SBC currently uses Microsoft server operating systems for less than 5% of its UMS server network, and anticipates this percentage will approach zero within the next few years -- provided there is full interoperability among servers, PCs, PDAs, phones and all other computing devices. If the proposed settlement is approved, however, SBC will probably have no choice but to replace its entire UMS server network at considerable cost with Microsoft server operating systems and software.
Microsoft's server operating systems are currently significantly more expensive than those of its competitors, and this price differential is likely to grow as Microsoft solidifies its position in the server market. Even at current prices, replacing the UMS server network with Microsoft server operating systems would significantly increase SBC's overall costs for UMS (including the cost of hardware, software, maintenance of hardware and software, staff, network management, and disaster recovery). In addition, if SBC were required to replace the hardware in its entire current server network within the next few years, this would dramatically increase costs for UMS during its critical first years on the market. Taken together, these cost increases will make UMS a much less appealing alternative to the Windows PC, as the consumer's cost savings will be much smaller.
UMS is not the only SBC network service to be negatively impacted. SBC's telephone network also utilizes many servers for various functions. For instance, SBC has servers that perform "operation support systems," as well as "operations administration maintenance and provisioning." Like UMS, under the proposed settlement these network support systems would be at risk, and the cost to convert them all to Microsoft server operating systems would be significant.
UMS is designed to save consumers money, because they can access the service through a "dumb PC" or handheld computing device, equipped only with a few features, like UMS, that will allow the consumer to use the Internet to perform essentially all major computing functions currently offered by a Windows PC at a fraction of the cost.
Assuming full interoperability, a "dumb PC" with a browser would be able to access and browse the Internet as well as a Windows PC. Moreover, by purchasing a product like UMS to accompany the "dumb PC," even more of the functions of a Windows PC (voice/data messaging, instant messaging, e-mail), would be available and the total price would remain substantially below a Windows PC. The cost savings to the consumer, when coupled with the other positive attributes of UMS, would make it a very attractive alternative to the comparable Microsoft products the consumer can only obtain by purchasing a Microsoft operating system.
By preventing full interoperability throughout the network of servers (including the browser), however, Microsoft can destroy any cost savings provided by UMS, because the consumer will have to buy a Windows PC or another device with a Microsoft operating system in order to obtain IE. That operating system already will have integrated or bundled Microsoft products that compete with UMS (MS Messenger; Outlook Express, Hotmail) which are included in the cost of the operating system. Thus, to use UMS, the consumer will have to pay an "add-on" or double cost in addition to the cost of the Microsoft operating system. The competitive disadvantage, for both SBC and the consumer, is plain.
SBC, or any Internet-based business, is highly dependent on the quality and speed of technological innovation in markets that supply the hardware and software used in new ventures, such as UMS, or established services. This is particularly true for UMS, which will rely on an extensive network of servers, computers and related devices and technologies to deliver its range of services. SBC's policy is to use a wide variety of suppliers for different products, including software, throughout the UMS networks and its other computer systems. To take full advantage of technological innovation, SBC chooses the "best in class." This term reflects SBC's philosophy to choose the provider for each particular product or service that best performs the specific functions needed by SBC. In this way, SBC can obtain maximum benefit from the speed and diversity of innovation to create the most competitive products possible.
In certain crucial markets, the proposed settlement will convert this world of choice into a world of one choice -- Microsoft. In the process, SBC's ability to employ its "best in class" strategy would be severely reduced. Indeed, technological innovation itself will be gravely hindered, particularly for server operating systems and server software -- two critical product areas for any network. This change will be driven by the proposed settlement's failure to ensure interoperability. As a result, consumers and businesses will be forced to purchase the Windows PC operating system, IE, and Microsoft server operating systems, or at least license Microsoft intellectual property, in order to guarantee full interoperability. The fact that Microsoft server products, especially in relation to the particular needs and functions of SBC, are more costly but by no means superior in quality or functionality, will no longer be determinative.
In the longer term, as competitive choices in these markets are diminished, Microsoft will be able to unilaterally control the pace of innovation. Currently, many different companies are working to innovate and develop different product functions and niche uses. SBC can take advantage of specialized innovations that are essential to supporting or improving its operations. In the world created by the proposed settlement, however, Microsoft will be the sole arbiter of what areas, products or uses should or should not be explored for technological advancement. Microsoft would be free to stifle innovation in a particular area that may be crucial to developing a product or service which competes with Microsoft.
The government made this very point to the District Court through the testimony of one of its expert economists, who stated that Microsoft's exclusionary practices had
Romer Decl. ¶ 5.
The effect of the proposed settlement on UMS is illustrative. SBC currently uses between 15 and 20 different providers for different products, including software, throughout the UMS network. In the "Microsoft world" that would be created by the proposed settlement, SBC expects that it would be limited to a total of 5 to 10 providers. Each of the lost providers will have to be replaced by Microsoft products because of interoperability obstacles created by Microsoft commingling and nondisclosure.
