The following paper was presented at a joint NIST/Kennedy School policy symposium held in Washington, D.C. in June, 1994. The proceedings of that symposium, including this paper, have been reproduced in Standards Policy for Information Infrastructure, Brian Kahin and Janet Abbate, eds. (MIT Press, 1995).
So many industry associations were formed in 1988 that a number of commentators dubbed it "The Year of the Consortium." Few of these organizations were intended to develop true de jure standards. Instead, they were most often formed as a species of marketing partnership in response to intense competition among the largest computer companies, each jockeying to dominate (or at least maintain their position) in volatile market segments such as Unix platform sales. Despite such overtly commercial purposes, these organizations frequently found that the development of de facto standards and specifications provided a crucial method for achieving their ends.
Less noticeably, a far greater number of consortia have been formed over the last ten years that have had a less public but more pervasive impact on the computer, information, and communications industries. These consortia have often been formed specifically to create a single standard of common need throughout the industry. Although proprietary interests may at times influence the operations of such consortia, these interests manifest themselves as often as not as skirmishes regarding small components of the central standard, rather than concerted efforts to drive the standard itself to overtly favor a single company or concentration of companies.
Remaining prominent on the standards landscape are the well-known de jure standard-setting bodies, such as the American National Standards Institute (ANSI) and the Institute of Electrical and Electronic Engineers (IEEE), whose mission it continues to be to set standards through well-evolved, representative processes relating to a larger standards set of goals.
The types of consortia and standard-setting bodies just described, and others discussed at length below, form a loosely coordinated, available, and existing "national standards infrastructure," which operates in an interlinked but independent fashion, and is both a product of and responsive to broad vendor and customer forces. Unlike the state- operated standards systems which exist in many other countries, the standard-setting process in the United States is both organic and dynamic, and is based on an evolutionary process that reaches back to pre-Civil War times.
Building the national information infrastructure (NII) will require the development of many new standards and interfaces, the adaptation of old standards and specifications, and the coordination of many standard-setting organizations. It is not feasible, and, this writer believes, not desirable, to seek to preempt or radically reform the existing method of standard setting in the course of facilitating the development of the NII. While such intervention in the past has quite appropriately led to the evolution of regulatory bodies and regulations in other areas, it is more preferable in the current context for government to seek to understand, support, and to use a computer industry term "optimize" the existing technology standard-setting infrastructure by judicious government participation and stimulation. This conclusion derives from several factors:
For all of these reasons, it is the view of the author that the appropriate activities of government in the standard setting area over the coming years are:
In undertaking these activities, government can effectively further national goals such as internationally promoting the U.S. industrial and services base, representing the rights and interests of those elements of society that lack an effective voice in commercial decisions, and, not least, promoting the interests of government itself as the nation's largest single purchaser of technology and services.
The purpose of this paper is to assist in understanding the types, motives, efficacy, and methods of the various types of consortia that may become involved (either by invitation or market force) in the development of the NII, and, in so doing, to provide insight to government useful in choosing appropriate partners for the task ahead. Similarly, this paper will seek to identify the factors necessary to take into account in forming new consortia necessary to complete the task of creating the NII.
With a better understanding of the dynamics that influence how standards evolve within different types of consortia and the factors essential to forming effective consortia, government will be better able to decide when to lead (perhaps by commissioning development by a given consortia of a standard, or by providing seed funding for the formation of a new consortium), when to follow (perhaps by causing one or more agencies to become members and financial supporters of a given consortium), and when to get out of the way (by recognizing that an existing consortium is an appropriate vehicle for supplying a necessary technology, and refraining from sponsoring a competing standard).
STANDARDS AND "STANDARDS": WHAT'S IN A NAME?
Although the word "standards" is much bandied about in a way that would seem to indicate common agreement on its meaning, in fact there is considerable divergence of opinion on what measure of respect is to be accorded to any given standard. Some would draw distinctions based on the source of the standard reserving the highest level of respect to "de jure" standards promulgated by organizations such as ANSI, and relegating standards maintained by industry associations to the lesser status of "de facto" standards. While some purists would not call consortia-developed specifications true "standards" at all, the impact of such de facto standards has often been as strong or stronger than de jure standards, and the role of market power-based de facto standards (such as DOS and Windows) can be most powerful of all.
