1625 Massachusetts Avenue, NW
Facsimile Cover Sheet
If you experience any problems with this transmission, call 202-955-6300 and speak with Jessica Green-Johnson.
Dear Ms. Hesse:
The attached are the Comments of NetAction and Computer Professionals for Social Responsibility on the Proposed Final Judgment in United States v. Microsoft. A copy has already been provided to you via E-mail.
Please feel free to contact me at 202-955-6300 with any questions or concerns.
THIS FACSTMILE IS INTENDED ONLY FOR THE USE OF THE ADDRESSEE
AND MAY CONTAIN CONFIDENTIAL OR LEGALLY PRIVILEGED INFORMATION.
BEFORE THE UNITED STATES DEPARTMENT OF JUSTICE
The Government's Competitive Impact Statement claims that "[t]he relief contained in the Proposed Final Judgment provides prompt, certain and effective remedies for consumers."1 However, any potential relief is far from "certain" or "effective" for the average (non-corporate) consumer, and relief certainly will not be "prompt" since it will arrive, if at all, only as Microsoft rolls out later versions of its Windows Operating System.
Unfortunately, the Proposed Final Judgment does not offer consumers any hope of relief in the market for the non-middleware software applications on which they rely, such as word processing and spreadsheets. Nor does the Proposed Final Judgment attempt to offer any hope of relief to consumers using Windows 95, Windows 98, Windows NT, or Windows 2000. By its terms, the Proposed Final Judgment only applies to Microsoft's dealings with third party developers for Windows 2000 Professional, Windows XP, and later Windows versions. Thus, to achieve even the uncertain benefits of the Proposed Final Judgment, consumers will have to pay a high price for new software and, as explained below, new hardware, including new computers. For these reasons, the Proposed Final Judgment is suspect in terms of both the public interest and the goals of antitrust relief described by the Government.2
Moreover, the Proposed Final Judgment must be rejected because the record does not permit the Court to determine, with any reasonable degree of comfort or certainty, exactly what relief the Proposed Final Judgment provides. The language of the Proposed Final Judgment, in combination with the numerous exceptions to its prescriptions, necessarily leaves the Court and the public at a loss to confidently predict what conduct is prohibited and what conduct is permitted. Indeed, in a number of specific instances, the exceptions provided for in the Proposed Final Judgment appear to enable Microsoft to escape large portions or even all of its obligations.
As the Court is fully aware, the Government's current lawsuit-now approaching its fourth anniversary-was triggered by Microsoft conduct that the Government thought it had prohibited in a previous consent decree. Only when it attempted to enforce the decree against that conduct did the Government discover that the language of the decree-language perhaps inserted by Microsoft and "protecting" its rights to innovate-could be interpreted to permit Microsoft to require that customers purchase its browser as a component of Microsoft's Windows Operating System, the exact conduct that the earlier decree ostensibly would have prevented.3 The very same sorts of ambiguity are evident in the Proposed Final Judgment. If it is approved, Microsoft and the Government will find themselves back in this Court yet again, arguing over interpretation, while non-corporate consumers are forced to endure yet another round of anticompetitive effects.
Because the Court has no power under the Tunney Act to modify the terms proposed by the parties, the Court must either reject the Proposed Final Judgment as inconsistent with the public interest, or order additional proceedings to clarify its terms.4 Such additional proceedings should provide the parties to the Proposed Final Judgment with an opportunity to prove that there has been an actual "meeting of the minds" with regard to the terms of the agreement. Only where that has occurred should the Court consider approving the Proposed Final Judgment; otherwise, rejection is the Court's only recourse.
NetAction is a national nonprofit organization dedicated to promoting use of the Internet for effective grassroots citizen action campaigns, and to educating the public, policymakers, and the media about technology policy issues. Among other projects, NetAction manages the Consumer Choice Campaign to focus public attention on Microsoft's growing monopolization of the Internet.
Computer Professionals for Social Responsibility ("CPSR") is a public-interest alliance of computer scientists and other interested individuals concerned about the impact of computer technology on society. CPSR provides the public and policymakers with realistic assessments of the power, promise, and limitations of computer technology and directs public attention to critical choices concerning the applications of computing and how those choices affect society.