SBC currently uses three separate server operating systems in its UMS operations, based on the particular needs and functions of different servers throughout the network. The three products are a UNIX operating system which is run on servers produced by several companies; Linux operating systems which run on a variety of servers; and, for certain limited functions, Microsoft server operating systems (which are run on less than five percent of the UMS network). In SBC's view, the Microsoft server operating systems are substantially less effective than competitors' systems in performing the functions needed by the UMS network. Among other things, Microsoft products have a poor security performance history, see pp. 90-91 supra, and thus it is SBC's policy that no server that faces the Internet is based on a Microsoft server operating system. As a result of the need for interoperability, however, under the proposed settlement, SBC could be forced to use Microsoft server operating systems throughout the UMS network.
SBC's choices will be similarly circumscribed for server software. For its mail server and directory server, SBC currently uses a range of non-Microsoft software products. For the functions needed by the UMS network, these products better meet SBC's requirements, and are less costly than, comparable Microsoft products. For SBC's web server, which is the critical primary link to the Internet and the Internet browsers, SBC uses server software from Apache. The Apache software is also preferable to comparable offerings from Microsoft for this purpose.
Similarly, in the areas of network management and voice over Internet protocol ("VoIP"), SBC could well be forced to switch to Microsoft products under the proposed settlement. Network management products essentially ensure that SBC's telephone network runs effectively and reliably, by monitoring the system for failures, analyzing configuration, and developing utilization statistics. The network management software currently used by SBC is more efficient and less costly than comparable Microsoft products. In the VoIP area, SBC's "gateway," which translates voice conversations into VoIP, uses a non-Microsoft, ITU ("International Telecommunications Union") H.323-compliant gateway, or translator. If the proposed settlement were approved, nothing would prevent Microsoft, by changing its server or browser programs to a non-compliant format, from forcing SBC to replace this translator with a Microsoft product or a product that incorporates Microsoft-licensed intellectual property to ensure the interoperability required for SBC's telephone network to function.
The flexibility provided to Microsoft to delay the disclosure of technical information could also significantly harm the competitive prospects of UMS and related products. As explained above, Microsoft can delay the release of technical information related to interoperability with a new Windows product or a change in an existing Windows product until very close to the time when the new or altered product is placed on the market. See pp. 80-85 supra. SBC, and other competitors, thus face the very real possibility that there will be insufficient time to ensure that a competing product, like UMS, will be fully interoperable when a new or altered Windows product is introduced. If UMS is not fully interoperable, the result would be that certain UMS customers who attempted to access their mailbox from a Windows PC would be "dropped" -- meaning their request for data or voice information would not be processed. Since SBC plans that UMS will have over 5 million customers by 2007, if UMS is not fully interoperable with Windows PCs for even one day, many UMS customers would be "dropped." The negative consequences should customers be unable to access their UMS mailboxes for any period of time are obvious.
By making it impossible for server operating systems which run websites on the Internet to interoperate with non-Windows operating systems, Microsoft will be able to force all businesses and all consumers to conform to the Windows world. With over 90% of the operating system and browser markets, for example, Microsoft can make its Passport product the dominant intermediary between consumers and e-commerce web sites, and can effectively impose a tax or toll on all transactions. If it wishes to, Microsoft will be able to use its browser dominance to ensure that any web portal in which it has a stake receives preferential display. And if it chooses to, Microsoft will be able to use its browser dominance to control the flow of information or content on the Internet by using its Internet gate-keeper position to prefer one type of message over another, for example, blocking access to sites that are critical of Microsoft. In all of these ways, Microsoft can use the developing world of Internet technology to protect and strengthen its PC operating system monopoly.
As shown, the proposed settlement allows Microsoft to continue exclusionary practices, like commingling, to easily evade virtually all of the proposed settlement's proscriptions and affirmative obligations, and, by simply doing what the settlement allows, or fails to enjoin, ensuring that the next generation of threats to the operating system monopoly in the form of Internet servers and web-based computing never leaves its crib. By the time the proposed settlement expires, Microsoft middleware will be the firmly entrenched medium of Internet communication, displacing open architecture with a closed proprietary system; all competition will be forced to use Microsoft's proprietary standards instead of the open architecture currently thriving as the medium for program development and communication; and OEMs will be even more beholden to Microsoft's demands. This is the teaching of an exhaustive trial record and a careful appellate review that affirmed the lower court's findings of a decade-long scheme of monopoly maintenance in violation of Section 2 of the Sherman Act.
Under the law of the District of Columbia Circuit, the proposed settlement falls far short of providing any meaningful remedy for this most serious of antitrust violations and for that reason alone does not satisfy the public interest standard. United States v. Microsoft, 56 F.3d 1448, 1458 (D.C. Cir. 1995) (The Court must "make an independent determination whether the proposed consent decree [is] in the public interest."). Worse still, because the proposed settlement operates to perpetuate this unlawful monopoly by permitting the continuation without sanction of conduct previously found to be exclusionary, the proposed settlement in fact injures the public interest. For these reasons, the District Court must reject it.