Further complicating the issue is the fact that one kind of standard will often be dependent on, or incorporate, another. For example, de facto standards often incorporate de jure standards (e.g., the X Window System incorporates certain IEEE standards), and in some cases de jure standards bodies accept de facto standards or link to them, such as Object Management Group's widely recognized Common Object Request Broker (CORBA) specification. Finally, organizations like X/Open seek to develop environments that use a mix of de jure and widely supported de facto standards.
As a result, the role of de facto standard-setting consortia is relevant and important to true de jure standard- setting organizations such as ANSI and IEEE. On the one hand, as a practical matter, a de jure organization may opt not to develop a competing standard if it is satisfied that the de facto organization is providing a useful standard. As a result, a de facto standard-setting organization is effectively permitted to control an important area of technology. On the other hand, the endorsement or incorporation by de facto bodies of de jure standards augments the effectiveness of the latter standards.
For these reasons, a high level of sensitivity should accompany increased government involvement in what has become a complex and largely successful interlocking web of organizations, standards, and environments. At the same time, government needs to thoroughly understand the degree of proprietary influence that may exist within a given consortium and that may affect its standard-setting activities.
DEVELOPING A TAXONOMY OF CONSORTIA
Consortia are to a degree merely new wine in old bottles. Trade associations have been formed for many years to promote the interests of their members in the building trades, for example this promotion has often involved influencing the development and promotion of building standards for local adoption, thereby benefiting the membership. However, these associations also have frequently engaged in lobbying and other activities primarily intended to protect the membership's business and market their collective wares. In the era of high technology, this traditional role has not been entirely abandoned. For example, the X Industry Association (XIA) was formed to promote the X Window System technology, simultaneously with and independently from the spinning out from MIT of a new organization, the X Consortium, which was created to continue the development of the standard itself outside of MIT. Today, XIA (and its marketing budget) is being merged into the X Consortium, as concerns mount over the avowed intention of Microsoft to supplant the Unix operating system with Windows NT.
Of greater interest than traditional trade associations are the following new types of consortia, many of which have had the development of standards or the development of new technology as their primary focus, albeit for varying degrees of strategic purpose:
A number of high-profile consortia (such as SemaTech and MCC) have been formed largely as a result of heightened national pro-competitive concerns. Some have succeeded, some have failed to coalesce, and some have wandered from their original purposes as they have sought to sustain themselves. In some cases the concerns (e.g., national paranoia over loss of dominance in DRAM chip production) that helped launch the enterprises have already abated to a degree, and smaller groupings of companies (sometimes including Japanese or European partners) have since entered into joint ventures to produce specific products without the high-profile, highly political trappings of their predecessors.
Although there may prove to be instances where government should play a role in forming additional high- budget, special-purpose consortia that would develop actual technology in addition, or incident to, creating standards, it is worth noting that by far the greatest number of standards have come from more modest cooperative sources. As the focus of the current discussion is standards and not technology development per se, research consortia are not analyzed in detail in this paper.
Existing groups such as VXIbus Consortium (VXI) and the MIDI Manufacturers Association (MMA), and numerous new consortia currently being formed such as the Distributed-Computer Telephony Group and GO-MVIP, the Global Organization for Multi-Vendor Integration Protocol are primarily concerned with assuring the development and maintenance of a usable, robust standard for the benefit of the industry generally. Essentially apolitical, they direct the greatest part of their efforts to evolving and supporting a standard, and have low budgets in consequence. Where they work best, they successfully avoid proprietary influences and pressures and implement the best technological methods to produce sensible, robust, practically implemented standards. While many such groups are formed and funded by vendors, some (such as the CAD Framework Initiative) are formed through the efforts of end-users, in order to lower acquisition costs of tools or other products that would otherwise too often be based on proprietary, non-interoperable technologies. Such groups are often formed to develop a standard to fill an important niche-industry technical gap that is not large enough to merit the attention of a recognized standard-setting body like IEEE for example, MMA was formed to provide an essential interface standard so that the electronic music industry could develop in an accelerated and orderly fashion.
