The dozens of private-sector and governmental position papers concerning the emerging global information infrastructure (GII) all agree that it must be an "open" system. As the Clinton Administration recognized in its 1993 Agenda for Action, the information infrastructure
will be of maximum value to users if it is sufficiently "open" and interactive so that users can develop new services and applications or exchange information among themselves, without waiting for services to be offered by the firms that operate the NII [national information infrastructure].2
The position papers similarly agree that the desired openness could be achieved only through standardized points of interconnection (in technical parlance, interface specifications). In the words of the Agenda for Action:
To assure interoperability and openness of the many components of an efficient, high-capacity NII, standards for voice, video, data, and multi-media services must be developed. Those standards also must be compatible with the large installed base of communications technologies, and flexible and adaptable enough to meet user needs at affordable costs.3
Further, the position papers all agree that governments should not impose the standards; rather, the private sector should develop them. All concur that standards organizations will have a primary role in establishing the GII standards (de jure standards), but some acknowledge that many of the standards inevitably will be set by the market (de facto standards).
At this juncture, the position papers begin to diverge. The divergence arises over the issue of proprietary control of the standards: how much control, if any, should the inventor of the standard be able to exercise over the practice of the standard by others? The amount of control that can be exercised over a standard is inversely related to how "open" that standard really is. During the course of 1994, different firms and trade associations articulated disparate views on the issue of proprietary control of standards, reflecting their narrow commercial interests. During 1995, governments and international bodies began to assert positions as well. Although all proclaim fealty to the goal of openness, a given entity's definition of openness turns on the extent of proprietary control it would allow over GII standards. This paper examines the different definitions of openness that have emerged over the past 2 years.
During 1994, many U.S. companies and trade associations began to express their opinions on "openness" on the information infrastructure. These opinions can be classified into four definitions ranging from the restrictive Microsoft definition to the expansive Sun Microsystems definition.
The Microsoft position starts from the assumption that hardware interfaces specifications are patentable and that software interface specifications can receive both copyright and patent protection. Microsoft further believes that there is no need to create special rules for the GII; rather, the current laws applying to the software industry (which have served Microsoft extremely well) are adequate for the GII environment.
Microsoft will achieve "openness" by voluntarily licensing its application program interface to all third party application developers. Thus, anyone who wishes to develop a product that attaches to a Microsoft operating system will be able to do so. Microsoft, however, has not agreed to license its interface specifications to firms seeking to develop operating systems that compete directly with Microsoft operating systems.4
Microsoft's narrow definition of openness—permitting attachment—flows from its business plan. Microsoft hopes to dominate the market for the operating system of the "set-top box"—the entry point of the information infrastructure into individual homes or businesses. By controlling the standard for the set-top box operating system, Microsoft will be able to exercise control over access to the entire infrastructure. Microsoft wants to encourage third party vendors to develop applications that will run on its operating system; the more applications, the more desirable the operating system becomes and the more likely that the market will adopt it as a de facto standard. At the same time, Microsoft wants to prevent the development of a competing set-top box operating system that is compatible with all the Microsoft-compatible applications.
The Computer Systems Policy Project (CSPP), whose members include computer systems vendors such as Apple and IBM, shares many of the same intellectual property assumptions as Microsoft. Thus, it believes that hardware interfaces specifications are patentable and software interfaces specifications are both patentable and copyrightable. The CSPP differs from Microsoft in that it believes that special rules should apply in the information infrastructure environment. Specifically, the CSPP believes that the owner of an interface that is adopted as an infrastructure standard should be required to license it on reasonable and nondiscriminatory terms, not only to developers of attaching products but also to developers of competing products. That is, the interface specifications should be readily available to all vendors so that they could "build products that are compatible with both sides of the interface."5 Further, the proprietor of an interface standard should be able to revise it only with timely notice or public process.
The CSPP position represents an underlying fear of Microsoft's market power. By requiring the licensing of the interface standards to the developers of competing as well as attaching products, CSPP hopes to prevent a Microsoft set-top box operating system monopoly even if the Microsoft interfaces emerge as the industry standard. Moreover, by permitting revision of standard interfaces only with timely notice, CSPP hopes to prevent the lead time advantages Microsoft's applications developers would otherwise have over third party developers. (These advantages are one subject of the ongoing litigation over the Microsoft consent decree.)