Indeed, the proposed settlement fails to meet the public interest standard on all of the bases set out in the 1995 Microsoft decision. As discussed above, and in further detail below in the context of recommended changes and additions to the proposed consent decree, the settlement now advanced by the government (1) is "ambiguous" in many respects and riddled with loopholes and exceptions to key provisions; (2) presents numerous "difficulties in implementation" that arise because so many provisions leave compliance in Microsoft's sole discretion; (3) has been subject to widespread condemnation by "third parties [who] contend that they would be positively injured by the decree," including SBC; and (4) in view of the remedies the government said were absolutely essential only two years ago -- after the District Court's detailed findings as to monopolization were upheld on appeal -- but which have now been omitted from the proposed decree, "on its face and even after government explanation, appears to make a mockery of judicial power." Microsoft Corp., 56 F.3d at 1462.
For these reasons, SBC submits that the proposed decree should not be entered. The proposal made by the Litigating States would, if adopted in its entirety, adequately serve the public interest in SBC's view. Alternatively, SBC respectfully offers the following detailed revisions that, if fully incorporated into the proposed settlement, would provide an appropriate remedy for Microsoft's adjudicated wrongdoing:
Sections 8 and 9 of the Litigating States' proposal provide an adequate remedy prohibiting retaliation by Microsoft against others. Alternatively, the following revisions should be made to the proposed settlement:
§§ III.A.1, III.A.2 and III.A.3
Proposed Additions to Follow RPFJ § III.A
A provision, such as Litigating States' proposal § 9, should be added following § III.A to specifically prohibit Microsoft from retaliating against any individual or entity who participates or cooperates in any way in any aspect or phase of antitrust litigation involving Microsoft. Such a provision will ensure that Microsoft does not retaliate against any individual or entity that has participated thus far, and will afford protection to any individual or entity that wishes to come forward with complaints against Microsoft based on the consent decree that is ultimately entered. In view of Microsoft's continuing dominant position, its history of retaliation, and the fear it has engendered throughout the marketplace, such a provision is both necessary and reasonable.
Litigating States' proposal § 14 should be included in the decree to require Microsoft to continue making and updating a version of its "Office" business productivity suite that can be ported to the Apple operating system, and to require Microsoft to auction licenses to third parties to facilitate the creation of versions of Office that port to operating systems other than Windows. Such a provision is necessary because the Apple Macintosh operating system at present is the only viable alternative to Windows as an Office platform, but if others develop, they should have access to this widely-used application software. Such a provision is justified by the specific findings, affirmed on appeal, that Microsoft used the threat of dropping support for the Apple version of Office to coerce Apple into using IE as its default browser.
The subject of uniform licensing is adequately addressed in § 2(a) and § 2(b) of the Litigating States' proposal. Alternatively, the RPFJ should be revised as follows:
For the reasons discussed in connection with RPFJ § III.A., the provision must apply not only to Microsoft's licensing to OEMs, but to all third party licensees.
The proposed decree should allow Microsoft to provide reasonable volume discounts only if they are based upon an independent determination of the actual volume of shipments. See p. 102 supra; Litigating States' § 2(a)(ii).
This provision and its three subsections should be eliminated. Instead, the provision should include an outright prohibition, such as that included in Final Judgment § 3(a)(ii) or Litigating States' § 2(a), against Microsoft's offering market development allowances ("MDAs") or discounts to OEMs or third party licensees. This loophole allowing MDAs makes it possible, as a practical matter, for Microsoft to engage in the very discrimination that the provision is intended to prevent.
Proposed Additions to Follow RPFJ § III.B
A provision should be added to the proposed decree that would require that Microsoft provide OEMs and other licensees with equal access to "licensing terms, discounts, technical, marketing and sales support, product and technical information, information about future plans, developer tools or support, hardware certification and permission to display trademarks or logos." See Litigating States' § 2(b); Final Judgment § 3(a)(ii). Without this provision, Microsoft will be able to keep such information secret, which will allow Microsoft to continue to take advantage of licensees' ignorance about what terms are available.
A provision should be added that prohibits Microsoft from enforcing any contract terms or agreements that are inconsistent with the decree. See Litigating States' § 2(a); Final Judgment § 3(a)(ii). This prophylactic measure is justified by Microsoft's proven history of evasion of antitrust regulation and anticompetitive conduct.
Section III.C attempts to address Microsoft's past illegal imposition of restrictions on OEM configuration of the desktop. These restrictions closed the OEM distribution channel to non-Microsoft middleware. Because the provision fails to address Microsoft's commingling of code, contains no affirmative requirement to offer a stripped-down version of Windows with a corresponding price reduction, and is riddled with loopholes and ambiguities that allow Microsoft to override both OEM and end-user choices regarding competing middleware products, section III.C fails to accomplish its goal. To effectively close these loopholes and reopen the OEM distribution channel in an effort to revive middleware competition, SBC recommends the following:
The Litigating States' proposal is adequate to satisfy SBC's concerns regarding the effectiveness of OEM configuration options. SBC recommends that the Litigating States' proposals addressing restrictions on OEM options be adopted to replace section III.C of the proposed settlement. See Litigating States' §§ 1, 2(c), 3, 8, 10. In the alternative, SBC recommends the following modifications to the proposed decree:
This subsection is meant to ensure that OEMs are free to install competing middleware products and services and to place icons and shortcuts to those products on the desktop. CIS at 30. To fulfill that purpose:
In this section concerning an OEM's freedom to promote a competing IAP, the settlement must either identify what a "reasonable technical specification" is or otherwise remove that determination from Microsoft's sole discretion. Otherwise, Microsoft will be able to block OEMs from featuring competing IAPs for virtually any reason, or else impose anticompetitive requirements, such as the use of Microsoft's proprietary protocols, before a competing IAP is allowed on the desktop.