Among all types of consortia, these may be the most appropriate for the government to work with, or create, in the course of coordinating and setting NII standards. Due to the comparative absence of proprietary pressures, the work product of these consortia tends to be the most responsive to broad practical and economic needs, and most often represents a broad consensus of the industry (both direct participants and non-participants, such as end-users, alike). For similar reasons, the standards implemented by this type of consortium may often be better technically, since there is no reason to artificially select or support a given technology or architecture.
It is worth noting that these consortia tend to have not only the smallest budgets (as little as $15,000 a year), but one of the highest success rates in creating standards that are widely followed by the industry. Accordingly, they represent unusually cost-effective vehicles for standards development. However, budget sizes may constrain the activities of these organizations, and the infusion of very modest amounts of government funding in this area could have dramatic effects in both speeding and guiding the pace of standards development. By way of example, all operations of such consortia are typically volunteer in nature. Employer time demands, communication inefficiencies, and increasingly tight travel budgets all slow progress. Small amounts of funding could subsidize meeting expenses, document production and interaction with other standards-setting bodies, thus accelerating progress.
For purposes of this discussion, "strategic consortia" are deemed to be consortia initially formed and funded by a limited number of companies for their individual benefit in order to promote the adoption of certain technology (such as a chip architecture) as "open" technology. The primary mechanism of many such consortia for seeking to achieve this end has been the development and promotion of some type of standard, the efficacy (as compared to the marketing) of which has varied widely in practice.
Perhaps the greatest and most interesting flowering of consortia of this type has occurred in the Unix marketplace, relating not only to the operating system itself (Unix International and The Open Software Foundation, or OSF), but with respect to chip architectures (e.g., in the RISC chip market, 88open, the Advanced Computing Environment (ACE), SPARC International, PowerOpen, etc.) and software environments (The X Consortium) as well. These consortia commonly were formed with one or more of the following specific characteristics and objectives:
In some cases, consortia have had roots of one type and have grown branches of another: the X Consortium, as noted, was formally born in 1988 out of concern over a Sun Microsystems software product initiative. The technology itself had been under development within MIT since 1984, and the industry began adopting it as a standard in 1987. The organization and the standard then enjoyed over five years of comparatively ecumenical existence dedicated to the continued evolution of a usable standard. Today, this consortium has left its home at MIT to become an independent entity, and the imminent release of Windows NT gives new significance to this organization as a vehicle for collectively asserting and promoting the continued vitality of Unix and the X Window System standard. Recently, it has decided to merge with XIA, as mentioned above, and has entered into discussions with OSF to take over further development of two other crucial components of the Unix desktop: Motif, a graphical user interface, and CDE.
Similarly, Object Management Group (OMG) was initially founded to develop a family of standards to facilitate the adoption of object oriented programming methods and products. Today, the organization has broadened into a wide range of activities dedicated to nurturing and promoting the evolving OOP industry, including development of six major annual trade shows in five countries on four continents, development of training sessions and publishing ventures, and evolution of innovative joint marketing and electronic distribution projects. At the same time, OMG has been successful in becoming recognized as the preferred source of many types of specifications for the OOP industry, and therefore the focal point for industry efforts in this area.
A hallmark of the success of several consortia has been their good fortune (or good foresight) in being "ahead of the curve"; in OMG's case, by seeking to establish a standard before any significant company had obtained a vested interest in promoting its own proprietary technology for adoption as a standard. As a result, the industry could follow OMG's lead, and cooperate to achieve the best technical result for mutual, rather than individual, benefit.