I. A Large Segment of the Consumer Market Will Be Unable to Avail Themselves of the Limited and Uncertain Benefits of the Proposed Final Judgment Unless They Invest Substantial Sums in Hardware and Software Upgrades
A. The Proposed Final Judgment Would Do Nothing to Increase Competition for Software Applications
Despite Microsoft's substantial dominance in the market for software applications-such as email, word processing, and spreadsheets-the Proposed Final Judgment limits its modest proposed remedies to the market for "middleware." The Government's loftiest description of its anemic proposed remedy promises only to "restore the competitive threat that middleware products posed prior to Microsoft's unlawful undertakings."5 The "bulleted" enumeration of those benefits offered by the Government further clarifies that only the market for middleware is targeted for relief.6
It is certainly true, as the Government points out, that middleware poses-or perhaps more accurately "posed" when the case was filed nearly four years ago-a significant threat to the dominance of Microsoft's Windows Operating System. By exposing its own "APIs," middleware allows software developers to write applications that will run on multiple operating system platforms, thus decreasing the importance of any particular operating system.7 And the middleware category is particularly important as computing evolves towards a model in which users go outside their desktop/laptop hardware not only for their data-as Internet has taught them to do-but also for applications by which to interact with that data.
While the Government certainly was correct to make middleware an important focus of its case, the Government certainly is not correct to make middleware the sole focus of its proposed remedy. It is one thing to say that the existence of a competitive market for middleware could undercut Microsoft's monopoly of operating system software if computing moves away from a desktop/server environment to a Net-based environment. It is quite another to say, as the Proposed Final Judgment does, that the public interest is satisfied when consumers can achieve some benefits of competition only if computing moves away from a desktop/server environment to a Net-based environment.
It is simply not in the public interest-certainly not in the non-corporate consumers' interest-to conclude what by now is nearly a decade of Government antitrust litigation by providing for some uncertain possibility of middleware competition while ignoring the monopoly position that Microsoft has built in the applications market over that same period.
B. Consumers Will Be Forced to Buy Software and Expensive New Hardware To Get Any Benefits From the Proposed Final Judgment
Even assuming that the Proposed Final Judgment has the potential, over time, to create a more competitive market in the narrow middleware product line which is its sole aim, consumers will have to buy a very expensive "admission ticket" to obtain any of those benefits.
The Government acknowledges that relief should, at a minimum, end the unlawful conduct, prevent its recurrence, and "undo its anticompetitive consequences."8 Curiously, despite this acknowledgement, the Proposed Final Judgment is purely prospective: by its terms it would apply only to the conduct of Microsoft and the opportunities of third party software and hardware vendors in relation to Windows 2000 Professional, Windows XP, and later generations of Microsoft's Windows Operating System.9 While not evident on the face of the Proposed Final Judgment, this result flows from the interplay among the Proposed Final Judgment's operative provisions and definitions. All of Microsoft's proposed obligations would be limited to conduct relating to a "Windows Operating System Product" which is defined as the "software code ... distributed ... by Microsoft ... as Windows 2000 Professional, Windows XP Home, Windows XP Professional, and successors[.]"10
Thus, by its terms the Proposed Final Judgment would not even attempt to "undo [the] anticompetitive consequences" of Microsoft's conduct for consumers who continue to use earlier versions of Windows, including Windows 95, Windows 98, and even Windows Me. Perversely, consumers will have to fill Microsoft's coffers by purchasing upgraded operating system software in order to obtain relief. Under the Proposed Final Judgment, only those consumers who upgrade to Windows 2000 Professional or a later version of the Windows Operating System would see any of the benefits of increased competition in the range of software choices available to them.11
And the cost of admission to the realm of greater choice is not just the price of the new software product. As anyone who has attempted a system upgrade on his or her own can attest, it is an endeavor best left to professionals. The process is enormously complex, takes hours of time, requires the user to make numerous decisions, often without adequate information, and is prone to crashing the computer, requiring professional help for recovery.12 Even the IS departments in corporate America are wary of upgrading their entire user community before they have thoroughly tested both the new operating system and the process of upgrading to it. And outside corporate America-that is, for the average (non-corporate) consumer-the task is so difficult that most consumers avoid it, continuing to use the operating system that came with their computers, and changing operating systems only if and when they buy a new computer.