Proposed Additions to Follow RPFJ § III.C
A provision that reads similarly to the Litigating States' proposed § 1 should be adopted to prevent Microsoft from repeating the illegal conduct that the Court of Appeals found it engaged in by commingling the code of its IE browser with the code for its Windows operating system. A restriction on the practice of binding middleware to the operating system is essential to restoring competition by making the OEM distribution channel a viable option once again for software vendors. See Litigating States' § 1. Such a provision will have the salutary collateral effect of preventing the exceptions contained in RPFJ § III.H from rendering the substance of sections III.C and III.H meaningless, as well as supporting innovation. See Shapiro Decl. at 23 (stating that an anti-binding provision in the Final Judgment, similar to the one proposed here, "strikes an excellent balance between the consumer benefits that can arise when Microsoft adds functionality to its operating system and the benefits that consumers enjoy when new and improved software is developed independently of Microsoft, especially if that software may serve a role in eroding Microsoft's monopoly position. By allowing OEMs to choose whether to make Microsoft's Middleware Products or rival software directly available to end-users, OEMs will have the incentive to experiment to best serve consumers' interest.").
Either as part of the provision dealing with the binding of middleware or elsewhere in the decree, there must be a provision requiring Microsoft to differentiate its product prices based on an OEM's selection of the Microsoft middleware products, if any, that it chooses to bundle with the Windows operating system. Such a mechanism must ensure that "stripped-down" versions of Windows cost less than the fully loaded version in a proportion that properly reflects the value of middleware products not included. Failure to provide a pricing mechanism, such as those contained in States' proposal §§ 1 and 3(b), removes any incentive OEMs have to create software packages composed of competing middleware products.
Several such mechanisms are possible. The Final Judgment provided that pricing be guided based on bytes of code. See Final Judgment § 3(g)(ii). SBC believes it would be preferable to allocate costs between the operating system and the removed middleware based on measurements of "function point code." The International Function Point Users Group Counting Practices Manual is a generally accepted, objective industry standard for measuring non-multimedia software (which excludes games and training software) and estimating software costs using an existing code base. See T. Capers Jones, Estimating Software Costs Function Point Analysis: Measurement Practices for Successful Software Projects (1998); David Garmus and David Herron, Function Point Analysis: Measurement Practices for Successful Software Projects, 34-61 (2000). Alternatively, SBC supports the use of a pricing mechanism based on the fully allocated product development costs for the operating system product and middleware products in question. See Litigating States' § 1.
SBC recommends adoption of Litigating States' proposed § 3. Section 3 mandates that Microsoft continue to license for 5 years its immediate predecessor version of Windows, at a price no higher than the last price at which the predecessor version was offered. This is a further means of preventing Microsoft from commingling its middleware products with Windows without offering OEMs, end-users and third parties the chance to buy a version of the operating system that is both cheaper and without Microsoft products bound to it. Prior versions of Windows typically have less Microsoft middleware product bundled with or bound to the operating system, and rely more heavily on accepted industry standards. As a result, predecessor versions may be more easily configured to include non-Microsoft middleware products.
Full interoperability is necessary to prevent Microsoft from perpetuating its monopoly of the PC operating system market by exercising control over server operating systems and software, and Internet browsers, and using that control to eliminate the nascent competitive threats posed by non-Microsoft server operating systems and embedded devices.
Proposed Additions To Follow RPFJ § III.E
The Litigating States' proposal properly requires Microsoft to distribute Java, free of charge, for ten years. Litigating States' § 13. The copy of Java that is distributed must be "a competitively performing Windows-compatible version of the Java runtime environment (including the Java virtual machine and class libraries) compliant with the latest Sun Microsystems Technology Compatibility Kit." Id. The proposed settlement does not require Microsoft to distribute a version of Java that is compliant with the latest technology from Sun Microsystems, and that is fully compatible with the most recent version of Windows. This requirement is critical to ensure full interoperability between IE and all non-Microsoft server operating systems, and will also help to erode the applications barrier to entry that shields Microsoft's monopoly power.
The government's own expert explained the need for an affirmative prohibition against such interference by Microsoft as necessary to prevent one of the more insidious methods of monopoly maintenance:
Shapiro Decl. at 22. The Final Judgment and Litigating States' proposal explicitly prohibit Microsoft from knowingly impeding or degrading the performance of non-Microsoft Middleware on a Windows PC. Final Judgment § 3(c); Litigating States' § 5.