The reverse, unfortunately, has also been true. 88open, the first RISC chip consortium, was brilliantly successful on the technical level, achieving the highest degree of true interoperability of any of the competing organizations. Nevertheless, its technical success was insufficient to overcome certain other handicaps, such as the fact that the chip technology upon which it was based was introduced too late into the marketplace to achieve the momentum of adoption from which it otherwise might have benefited. The new PowerOpen Consortium (styled in many ways on the earlier 88open Consortium) is virtually assured of avoiding this fate, since both IBM and Apple (among others) have already announced products based on the PowerPC series RISC chips being produced by Motorola, and the sales momentum that those companies represent should be more than sufficient to ensure speedy and efficient porting of significant amounts of software in short order.
However, the success of a consortium tends to correlate strongly and inversely to the degree of proprietary advantage that its founders sought to gain. This results from a variety of factors: not only are there more losers than winners in industries dominated by a small number of large players, but the word "standard," stretched to the limit of anyone's most liberal commercial definition (e.g, VCR formats or PC operating systems), cannot usually accommodate more than one or two major instantiations on a long term basis. Again, many strategic consortia were formed as much for publicity as for actual technological purposes, or were announced before a clear plan of action had been evolved, or fell apart as the shifting strategic sands upon which they were based continued to move (e.g., the ACE Consortium, which suffered from other problems as well). Finally, strategic consortia are more likely to be formed by companies that perceive that they are already at a disadvantage, and the odds are therefore often against success from the beginning.
Conversely, the success of consortia correlates highly and directly to the absence of perceived proprietary advantage to any individual company (e.g., VXI, discussed above, or OMG, which was formed long before any significant number of object oriented products were available, or even in the design phase) or, in some cases, with a very high degree of agreement among a critical mass of companies as to the existence of a common enemy (e.g., the wholesale and rapid adoption of the X Window System Standard in response to Sun).
Working with strategic consortia must necessarily involve more careful investigation on the part of government than supporting purely specification-oriented organizations. Important in this regard are the following:
If government wishes to play a comparatively passive, rather than a heavy-handed role in working with consortia, it may therefore wish to evaluate at the outset, and weigh in the balance of its decisions, whether a given strategic consortium is wide-based in its support or has competitors in the marketplace, whether it is promoting or supports a technology which is succeeding or failing, and whether its management's and its members' goals are conducive to the national interest. It may be that adversely weighting proprietary purpose may be more desirable than imposing heavy layers of regulation on the activities of consortia that accept government funding. How such a comparatively ad hoc judgment would be formulated is admittedly not without its own problems, and may best be left to the judgment of the governing body of a "Consortium of Consortia," as discussed below.
WEIGHING CONSORTIA AS GOVERNMENT PARTNERS
In weighing whether to support the activities of a given consortium, or to facilitate the formation of new consortia, government should also be mindful of the factors that lead a consortium to flourish or fail. Similarly, government will not wish to become involved with groups that are not operating with due regard for applicable antitrust law. For example, any successful consortium must carefully address a number of structural issues in order to succeed and avoid adverse consequences. The approach to these issues will vary for a given organization depending on the goals, competitive position, and other unique features of the particular group of companies involved.
Frequently, the first reality that the organizers of a consortium must confront is economic. Where goals are ambitious, such as mounting certification as well as standards development programs, funding needs can be extreme and can only be satisfied either by enrolling very large memberships (e.g., The X Consortium, or OMG) or by requiring very large contributions from individual members (e.g., OSF and PowerOpen). Typically, large-budget consortia have been strategic consortia, with proprietary vendors (i.e., computer or chip manufacturers) contributing the lion's share of funding.
Substantial aid (and, where perceived to be necessary, substantial diminution of the influence of large companies, which usually receive a disproportionate number of board seats in exchange for funding) could be provided by government assistance of modest size. Using 200 as a rough estimate of the number of all computer and related industry consortia which are currently active and $200,000 as an estimate of average annual aggregate cash contributions ($50,000 may be closer to a median number) yields a total funding requirement for the consortium portion of the national standards infrastructure of only $40,000,000. Accordingly, the magnitude and cost-effectiveness of this in-place resource, and the degree of influence that comparatively modest amounts of funding might produce, should not be ignored.