To compound the problem and increase the price of admission even further, each new generation of operating system is significantly more resource-intensive than the last, as clearly indicated by the "minimum system requirements" notice Microsoft includes on its packaging and on its web site. The "minimum system requirements" for random access memory (RAM) have doubled with each succeeding consumer version of Windows. Windows Me required 32 megabytes (MB) of RAM; Windows 2000 Professional required a minimum of 64 MB; Windows XP recommends at least 128 MB.13 The "minimum" CPU and hard disk requirements have accelerated even more rapidly.14 And as every computer user knows, the "minimum" hardware requirements rarely provide adequate performance under the new operating system. So consumers are forced to buy enhancements to their existing computers, such as more memory and larger hard drives, to "catch up" with the demands of the new operating system. Even the consumer who upgrades her computer faces significant performance limitations stemming from the processor and bus architecture of existing systems. This history of Microsoft operating system evolution is that a consumer will, as a practical matter, need to buy a new computer to make effective use of a new operating system.
The reality, then, is that even if the Proposed Final Judgment would allow the development of a more competitive market for middleware, consumers would have to purchase new software and new computers (or spend almost as much to upgrade their existing systems) to obtain any benefits. As a result, a large percentage of consumers would simply be unable to afford the price of admission in their homes, or in their schools, or in their libraries. No proposed antitrust remedy can be in the public interest when it excludes the most vulnerable members of the public from any potential benefits.
II. Important Provisions of the Proposed Final Judgment Will Not Accomplish Their Intended Purposes
The Proposed Final Judgment is vague or subject to evasion by Microsoft to such an extent that it is not in the public interest as it currently stands. In Tunney Act proceedings-particularly where the litigation was required by a failure of precision in a prior settlement decree-a court should insist on precision in the proposed decree, so that the task of enforcing the decree does not become unmanageable.15 As discussed below, however, the Proposed Final Judgment is riddled with provisions that make its interpretation and enforceability highly problematic.
A. Microsoft I Illustrates the Importance of Eliminating Ambiguities and Loopholes from the Proposed Final Judgment
In assessing the Proposed Final Judgment, there are important lessons to be learned from past dealings between the Government and Microsoft, in particular the unfortunate history of the first Microsoft consent decree. That decree provided, among other things, that Microsoft can not require OEMs, as a condition of a license to an operating system, to license another Microsoft product.16 Soon after that decree was approved, however, Microsoft integrated its web browser code into the Windows Operating System, causing the Government to seek an injunction. The decree, however, contained an exception that doomed its very object: it could not be "construed to prohibit Microsoft from developing integrated products."17 Because the appellate court found that the Government was unlikely to prevail on this question of integration, Microsoft was free to integrate its web browser into the operating system.18
It was only last year, after extensive litigation, that Microsoft's integration of the browser code into the Windows Operating System was finally found to be illegal as an improper effort to prevent entry by rival browsers.19 By that time, however, Netscape Navigator, the browser at which Microsoft had directed many of its tactics, was no longer a threat to Microsoft's monopoly. In a very real sense, the presence of a vague exception for "integration" in the first Microsoft decree, and Microsoft's ability to exploit that exception, made possible the monopolistic behavior and anticompetitive distortion that was at issue in Microsoft II.20 It is crucial that that scenario not be repeated here, and that the Proposed Final Judgment be cleansed of similar opportunities for Microsoft to "design around" the decree's restrictions. Regrettably, there are many aspects of the Proposed Final Judgment that require such cleansing.
B. The Proposed Final Judgment Will Allow Microsoft To "Design Around" Its Obligations
Despite the fact that ambiguous language is exactly the reason this case was initiated, the Proposed Final Judgment suffers chronically from the same defect. Loose language and a plethora of exceptions would permit Microsoft to "design around" the restrictions that are currently being proposed by the Government:
Each of these shortcomings is discussed in further detail below. Given the ambiguities and exceptions, consumers have no reason to expect that the Proposed Final Judgment would effectively prevent the Microsoft monopoly from expanding beyond the operating system and browser markets into every facet of digital life. For these reasons, the Proposed Final Judgment should be rejected or additional proceedings ordered by this Court.