The Litigating States' proposal also properly requires that if Microsoft takes any action that would "interfere with or degrade the performance of non-Microsoft Middleware," it must give 60 days advance notice to the affected ISV. Litigating States' § 5. The proposed settlement does not contain a knowing interference provision. Since the Court of Appeals specifically affirmed the findings upon which this remedy was based, the decision to delete it is difficult to understand. CA at 65-66. To the contrary, the proposed settlement actually gives Microsoft the incentive to make slight changes to its operating system product, as part of a "minor upgrade," that would have the effect of impeding the interoperability of non-Microsoft middleware with a Windows PC operating system. See RPFJ §§ III(D), VI(J). If the change is a part of a "minor upgrade," Microsoft is not required to disclose the APIs and other technical information required to ensure full interoperability. Id.
To create a level playing field and foster competition, a provision must be added to ensure that open or industry standards continue to be promoted and used by Microsoft as part of the Windows PC operating system environment. An industry standard is any technical standard that has been approved by (or has been submitted to and is under consideration by) any independent, publicly recognized organization or group that sets standards. If Microsoft can replace an open industry standard with its own proprietary codes, it will prevent full interoperability and thus reinforce the applications barrier to entry.
As a result of Microsoft's monopoly power in the PC operating system market, it is able now, and in the foreseeable future, to depart from industry-recognized standards for its own competitive advantage. This is accomplished in two ways. First, Microsoft has in the past made subtle, undisclosed changes to a number of recognized industry standards that are used to execute functions by the Windows operating system. Even a small modification can severely impede the ability of a competing operating system or middleware product from interoperating with a Windows operating system product.35 Second, any new or modified standards implemented by Microsoft become, as a practical matter, an industry standard within a very short period of time because of the high percentage of Windows users.
Microsoft's Brad Silverberg explained this Microsoft strategy in the context of a previous standards battle with Novell's Netware:
Henderson Decl. ¶ 35 (internal citations omitted).
To ensure that Microsoft's practices are changed and to ensure full interoperability, the settlement must include a provision that requires:
(i) that Microsoft continue to use and promote all open or industry- recognized standards;
(ii) that Microsoft not alter or modify an industry standard in any way, except to the extent that such modification is compliant with, and approved by, an independent, internationally recognized industry standards organization;
(iii) that Microsoft disclose any change it implements to an open or industry-recognized standard, in a "timely manner," as that phrase is defined in the Litigating States' Proposal § 22 (pp);
(iv) that Microsoft assist any other software provider to achieve interoperability with any protocol Microsoft uses in any such situation in which Microsoft is the holder of the reference protocol implementation; and
(v) that Microsoft work with all other holders of reference protocols to achieve and ensure interoperability with any protocol Microsoft uses, in any situation in which Microsoft is not the holder of the reference protocol implementation.
There are over 300 separate standards used by any PC operating system to function within a local area network or on the internet. The following protocol families are among those that are particularly important to Internet-based computing: (1) the TCP/IP protocol family, which is universally used to transmit data and services on the Internet; (2) the H.323 protocol family as defined by the ITU, which supports video and voice communications and is often referred to as a Voice over IP (VoIP) protocol; (3) the SIP protocol family, which supports video and voice communications, as well as instant messaging; and (4) the HTML/HTTP protocol family, as defined by the World Wide Web Consortium (W3C), which supports web browser and server protocols.
Microsoft's control of IE is an integral part of the anticompetitive conduct that has maintained Microsoft's monopoly over the PC operating system market. As the Litigating States propose, to remedy these anticompetitive acts and prevent recurrence, the source code for IE must be disclosed on a royalty-free and non-discriminatory basis. See Litigating States' § 12.
Sections 8 and 9 of the Litigating States' proposal adequately address retaliation issues. Alternatively, the following revisions should be made in the RPFJ:
Proposed Additions to Follow RPFJ § III.F
A provision should be added, such as Litigating States' proposal § 11, prohibiting Microsoft from offering consideration to any competitor in exchange for the competitor's agreeing to refrain from developing or distributing any product or service that competes with any Windows Operating System or Middleware Product. See also Final Judgment § 3(h). Such a provision is necessary to prevent Microsoft from seeking anticompetitive contracts that divide markets or otherwise limit competition, regardless of whether the "terms" are reasonable. See discussion of RPFJ § III.F.2 supra.
SBC believes that Section 6 of the Litigating States' proposal is consistent with the public interest on the issue of exclusive dealing. Alternatively, the following changes should be made in the RPFJ:
Section III.H focuses on OEMs' ability to offer and promote and end-users' ability to choose competing middleware products. Yet the provision undermines this purpose in several ways, including: (1) preventing either OEMs or end-users from removing Microsoft products from the operating system; (2) permitting Microsoft to override or alter OEM and end-user choices of competing middleware products; and (3) delaying the implementation of the provision to such an extent as to render it meaningless for a fifth of the lifespan of the decree. To remedy these flaws, the following changes are suggested:
The "Notwithstanding" Clauses In This Provision Must Be Deleted
The "Notwithstanding" clauses at the end of section III.H allow Microsoft to disregard OEM and consumer choice whenever Microsoft decides that its products must be invoked to operate with its servers or when the Non-Microsoft Middleware Product fails to implement a "reasonable technical specification" (a term that is left to Microsoft to define). The clauses should be eliminated in their entirety. The exceptions contained in these clauses are so broad that they threaten to render the substance of section III.H meaningless by permitting Microsoft to override an OEM's or end-user's middleware default choice at will.