A second economic concern is to ensure that those companies become members whose participation is essential in order to achieve success. Accordingly, a consortium's dues structure commonly reflects this concern, with some classes of members often bearing a disproportionate contribution obligation. The motivation of certain types of members to take up this burden is to increase the likelihood of those members later reaping a disproportionately high economic benefit in the marketplace. For example, while sponsor-level membership in PowerOpen (a significant goal of which is to foster rapid porting of software to the PowerPC environment) requires $250,000 in annual fees and initiation dues of $750,000, ISVs may participate at a limited level for only $100 a year.
Conversely (as discussed in greater detail below under "Governance"), all types of consortia are likely to maximize revenue income by offering various levels of membership, with higher dues being required of those levels of membership that receive early access to technology and other benefits. Many consortia also charge different rates for companies with different revenue levels, in order to permit smaller companies to participate (or in order to charge more to their larger members, depending on your point of view).
Although many consortia might not be formed at all if there were no strategic advantage to be gained by those who take the initiative to form them, it is also true that only those interest groups will be invited to join at what amount to subsidized rates whose participation the founding and funding members deem to be essential to achieving the sponsor members' own ends. Accordingly, government funding might be conditioned on giving membership to other interest groups, or used to fund such membership, and therefore input, in the standards process.
Besides problematic membership profiles, large-budget strategic consortia suffer from a number of other diseases peculiar to the species, one of which may be described as the "cliff syndrome." Simply put, the larger the budget, the greater the likelihood that such an organization will find itself unable or unwilling to downsize when its grand objective has succeeded or failed. In such a situation, such an organization tends to hurtle at full budget speed off a financial cliff of its own making and into oblivion. The origin of the problem is that while the grand objective remains the focus of the dues- paying members, little attention is given to whether their collective investment in the consortium merits the development of any long term, less dramatic goals as well.
Often, in fact, it does not, and the consortium should be promptly and efficiently wound down when the major objective has been achieved or failed. However, in other cases, such organizations could have useful long-term lives as user groups, providers of certification services and the like, some of which may be of greater importance to the public. Unfortunately, when the pursuit of the major objective has passed into history, or if the marketplace or technology shifts have made the objective irrelevant, it is often too late to downsize the organization to a sustainable level, since the original strategic proponents are no longer interested in providing the funding necessary to transition the organization to fulfill more humble objectives. Organizations that successfully make the transition are those which have used their high-profile years to quietly backfill their budgets with income and services by establishing more humble activities, such as publishing and joint marketing programs.
The most stable and longest-lived consortia therefore tend to be those that require "modest" annual fees (modest being defined for such purposes as being on the order of $5,000 for a company with under $10 million in sales, and up to $50,000 for companies with sales of over $500 million). In this case, members tend not to send upper-level management to meetings or to examine the benefits received as aggressively or frequently. In consequence, management of the consortium can usually take a more aggressive part in the steering of the organization. Human nature being what it is, the direction of that guidance will usually be toward building a long-term, expanding organization.
Where this process works best, innovative executive directors roll out new programs on a regular basis to keep members happy and renewing. Where the process works poorly, the organization rightfully withers and dies.
In the middle case, where due to a failure of forethought or other reasons a transition has not been achieved, government funding might either assist in a down-sizing transition (where, for example, a standard remains important and requires continuing maintenance) or aid in the transfer of these duties and the underlying technology and staff to another consortium.
As with the evaluation of strategic pressures, a careful review of funding structures will therefore also assist government in assessing the appropriateness of a given consortium as a partner in standards development. Those consortia that have modest budgets and broad-based financial support are likely to be better long-term partners than those whose funding is highly dependent on the goals and economic viability of a small number of large hardware companies.