Under Section VI.U of the Proposed Final Judgment, "[t]he software code that comprises a Windows Operating System Product shall be determined by Microsoft in its sole discretion." This overarching provision would permit Microsoft to unilaterally alter one of the bedrock terms of the Proposed Final Judgment and, thus, to alter the terms of the agreement. As in 1995, the actual implications of this provision of the decree are unclear (except, perhaps, to Microsoft). Nevertheless, it is not difficult to imagine instances in which Microsoft would attempt to, for example, integrate potential Microsoft Middleware Products into its Windows Operating System in order to escape its disclosure obligations (a possibility discussed further below).21 The provision would also allow Microsoft to implement code designed to make competing middleware products incompatible with the Windows Operating System and thus prevent consumers from using that product as is ostensibly permitted under Section III.H. Indeed, the overarching problem with this provision for both competitors and consumers is its ambiguity and the uncertainty that is associated with it. At a minimum, definitional control of the Proposed Final Judgment should reside first with the agreement itself and, next, with the Court; it certainly should not reside with Microsoft. There is hardly any point in a decree that gives the defendant the right to determine its meaning.
Similarly, under Section III.D of the Proposed Final Judgment, Microsoft must disclose to competitors any and all APIs used by Microsoft Middleware to function effectively on the Windows Operating System. However, because the definition of Microsoft Middleware appears to be limited to "software code that Microsoft distributes separately from a Windows Operating System Product,"22 the possibility arises that Microsoft could avoid the required disclosure of its APIs simply by integrating potential Microsoft Middleware into the operating system.23 Integration would have the same effect upon potential competitors as nondisclosure: the applications barrier to entry would remain impenetrable and innovation by anyone other than Microsoft would be prevented. Moreover, further integration of middleware products would permit Microsoft to extend its monopoly power into adjacent markets.24
Moreover, assuming that Microsoft determines to release future Microsoft Middleware, Section VI.J would egregiously permit it to determine when or even if its API disclosure obligations are triggered. Specifically, this provision would allow Microsoft to unilaterally determine which releases are "updates" to existing Microsoft Middleware and which are "new major version[s]" of such.25 Microsoft would avoid disclosure (triggered by release of a new major version) simply by denominating the release anything other than "a whole number or ... a number with just a single digit to the right of the decimal point."26 The implications of this type of control on the part of Microsoft make the API disclosure provisions of the Proposed Final Judgment effectively meaningless.
Even where disclosure of APIs to competitors was to occur under Section III.D, it would not be required until "the last major beta test release" of the relevant Microsoft Middleware.27 Microsoft would thus have two incentives with regard to the release of the updated version: (1) to push the date of release of its last beta test as close as possible to the release of a commercial product, and (2) to use the interval until the last beta test to plan a subsequent and improved version of the same software to be released once competition with the updated version is threatened. After release, potential competitors would hurriedly attempt to implement the APIs to enable their products to interoperate with the Windows Operating System. Meanwhile, Microsoft's product would have been commercially released and gaining market share. Thus, before competitors would have been able to convince consumers of the value of their products, Microsoft would have developed a subsequent and improved version of the product, effectively foreclosing competition. The timing of competitors' access to APIs is crucial for compettion to have any chance of developing. But the timing provided for in the Proposed Final Judgment does not accomplish its purpose.
Section III.H ostensibly "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[.]"28 However, the same section would provide Microsoft with yet another exception that effectively swallows the obligation: Microsoft would be permitted explicitly to override a consumer's default choice of middleware product with its own Microsoft Middleware Product where the non-Microsoft product "fails to implement a reasonable technical requirement ... that is necessary for technical reasons to supply the end user with functionality consistent with a Windows Operating System Product[.]"29 Even a superior middleware product could be preempted where it did not conform to Windows Operating System Product code, over which Microsoft would have exclusive control (described above). It is not difficult to imagine Microsoft creating a series of "technical requirement" obstacles that would need to be navigated by competitors and consumers in order to permit them choice in middleware products. Granting Microsoft such control over this important provision would permit Microsoft to avoid its impact entirely, continuing the exact anticompetitive conduct this provision was ostensibly designed to prevent.