The proposed settlement allows Microsoft to charge a royalty for the required license of technical information concerning interoperability, and to obtain a cross-license to the licensee's technical information used to interoperate with a Windows PC operating system or Microsoft Middleware. As the government recognized in the earlier remedy proceedings, royalty and cross-licensing requirements are anticompetitive. Final Judgment § 3(i); see also Gov't D.Ct. Sum. Resp. at 14. As the Litigating States have proposed, Microsoft should be required to license the necessary technical information on a royalty-free basis, and without the right to a cross-license from the licensee. Litigating States' § 15.
Section III.J of the proposed settlement is a loophole and must be deleted. All APIs, communications interfaces and technical information that must be disclosed to ensure interoperability serve, at least in part, an authentication or encryption function related to the security of the operating system. Microsoft should not be given an excuse to withhold disclosure of crucial technical information for potentially anticompetitive purposes. Neither the Final Judgment nor the Litigating States' proposal contains a similar provision.
In contrast to the proposed settlement, certain aspects of the Litigating States' proposal would be far more effective in ensuring that the intent and spirit of the final relief entered in this action be effectively enforced:
The way the proposed settlement defines key terms significantly restricts, and in many instances eliminates, the effect of the proposed settlement's substantive provisions. SBC generally recommends the adoption of the definitions contained in the Litigating States' proposal § 22. Some of the problems posed by the proposed settlement's definitions relating to middleware are as follows:
For the reasons stated herein, the proposed settlement with Microsoft is contrary to the public interest and should be substantially modified or rejected entirely.
January 28, 2002
1 As recently as January 2002, Microsoft controlled over 96% of the entire PC operating system market, and Apple's Macintosh operating system had only a 2% share. Is Apple Out of the Running in the Operating Systems War? (Jan. 8, 2002), at http:// www.websidestory.com/cgi-binwss.cgi? corporate &news&press_1_163.
(1) the competitive impact of such judgment, including termination of alleged violations, provisions of enforcement and modification, duration of relief sought, anticipated effects of alternative remedies actually considered, and any other considerations bearing upon the adequacy of such judgment;
(2) the impact of entry of such judgment upon the public generally and individuals alleging specific injury from the violations set forth in the complaint including consideration of the public benefit, if any, to be derived from a determination of the issues at trial.
15 U.S.C. §§ 16(e)(1)-(2).
4 The terms "binding" or "commingling of code" refer to including software or a link to web-based software in an operating system product in such a way that either an OEM or end-user cannot readily remove or uninstall the code without degrading the performance or impairing the functionality of the operating system. "Bundling" refers to the sale or marketing of different software products in a single package, but without commingling of their codes.
5 Even the settlement's definition of "trademark" is so broad as to further limit the scope of the decree: "`Trademarked' means distributed in commerce and identified as distributed by a name other than Microsoft® or Windows® that Microsoft has claimed as a trademark or service mark by (i) marking the name with trademark notices, such as ® or TM , in connection with a product distributed in the United States; (ii) filing an application for trademark protection for the name in the United States Patent and Trademark Office; or (iii) asserting the name as a trademark in the United States in a demand letter or lawsuit. Any product distributed under descriptive or generic terms or a name comprised of the Microsoft® or Windows® trademarks together with descriptive or generic terms shall not be Trademarked as that term is used in this Final Judgment. Microsoft hereby disclaims any trademark rights in such descriptive or generic terms apart from the Microsoft® or Windows® trademarks, and hereby abandons any such rights that it may acquire in the future." RPFJ § VI.T. Thus, Microsoft may release a new middleware product entitled WindowsTM Telephone, for example, and because the name is descriptive rather than trademarked, it would not be considered "Microsoft Middleware" under the terms of the decree.
6 Significantly, by omitting Microsoft Office from the list of middleware products, the government has eliminated from the proposed settlement a middleware product that provides Microsoft with a de facto monopoly in the middleware market. As of March 1997, Office's market share had reached 90%, a figure that has likely grown since that point. See Jesse Berst, Office Suites for Free, ZDNet AnchorDesk (March 7, 1997), at http://www.zdnet.com/anchordesk/story/story_743.html; see also, Benjamin Woodhead, Microsoft's Australian Monopoly? Let the U.S. Handle It, iTNews (Nov. 17, 1999), at http://www.itnews.com.au/story.cfm?ID=507 (referring to the lack of recent statistics on Office Suite's market share, "We don't bother to measure that market anymore because Lotus and Corel have been squeezed out of it ... No one will pay for that sort of research because everyone knows what the answer is.").
7 Describing the functionality of a product in terms of the categories of applications, rather than the operation of the product, also limits the effectiveness of section III.H of the proposed settlement, which relies heavily upon the definition of "Microsoft Middleware Product" to set the parameters of non-Microsoft middleware access to the OEM distribution channel.
8 An example of cross-dependency is the link between IE and Microsoft Word in the Windows Operating System Product. Even if an end-user has selected Navigator as her default browser, IE may automatically launch if the user clicks on a URL, (i.e., an Internet address) that is contained in a Word document.