There are numerous models for developing technology, each with its own advantages. The spectrum of methods runs from the very expensive method of total in-house development (88open designed and built almost all of its own test suites), through use of contract parties (such as UniSoft or ApTest, an 88open spin-off company) and contributed employees (used by CAD Framework Initiative), to employment of an RFT (Request for Technology) process (the OSF and OMG model). Other consortia, such as the X Consortium, use a mix of techniques, including reliance on member labor externally, internal staff, and contributions of technology.
Comparing the principal advantages and disadvantages of these methods, the following features stand out: In-house development can maximize control and planning and minimize time to completion, but requires the highest level of funding. Contributed employees are sometimes hard to motivate, may be preoccupied with what is going on at the home office (if they are on site at the consortium), and are often not given ample time or credit by their employers to properly complete their assigned consortium tasks. Issuing an RFT can reap a rich offering of responses, but working through to a final acceptance can be a laborious and sometimes tendentious process to operate and administer. It also requires the most scrupulous legal scrutiny to avoid anti-trust problems.
Again, where a particularly important piece of the standards mosaic may be involved, judicious economic support from government sources may facilitate production by permitting greater flexibility in choosing the mode of development.
Effective, efficient and representative evolution of standards by consortia is impossible without an appropriate structure of administration and technical decision making. When the author's law firm first began representing consortia, it performed a wide examination of possible forms under various jurisdictions, and settled eventually on the Delaware not-for-profit, non-stock membership corporation. This structure permits members to join and leave with ease, provides members (and their counsel) with a convenient and familiar body of corporate law to rely upon, and provides a unique degree of flexibility of operation. Most of the largest East Coast-based consortia are now formed on this model, while many consortia formed on the West Coast use a similar format under California law. This structure has stood up extremely well in practice.
The heart and soul of any consortium may be found in a humble home: its by-laws and charter. Although a few important rules may come to rest in a membership application, most of the regulations and rights of the organization will be found in these legal documents. Whether or not they are carefully conceived will determine whether the organization is easily managed, whether it incurs needless exposure to its members under the antitrust laws, whether its members feel themselves fairly represented and therefore renew their membership, and whether or not the organization is sufficiently flexible to evolve and flourish.
A crucial factor in reaching a stable and sustaining organization is determining the appropriate method of representing members on the Board of Directors. Organizations that we have represented have developed a variety of radically different formulae (sometimes several divergent methods have been adopted by the same organization at different points in its evolution) that have worked. These have included avowedly economic models (those who pay the highest dues get the Board seats), arbitrary solutions (the first members to join receive the seats, while later members may stand in line for a turn-over opportunity), as well as models which place a premium on democratic values (seats are allocated to types of members - ISVs, hardware manufacturers and academics, in order to ensure that all interest groups are heard from) or neutral objectives (e.g., only non-members can be directors, to ensure that standards adopted are "pure").
Most (but not all) consortia have different categories of membership, with differing classes having different rights. These rights may include voting for (or nominating) directors, the right to participate in technical or other committees, early access to technology or standards, and reduced (or free) access to standards, certification or other services. In some cases, access to such rights is deliberately conditioned on paying higher dues, in order to boost funding levels. In all cases, setting the benefits appropriately in relation to the dues is essential in order for the enterprise to succeed, and some trial and error is sometimes engaged in before the appropriate formula is found by a given organization. Similarly, periodic readjustments are often necessary to keep members satisfied; many consortia have gone through periods of declining membership before realizing that recalibration of dues and services was necessary.
Typically, the actual activities of consortia are implemented by technical committees, work groups, and special interest groups, as well as by non-technical governing bodies, such as business, audit and executive committees. In some consortia, most technical activities take place primarily or exclusively by electronic mail interchange rather than at face to face meetings.
Based on the experience of this author, it would be quite workable for most consortia to either accommodate government agencies as members within existing member classes, or to create or amend a class to permit government participation.
Although this author would not advocate that extensive regulations or a stringent "litmus test" be imposed on consortia before government involvement would become possible, government should nevertheless be entitled to exercise the same degree of review and influence that any other member would be granted to ensure that the vehicle that it was funding met its objectives.