The CSPP approach may work well enough for de jure standards set by a standards body. The standards body may, through negotiations, extract significant concessions from the proprietor. But what if the market, as opposed to a standards body, adopts Microsoft's set-top box operating system as a de facto standard? Pursuant to what authority will Microsoft be forced to license its interface specifications to competitors, or provide timely notice of upcoming revisions? Moreover, who will determine the "reasonableness" of the license fees demanded by Microsoft? Indeed, CSPP itself recognizes the shortcomings of its approach. It has conceded that in GII markets that are not competitive, the government may need to intervene "to ensure that critical interfaces are open."6 Nonetheless, the CSPP continues to insist that the government "refrain from implementing compulsory licensing related to standards."7
The American Committee for Interoperable Systems (ACIS), whose members include Storage Technology, AT&T Global Information Solutions, Amdahl, and Broderbund Software, starts from a somewhat different intellectual property assumption than Microsoft and CSPP. While it agrees that hardware and software interface specifications are patentable, it doubts that many such specifications will meet the rigorous statutory requirements for patentability. Moreover, unlike Microsoft and CSPP, ACIS believes that copyright cannot protect software interface specifications. It reaches this conclusion in reliance on the recent appellate court decisions in Computer Associates v. Altai8 and Lotus v. Borland.9 Further, ACIS believes that any incidental copying that occurs during software reverse engineering is lawful pursuant to Sega v. Accolade.10
ACIS, accordingly, believes that existing intellectual property law, as understood by ACIS, permits sufficient openness in the GII. Because few interface specifications would receive patent protection, and no software interface specifications would receive copyright protection, the firms that develop both de facto and de jure GII standards would rarely be able to exercise proprietary control over them. Additionally, the ability to reverse engineer the interfaces reduces the necessity of mandatory disclosure.
Sun, like ACIS, believes that copyright does not extend to software interface specifications. Nonetheless, Sun does not conclude that the current intellectual property laws provide sufficient openness on the GII. It fears the ability of the de facto standard inventor to obtain patents, to keep certain interfaces hidden, and to change the specifications without notice. Because of the importance of the information infrastructure to the world economy, Sun believes that the government should overcome these obstacles to openness by designating critical infrastructure interface specifications as "barrier-free." Other developers would have complete access to these specifications for nominal consideration.
Microsoft and CSPP argue that the ACIS and Sun positions—which at a minimum deny copyright protection for software interface specifications—would eliminate an important incentive for innovation. According to Microsoft,
Without the incentive offered by the ability to license intellectual property, the information infrastructure would not get built. R&D of the type needed to develop complex products like interactive television requires the investment of hundred of millions of dollars. Companies must be able to recoup those investments by licensing the rights to use the fruits of those investments. In addition, public domain standards give international competitors free ride on technology and intellectual property developed here in the U.S.
Similarly, CSPP states that "[d]evelopers of specifications for interfaces must be able to retain ownership of and benefit from the intellectual property that goes into the specifications, in order to maintain incentives to develop new technologies."12
Sun and ACIS reply to this criticism by drawing a distinction between interface specifications and interface implementations: "Interface specifications are pieces of paper; implementations are actual products or services."13 Removing protection from an interface specification does not lead to removal of protection from a particular implementation of that specification. Indeed, proprietary implementations of nonproprietary specifications provide the bases for rigorous competition between providers: "This combination of nonproprietary interface specifications and proprietary implementations meets the imperative of balancing the requirement of providing incentives to developers of new technology with the societal need for interoperability along the information infrastructure."14
Governments and international organizations began to adopt positions on GII openness in the weeks leading up to the G-7 Ministerial Conference on the Information Society at the end of February 1995. These positions are somewhat vaguer than those articulated by the U.S. private sector.
In January 1995, information technology trade associations from Europe (EUROBIT), the United States (Information Technology Industry Council; ITIC) and Japan (JEIDA) met to create a joint position paper they hoped would influence the G-7 meeting the following month. The first section of the joint paper dealt with interoperability and openness:15
The key to interoperability is the development and implementation of open interfaces. An interface is open if its specifications are readily and non-discriminatorily available to all vendors, service providers, and users, and if such specifications are revised only with timely notice and public process.