"Middleware" means software that operates, directly or through other software, between an Operating System and another type of software (such as an application, a server Operating System, or a database management system) by offering services via APIs or Communications Interfaces to such other software, and could, if ported to or interoperable with multiple Operating Systems, enable software products written for that Middleware to be run on multiple Operating System Products. Examples of Middleware within the meaning of this Final Judgment include Internet browsers, e-mail client software, multimedia viewing software, Office, and the Java Virtual Machine. Examples of software that are not Middleware within the meaning of this Final Judgment are disk compression and memory management.
Final Judgment § 7(q).
10 The definitions of "Windows Operating System Product" and "Personal Computer," read together, also create an ambiguity that places in doubt whether future versions of Microsoft's operating system will even qualify as a "Windows Operating System Product" under the proposed settlement. Windows XP is Microsoft's first PC operating system designed for shared or multiple person use. Microsoft has promoted XP's ability to facilitate home networks where many people can share devices and Internet connections. See Experience the Connected Home: Share One or Many Computers (May 9, 2001), at http://www.microsoft.com/windowsxp/home/evaluation/experiences/connectedhome.asp. Because Windows Operating System Product is defined as software "distributed commercially by Microsoft for use with Personal Computers," RPFJ § VI.U, and the definition of "Personal Computer" means "any computer configured so that its primary purpose is for use by one person at a time," RPFJ § VI.Q, if XP or its successors are distributed primarily for multiple users or employed for construction of mini-networks or servers, successor products could fail to meet the definitional requirements to be covered under the decree. See RPFJ § VI.Q (expressly excluding servers and other computing devices from the definition of Personal Computer).
11 "Binding harmed consumers who did not want Internet Explorer, by causing `performance degradation, increased risks of incompatibilities, and the introduction of bugs.'" Felton Decl. ¶ 84 (citing D.Ct. at ¶ 173).
12 In addition to the settlement's failure to prohibit commingling of code, the settlement also condones Microsoft's bundling of products with its operating system. Section III.C presents OEMs with a laundry list of options they may adopt in installing, displaying, and distributing Non-Microsoft Middleware, but nothing in the proposed settlement prevents Microsoft from forcing OEMs to accept additional products as part of the Windows Operating System Product that are included with the operating system. As a result, under the proposed settlement, OEMs can be forced to accept a complete package of Microsoft products with each license of the Windows operating system.
14 It is noteworthy that the binding of applications is not limited to browsers and internet-related services, but also includes common applications such as word processing. For example, Microsoft has used its operating system monopoly to motivate consumers to use Microsoft Word instead of Corel's Word Perfect. Regardless of the quality or perceived attributes of Word Perfect versus Word, many businesses and individual consumers use Word simply to avoid incurring the additional trouble and expense of licensing a second word processing application when the PC operating system already comes equipped with such a function.
15 The proposed settlement contains one exception to its blanket prohibition on third party alterations: it would permit a Non-Microsoft Middleware producer to designate that its product be invoked automatically in place of a Microsoft Middleware Product. However, the mechanism by which the producer may accomplish this is at Microsoft's discretion, and Microsoft may require confirmation from the end-user that he or she would like to accept this option. See RPFJ § III.H.2.
16 Section III.C.4 allows OEMs to offer alternative operating systems. While seemingly procompetitive, the government fails to acknowledge that there currently is no market for alternative operating systems. See CIS at 32. As the Court of Appeals explained, due in large part to network effects, there is no incentive for consumers to use or for ISVs to write programs for PC operating systems other than Windows. See CA at 49-50, 55. Moreover, it is unclear that it is technologically feasible to include multiple operating systems on the same PC without sacrificing significant amounts of storage capacity or speed. No similar provision appeared in the Final Judgment, a fact which suggests that it is mere window-dressing (no pun intended) and does nothing to eliminate the barriers to competition erected by Microsoft.
17 The Court of Appeals found no justification for the restrictions on OEM configuration generally, but did hold that "a shell that automatically prevents the Windows desktop from ever being seen by the user is a drastic alteration of Microsoft's copyrighted work, and outweighs the marginal anticompetitive effect of prohibiting the OEMs from substituting a different interface automatically upon completion of the boot process." CA at 63.
18 OEMs incur increased costs as a result of customer "hotline" calls to the OEM. CA at 61. The additional program code also reduces the storage capacity of the computer and the speed of the processor. This is yet another way that Microsoft is able to erode OEM and consumer incentives to use competing middleware products.
19 Although the Court of Appeals did not affirm the District Court's blanket conclusion that IE's override of competing default browsers was illegal in all circumstances (for example, when accessing Windows "Help" resources and updates on the Internet), CA at 65-67, the proposed settlement swings much further in the other direction in permitting Microsoft to write the rules of when such an override of a user's designated default middleware product will be permitted.
20 The last part of the "technical requirements" clause, moreover, puts the onus on ISVs to request the reason for the technical failure. Because ISVs are unlikely to be immediately aware that there is a technical failure on the part of their middleware, the burden must be placed on Microsoft to explain such overrides.
21 "Beta test" refers to the last round of testing for a new software product that is typically performed by sending the software product out to consumers and industry insiders both as a means of ironing out the kinks in the product and obtaining publicity for the impending release. There is no set date within the industry of when these "tests" are performed. They can occur months before or immediately proceeding a pending product release date.