Should government become involved in the development of standards through consortia, it will come into contact with issues of ownership, protection, and control that superficially appear to be philosophically counter to certain traditional government objectives. For example, although government will presumably have no interest in assisting in the creation of any standards that are not manifestly available to all, it will still share the concerns of profit-making consortium participants in making sure that such standards are effective. Should vendors falsely claim that a product is compliant with a standard, then the standard will lose credibility in the marketplace, and the standard-setting effort will be in vain. Similarly, if there are competing specifications in the marketplace, each claiming to be a standard and each using the same name, confusion will follow.
To truly control standards, a rigorous program to select, register, and maintain trademarks, service marks, and/or membership marks is therefore often advisable. Where certification is intended, registration of certification marks is even more important.
Consortia may or may not plan to develop technology when they are formed, but they often nevertheless find themselves creating intellectual property as they evolve. Since members may come and go, and the organization itself may at some point dissolve, careful consideration must be given to the current and eventual ownership of technology. Where government funding is involved, the analysis will require another dimension, and any participation by government should address this concern at the outset. Whether or not traditional rules of government funding and ownership will be appropriate will depend on the situation, and therefore the mode of actual participation by government (e.g., membership, grant support or actual contract development) will be highly relevant, due to the requirements of existing law and regulations.
A related issue arises where consortia make use of contributed employees (a frequent practice). Usually, those employees will be subject to agreements with their employers that provide that all developments created by them will belong to the employer. Similarly, many consortia permit the contribution of technology that becomes part of a specification, a standard, or an implementation of a standard, which is then made freely available to the world. In situations where technology is developed over time as a result of the contributions of many individuals and corporate members, unscrambling the ownership of that technology and the ultimate liability for any error or infringement can become nightmarish (see the "Sources of Technology" chart above).
To give one example of how complicated such matters can become, the X Window System Standard was initially developed within the Massachusetts Institute of Technology by university staff, and then supported by industry financial contributions. Member (and other) technical contributions were then often accepted, modified, and incorporated into the standard (although actual ownership of the technology contributed was not necessarily transferred), at the same time that further technology was being developed within the X Consortium (still by MIT staff). Since certain software representing the standard is licensed by the Consortium to the world at large at a minimal fee, actual code is thus released to commercial users, who then incorporate it into their products. The software distributed by the Consortium also includes third party commercial software (such as fonts), which is similarly distributed and reused. To complete this complex picture, MIT has now transferred its ownership rights to the new, independent X Consortium.
Where the avowed goal of a consortium (and the government) is to develop and support an open, non-proprietary standard, any actual ambiguity about ownership and control are much to be avoided, even if developmental convolutions like those described above may prove to be unavoidable. As a result, every consortium should institute a careful program of technology procurement, documentation, protection and labeling, as well as the inclusion of conclusive warranty disclaimers.
Similarly, a consortium must decide whether or not to throw its standards into the public domain, or copyright them and make them freely available. Usually, the latter is the better course, since it permits the consortium to maintain better control of the standard. The interests of government and industry are congruent on this point.
Although taxation of consortia will not be a matter of primary concern to government in evaluating consortia as partners, government will need to be sensitive to the reality of tax planning by those who form consortia, and the way in which tax concerns affect how consortia structure themselves and operate.
Many consortia elect to operate as tax-exempt trade associations under Section 501(c)(6) of the Internal Revenue Code; a few register under Section 501(c)(3) as public charities. As tax exempt organizations, most of the activities of these consortia may be carried on without payment of income taxes on dues or related income. However, a number of activities (such as the inclusion of paid advertising in consortium publications) are deemed by the IRS to stand outside the tax exempt purposes of the organization, and to constitute "unrelated business taxable income." The income from these activities is taxable, and if too great a level of this type of income develops it may endanger the tax-exempt status of the entire consortium. Accordingly, if it is important for the consortium to engage in some taxable activities, it often decides to carry on those activities through a separate corporation. However, many common consortium fee-generating activities that are of greatest relevance here, such as certification testing and standards publishing, will usually be deemed to be related to the purpose of the consortium and thus fall within tax exempt parameters.