The joint paper stresses that those open interfaces should be the product of "private-sector-led voluntary consensus standards development processes."16 It also draws a distinction between interface specifications and implementations: "Interface specifications provide the information and technical parameters for how systems, products and services communicate with each other, but should be limited to that information necessary to achieve interoperability, allowing suppliers to develop different implementations with distinguishing characteristics."17 Nonetheless, the joint paper recognizes a role for proprietary technologies in GII standards: "When a proprietary technology is incorporated into a standard, the developer must voluntarily agree to license the technology on reasonable terms and conditions, demonstratably free of discrimination."18
The joint paper's views in general run parallel to those of CSPP: Intellectual property rights can reside in technology included in a GII standard, but the proprietor must license the technology on reasonable, nondiscriminatory terms and revise specifications only with timely notice. The joint paper unfortunately also shares the infirmity of the CSPP paper in not explaining how the proprietor of a de facto standard will be required to make its interface specifications available to competitors. But the joint paper does not explicitly state that copyright protects software interface specifications. Thus, the references to proprietary technology could be interpreted as referring to patented hardware and software interface specifications. The paper was probably intentionally left ambiguous on this point. While many ITIC and CSPP members historically have supported copyright protection for software interface specifications, many JEIDA members have opposed it. By discussing "proprietary technology" in the abstract, the joint paper could satisfy both constituencies.
In February 1995, the Clinton Administration's Information Infrastructure Task Force issued its Agenda for Cooperation amplifying on themes Vice President Gore had articulated in a speech in March 1994, to the International Telecommunications Union in Buenos Aires. One of the five core principles for the GII recognized by Vice President Gore and the Agenda for Cooperation is providing open access. The Agenda for Cooperation further recognizes that open access can be achieved only through interoperability and standardization:19
An essential technical element of the open access concept is interoperability, i.e., the ability to connect applications, services, and/or network components so that they can be used together to accomplish tasks. As the GII will be based on many different existing and emerging components at local, national, and global levels, it is imperative that these components be interoperable. The key to interoperability is global standards.
In listing recommended government action to achieve the goal of open access through global standards, the Agenda for Cooperation states that the U.S. should join with other countries to "[e]ncourage an open, voluntary standards-setting process that does not denigrate intellectual property rights . . . ."20 Like the EUROBIT-ITI-JEIDA formulation, the U.S. government appears to recognize intellectual property rights in standards adopted by standards bodies without taking a specific position on the copyrightability of software interface specifications. The discussion in the Agenda for Cooperation is so general, however, that it does not contain the protections included in both the joint paper and the CSPP paper: licensing on reasonable and nondiscriminatory terms, and revision with timely notice and public process.
Interestingly, a separate section of the Agenda for Cooperation appears to address this issue, as well as the problem of proprietary rights in de facto standards. When describing the need to create a flexible regulatory framework that fosters competition, the Agenda for Cooperation states that such a regulatory framework should clearly indicate:21
Here, the U.S. government appears to go beyond the joint paper's and CSPP's call for voluntary licensing; the quoted passage reflects an intent to mandate licensing by law. In other words, the Agenda for Cooperation takes the problem of proprietary rights in de facto standards seriously.
The European Commission in September 1994, issued an "initial theme paper" for the G-7 meeting. The theme paper recognizes the importance of interoperability in permitting competition in the development of the infrastructure, which in turn will ensure that users receive the widest range of services at the lowest possible cost. To this end, "G-7 governments are invited to express their support for a consensual standardization process which is open, voluntary, and private sector-led."22 The Commission notes that such a process "would help avoid two pitfalls. On the one hand, the unilateral imposition of mandatory standards by public authorities and, on the other, the emergence of de facto standards from monopolistic market positions."23
Although it is properly concerned about de facto standards and resultant monopolization, the Commission is unrealistic in its belief that standards bodies alone will solve the problem. Standards bodies tend to work slowly, and thus probably will not keep pace with the highly complex, rapidly evolving GII. Accordingly, de facto standards will emerge notwithstanding the best intentions of government and industry.
At the conclusion of the G-7 Ministerial Conference, the G-7 parties issued a statement strongly supporting openness. The current regulatory framework must evolve to "put the user first."24 The framework "must be designed to allow choice, high quality services and affordable prices."25 This consumer welfare will flow from "dynamic competition,"26 which in turn will result from "interconnectivity and interoperability."27
The G-7 parties specifically commit themselves to "[p]ursue the interconnectivity of networks and interoperability of services."28 This goal will "be achieved through the promotion of a consensual standardization process which is market-led and which encourages open interfaces."29 The G-7 partners recognize the need to accelerate the standardization process so that it can develop "timely and market-responsive standards."30
The statement does not explicitly discuss de facto standards. Nonetheless, it reflects a concern with the possible monopolization of the GII. Accordingly, the statement indicates that competition rules need to be interpreted and applied so as to encourage new entrants and promote global competition. Further, competition authorities must "shield[ ] against . . . risks of abuse of market dominance."31
Moreover, the G-7 statement contains no reference to the incorporation of proprietary technology in standards. The silence on this topic appears to be a result of a compromise between the United States and the European Union.32
The United States, the European Union, and the G-7 have all rejected the Microsoft definition of openness; in their view, openness requires that the standard interface specifications be available to all vendors, not only those developing attaching products. The United States and the G-7, at least implicitly, have recognized the shortcomings of the CSPP/EUROBIT-ITI-JEIDA approach; they understand that standards bodies alone cannot eliminate the danger posed by proprietary control of de facto standards. The G-7 solution appears to be vigilant antitrust enforcement. In contrast, the U.S. solution appears to be statutorily imposed compulsory licensing.