22 As a general matter, interoperability is the ability of different computers, servers or other devices, regardless of whether they use the same software and hardware, to freely transmit and receive information to and from each other.
23 With regard to server interoperability, the proposed settlement only requires "client protocols" to be disclosed. As is fully explained below, to achieve full interoperability between a PC ("client") and a server, there must be disclosure of both client and server protocols, so that the server can accept and transmit data and services to the PC. By disclosing only the client protocol, only one-half of the transaction (PC to server) is achieved, thus defeating the server's ability to fully interoperate with the PC. See pp. 70-72 infra.
24 A protocol is a piece of an operating system's software code that allows the operating system to translate, and thus understand, the language of another computer or server that is attempting to transmit data. When a PC ("client") and a server are transmitting and accepting information or services between each other over the internet, server protocols allow the server operating system to accept and understand the information or services being transmitted from the client. In other words, the server protocols allow the server to transmit information to the PC by converting the information from the server's computer language to the PC's computer language. The client protocols perform the opposite task, allowing the PC to fully interoperate with the server. In order to process information from PC to server, and from server to PC, it is essential that both server and client protocols be provided. Without knowledge of the appropriate server protocols necessary to interoperate with a Windows PC, an ISV cannot design an operating system for a server which will properly interoperate with the Windows PC operating system.
25 In October 2001, Microsoft dispelled any doubt whatever that it would use its control of Internet Explorer and Microsoft server operating systems to exclude competing browsers, when it blocked access to its MSN.com server for Netscape's and Opera's competing browsers. See p. 80 infra.
26 The Litigating States' proposal properly requires full disclosure of all technical information necessary to design a non-Microsoft server operating system that would be fully interoperable with a Microsoft server operating system. Litigating States' § 4.
27 The Litigating States' proposal justifiably requires full disclosure of all technical information relevant to the interoperability between a Microsoft PC or server operating system and any embedded device. Litigating States' § 4.
28 When browsers connect to a web server, they send information identifying specifically which browser it is and the capabilities of that browser. Programmers often code their web servers to be aware of browser differences so that the web server can provide a richer end-user experience. It is unusual, to say the least, to use browser and web server capabilities in this way to deny access.
30 The listing of categories of protected parties, instead of applying the ban to all third parties, also presents a significant risk of omitting some competitors (e.g., SBC and Sun Microsystems) from the decree's protection because their businesses do not fit neatly into any standard category.
31 Microsoft Platform Software is defined as Microsoft Middleware and the Windows Operating System. The limitations of those defined terms which, among other things, exclude applications (including applications on Internet-based servers that would threaten the Windows monopoly), render the "retaliation" ban far too narrow. See pp. 36-40, 42-43 infra.
32 The proposed settlement does purport to limit Microsoft's ability to enter into agreements with ISVs which require the ISV to refrain from "developing, using, distributing, or promoting any software that competes with Microsoft Platform Software." RPFJ § III.F.2. However, as discussed supra at pp. 97-98, that same section creates an exception that permits such agreements with ISVs if they are in relation "to a bona fide contractual obligation of the ISV." Therefore, even as to ISVs, the restriction is rendered potentially meaningless.
33 Based on a review of the published cases, every consent decree that the government has entered in a Section 2 case since 1978 has been ten years or longer in duration, with the exception of the first Microsoft decree. That decree was not entered after a full trial on the merits and a finding of unlawful monopoly maintenance.
34 Prior to 1996, SBC was subject to line of business restrictions imposed by the AT&T Consent Decree or Modified Final Judgment ("MFJ"). These prevented SBC's entry into markets such as long distance telephone and imposed numerous affirmative obligations to assist actual and potential competitors. See U.S. v. AT&T, 522 F. Supp. at 186-95 (setting forth line of business restructions); United States v. Western Electric Co., 673 F. Supp. 525 (D.D.C. 1987) (upholding "core" line of business restrictions on local telephone companies), aff'd in part, rev'd on other grounds, 900 F.2d 283 (D.C. Cir. 1990). These provisions, which were of indefinite duration, were ultimately superseded by the Telecommunications Act of 1996, which replaced the MFJ with detailed regulatory obligations to preserve non-discriminatory access to the local telephone network, to require SBC to share its network elements at regulated prices, and to take affirmative actions to open its local network to competition as the price for entry into the long distance market. See Telecommunications Act of 1996, 42 U.S.C. §§ 251-59, and § 271 et seq. The duration, detail and substantial scope of those affirmative requirements and prohibitions, even when they were embodied in the MFJ, stand in marked contrast to the trivial and temporary prohibitions applied to the Microsoft monopoly.
35 For example, if Microsoft made subtle changes to the industry-recognized audio Codec standard, applications that used audio features, such as Real Player, would not be able to interoperate with a Windows PC operating system. If Real Player continued to employ the industry-standard Codec in its program, Windows PC users would be able to download that Codec to their Windows operating system, but would face the very real possibility that the program would not function with their Windows PC operating system as well as the competing Microsoft product, Media Player, which would, of course, be designed to run with Microsoft's modified Codec.