Finally, the Proposed Final Judgment would grant Microsoft an overarching exception to all of its disclosure obligations where, in Microsoft's determination, "the disclosure ... 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[.]"30 In an environment where each and every piece of software is increasingly integrated with both the underlying operating system and companion programs, each piece of software could be interpreted by the platform provider as a vehicle for potential interference with vital systems. This is a common argument of monopolists, for it is designed to delay or prevent competition in network services. Given this amount of discretion, Microsoft would be able to effectively prevent competitors from introducing products based upon Microsoft's "determination" that such products would be "dangerous" to the platform or other components.
The litany of ambiguities described in the preceding paragraphs are not a complete listing of the faults of the Proposed Final Judgment; rather, they are indicative of a systematic failure to consider the machinations of Microsoft and the extent to which even the smallest exception will undoubtedly be employed to preclude competition. Ultimately, consumers will suffer the most from these anticompetitive tactics because, with few alternative resources, they will be forced to buy what the incumbent has to offer, without the benefits of innovation and competition. For these reasons, the Court should reject the Proposed Final Judgment or order additional proceedings as described below.
III. The Court Should Either Reject the Decree As It Currently Stands or Order Additional Proceedings to Eliminate Evident Ambiguities
Unfortunately, under the Tunney Act, the Court has no power to modify the terms agreed to by the parties to the Proposed Final Judgment.31 This leaves the Court with the difficult decision of whether to accept or reject the Proposed Final Judgment in its entirety. NetAction and CPSR respectfully recommend that the Court not shy away from rejection of the Proposed Final Judgment where it is apparent that ambiguous provisions, described above, will not ameliorate the extant competitive situation for consumers. Such approval would not be in the public interest as required by statute.32 As an alternative to outright rejection, the Court should consider instituting additional proceedings to assure itself of a "meeting of the minds" among all parties to the agreement.
Microsoft, the Government, and the state parties to the Proposed Final Judgment will undoubtedly argue that the expenditures in time and resources necessary to come to a workable agreement justify approval of the Proposed Final Judgment at this time. On the contrary, however, the time and energy spent upon formulating a solution to a competitive problem that has plagued competitors and consumers for the better part of a decade argues for, and not against, an agreement that is stable, workable and not subject to multiple interpretations. The Court should not rush to approve an agreement that will return to its docket in the near future as a result of ambiguities.
Instead, the Court must be careful to define the terms of the Proposed Final Judgment such that it neither inadvertently accepts an agreement that will not solve the problem nor rejects an agreement that would successfully ameliorate the problem and benefit consumers in the marketplace. Such terms have not yet been defined with respect to the Proposed Final Judgment. Fortunately, under the Tunney Act, the Court has broad powers to order further proceedings to ensure that the public interest is served.33 In aid of its enforcement authority, then, the Court should order additional proceedings in this case to "pin down" the meaning of the various provisions of the Proposed Final Judgment that appear to be subject to dispute.
In order to assure itself that Microsoft, the Government, and the state parties have reached an actual "meeting of the minds," the Court should permit participants to the Tunney Act process to submit written questions to each of the three parties to the Proposed Final Judgment. Such questions should only cover what is or what is not permissible under the provisions of the Proposed Final Judgment. Each of the three parties should answer separately, with no consultation among them, as to whether the action in question is permissible. In order to prevent any "backsliding" in interpretation, each party should be required to submit an affidavit agreeing to be bound by its answers in any additional proceedings. If the three separate answers are in agreement as to the questions posed, the Court should recognize that a "meeting of the minds" has occurred and approve the Proposed Final Judgment forthwith. If the answers are not in agreement, the Court should reject the Proposed Final Judgment until such a "meeting of the minds" is reached and conclusively proven.
The Proposed Final Judgment, as it currently stands, does not offer consumers any hope of relief in the market for non-middleware software applications. Nor does it attempt to offer any relief to consumers using Windows 95, Windows 98, Windows NT, or Windows 2000. To achieve even the uncertain benefits claimed by the Proposed Final Judgment, consumers will have to buy new software and hardware, including new computers.