It is unlikely that Congress would require compulsory licensing for critical GII interface specifications. Even if it did, such licensing might serve as an effective remedy only if the government remained actively involved to arbitrate disputes concerning the reasonableness of license fees and the timeliness of disclosure. Further, active government participation might not succeed in preventing costly and time-consuming litigation over these issues. Vigilant antitrust enforcement has the same deficiency: It is cumbersome and resolves only one narrow dispute at a time.
For these reasons, the most effective approach may well be that of ACIS. If governments clearly affirmed the recent U.S. case law denying copyright protection for software interface specifications, much of the problem posed by de facto interface standards would vanish. To the extent problems remained on the margins—e.g., the occasional patent over an interface specification—compulsory licensing or antitrust enforcement could address them.
In any event, governments need to focus more attention on this issue immediately. The private sector has forged forward with the GII, already outpacing the standards bodies. As a practical matter, it will be much easier to solve the problem of de facto standards before the standards emerge on the GII. Once they do emerge, they will create vested interests willing to spend significant sums to resist any change to the status quo. Notes
1. Jonathan Band is a partner in the Washington, D.C., office of Morrison & Foerster. Sections of this article appear in Band and Katoh. 1995. Interfaces on Trial. Westview Press, and the forthcoming Japanese language version of this book.
2. Information Infrastructure Task Force. The National Information Infrastructure: Agenda for Action. Information Infrastructure Task Force, Washington, D.C., September 15, p. 9.
3. Id.
4. See Myhrvold, Nathan P. 1994. "Interactive Video Systems," statement before the House Subcommittee on Telecommunications and Finance of the Committee on Energy and Commerce, February 1.
5. Computer Systems Policy Project. 1994. Perspectives on the National Information Infrastructure: Ensuring Interoperability. Computer Systems Policy Project, Washington, D.C., February.
6. Kerkeslager, Ellwood R. 1994. "Electronic Commerce and Interoperability in the National Information Infrastructure," statement before the House Subcommittee on Technology, Environment and Aviation of the Committee on Science, Space and Technology, May 26, p. 8.
7. Computer Systems Policy Project. 1995. Perspectives on the Global Information Infrastructure. Computer Systems Policy Project, Washington, D.C., February, p. 6.
8. 982 F.2d 693 (2nd Cir. 1992).
9. 49 F.3d 807 (1st Cir. 1995), petition for cert. filed (U.S. June 7, 1995).
10. 977 F.2d 1510 (9th Cir. 1992).
11. Myhrvold statement, p. 23.
12. Kerkeslager statement, p. 3.
13. See Rosing, Wayne. 1994. "Interactive Video Systems," statement before the House Subcommittee on Telecommunications and Finance of the Committee on Energy and Commerce, February 1, p. 2.
14. American Committee for Interoperable Systems. 1994. Comment on International Aspects of the NII, August, p. 4 n.2.
15. EUROBIT-ITI-JEIDA Paper (Jan. 1995), p. 5.
16. Id., p. 7.
17. Id., p. 5.
18. Id., p. 7.
19. Information Infrastructure Task Force. 1995. Global Information Infrastructure: Agenda for Cooperation. Information Infrastructure Task Force, Washington, D.C., February, pp. 14-15.
20. Id., p. 16.
21. Id., p. 17.
22. European Commission. 1994. Initial Theme Paper. September.
23. Id. Canada also seems to rely entirely on standards bodies to achieve the goal of interoperability. See The Canadian Information Highway (April 1994), pp. 13-14 and 23-24.
24. Documents Resulting from the G-7 Ministerial Conference on the Information Society, issued Feb. 26, 1995, reprinted in Daily Executive Report (Bureau for National Affairs, Washington, D.C.), Feb. 28, 1995, at M-5.
25. Id.
26. Id.
27. Id.
28. Id.
29. Id.
30. Id.
31. Id.
32. See Hudson, Richard L., and James Pressley. 1995. "G-7 Nations Make Gains in Facilitating
Access to Information Superhighway," Wall Street Journal, February 27, p. A7A.