Moreover, the Proposed Final Judgment is disturbingly reminiscent of Microsoft I in its ambiguities. Indeed, it was an ambiguity in Microsoft I that led to this proceeding and forced consumers to endure five long years of legal wrangling over an issue that the Government thought had been decided between the parties. As it stands, the Proposed Final Judgment does not permit the Court to determine, with any reasonable degree of comfort or certainty, what relief in fact would be provided by the Proposed Final Judgment. Indeed, in a number of specific instances, the exceptions provided for in the Proposed Final Judgment appear to permit Microsoft to escape large portions or even the entirety of its obligations.
On the present record, the Proposed Final Judgment cannot be found to be in the public interest. The Court should either reject the Proposed Final Judgment outright, or order additional proceedings, as described herein, to definitively clarify its terms.
Dated: January 28, 2002
7 This would be true so long as several conditions obtain: (1) those applications depend only on the middleware's APIs, and not on the APIs of the underlying operating system, (2) the middleware runs equally effectively on multiple operating system platforms, and (3) all of every user's applications run on middleware.
11 A massive migration to later Windows operating system products carries additional problems for consumers. Indeed, another of the hidden costs to consumers of both Microsoft's anticompetitive conduct and the Proposed Final Judgment is the cost in network security. The continued dominance of the Windows Operating System and related applications means that Microsoft is a target for hackers and all those who would compromise the privacy and security of network systems. Elinor Mills Abreu, Hack this! Microsoft and its critics dispute software security issues, but users make the final call, InfoWorld (Sept. 27, 1999) http://www.infoworld.com/articles/hn/xml/99/09/27/990927hnnchack.xml. Thus, a breach of privacy or security in a given Microsoft product will be visited upon millions of consumers worldwide. Absent continued anticompetitive conduct, leading to continued dominance in operating system, browser and office applications, such costs would be significantly decreased as the risk of breach would be more dispersed among several operating systems. See Letter from the Electronic Privacy Information Center to United States Senator Patrick Leahy, Senate Judiciary Committee (Dec. 11, 2001), http://msdn.microsoft.com/msdnews/2001/July/hess/print.asp.
12 See, e.g. Mark Hammond, Hidden upgrade woes found in Windows 98 , eWEEK (June 25, 1998) http://zdnet.com.com/2102-11-510244.html, A. Kandra, Consumer Watch: Avoiding the Upgrade From Hell, PC World (August 2001) http://www.pcworld.com/features/article/0,aid,52348,00.asp.
13 Compare http://www.microsoft.com/catalog/display.asp?site=10451&subid=22&pg=3 (Windows Me); http://www.microsoft.com/catalog/display.asp?site=657&subid=22&pg=3 (Windows 2000 Professional); and http://www.microsoft.com/catalog/display.asp?site=11052&subid=22&pg=3 (WindowsXP).
21 "A monopoly in operating system software is a platform for unprecedented control over the flow of information to consumers. Control over this software can be leveraged to near total control over the computer screen. Dominating the screen means controlling ... what [consumers] see and when they see it." The Project to Promote Competition & Innovation in the Digital Age, At the Crossroads of Choice, http://procompetition.org/research/crossroads/crossexec.html.
23 Comments of Robert Litan, Roger Knoll and William Nordhaus on the Revised Proposed Final Judgment (filed Jan. 17, 2002); see also Jonathan Krim, Wording of Microsoft Deal Too Loose, Analyses Say, The Washington Post, E1, E10 (Jan. 18, 2002).
24 The list of markets into which Microsoft is attempting to expand its dominance is growing. Consumer Federation of America and Consumers Union recently published a report, which describes Microsoft's current bundling strategy, which includes integrating email, instant messaging, calendars and contact lists, audio and video media players, digital photography, digital rights management, and identity verification. Dr. Mark N. Cooper, Consumer Federation of America, and Christopher Murray, Consumers Union, Windows XP/.NET: Microsoft's Expanding Monopoly, How It Can Harm Consumers and What the Courts Must Do to Preserve Competition (Sept. 26, 2001) http://www.consumerfed.org/WINXP_anticompetitive_study.pdf.