Although tax avoidance (as compared to tax evasion) is a legal and hallowed part of the American psyche, it is unfortunate that those forming a consortium often fail to remember that tax avoidance was not their primary purpose for forming the organization to begin with. Where a consortium has an admittedly strong proprietary purpose, that purpose or affiliation will usually make obtaining tax exempt status impossible without crippling the purposes of the organization. Similarly, where an organization intends to embark upon many taxable activities anyway, the amount of tax structuring that may be necessary to avoid taxation of other activities can rise to a level of artificiality and inconvenience that ultimately outweighs the benefit of carrying on its other activities on a tax-exempt basis.
In balance, where an organization is a true standards- setting organization, seeking tax exempt status will usually be well worthwhile. Securing tax exempt status facilitates budgeting and financial management, and permits management to ignore the possibility of taxable year-end profits. Where the purpose is proprietary or the plans complex, it sometimes proves to be better to forgo filing, or to later abandon tax exempt status, if this can be accomplished without adverse consequences. Where tax exempt status is not sought or is surrendered, taxes may be minimized or eliminated by managing the budget as closely as possible to a break-even point annually. Where feasible, multi-year member commitments are often sought to facilitate long-term planning relating to issues such as hiring of management and entering into leases (OSF's multimillion dollar, multi-year commitments led the pack in this regard for some time). Obviously, long-term issues are most problematic for the "cliff syndrome" consortia mentioned above, and least troublesome for stable, large member base consortia with smaller annual dues, such as OMG, which is now the world's largest software consortium with over 400 members.
As with tax issues, antitrust concerns (quite properly) affect the way consortia structure themselves and operate. By definition, consortia are combinations of competitors. As such, they must operate with careful regard to the federal and state laws that they are subject to, as well as to the threat of treble damages suits from private parties.
At times, consortia seek to carry out new practices for which there are no (or only poorly analogous) precedents, and careful analysis must then be undertaken to ensure that risks are not inadvertently taken. In such cases, consortia encounter legal expense and uncertainty that may impede their actions. Happily, standard setting is presumptively deemed to be an appropriate activity for joint activity. Engaging in certification services, while less explicitly favored, may be carried on so long as proper controls are established to ensure non-discriminatory pricing and access to testing.
As a generality, the legal community believes that non-members may be charged higher fees for certification testing and other services and products than members, so long as there is a rational relation between these fees and dues already paid by members and expended on the establishment of the services or products in question. The reason that these practices are engaged in is that some incentives must be given to the companies that fund the activities of the consortium. The alternative is "free ridership," where individual companies see no advantage to funding what they hope others will pay for. Government support for the standards development activities of consortia would help remove the advantage that well-to-do companies have over those whose budgets do not permit participation in multiple consortia.
Government assistance in the antitrust area to date has been most marked through passage of the recently augmented and renamed National Cooperative Research and Production Act of 1993 (the NCRPA; originally, the National Cooperative Research Act of 1984). Under the NCRPA, certain types of activities (such as, arguably, the development of standards) are protected from the risk of treble damages and liability for a plaintiff's attorneys' fees.
However, the NCRPA is not without problems and limitations. Although the amendments enacted in June of 1993 expand the activities that the NCRPA covers, there is still doubt as to the exact coverage involved. Also, public disclosure of members is required, and this is not always acceptable for a strategic consortium where members may not yet have announced their commitment to a given technology or architecture; in such a case, the ability of their competitors to discover their strategic direction as a result of their membership in the consortium may be seen to outweigh the benefits of registration under the NCRPA. Finally, as with many pieces of legislation, there are also a number of unartful, or politically influenced, aspects to the NCRPA which make it difficult to tell whether liability may be eliminated in a given situation or for a given member or activity. For example, the ability of some non-US members of a consortium to take advantage of certain aspects of the NCRPA's protection will probably be impossible to reliably determine. Although it is doubtless too soon to look for further amendment of the NCRPA, there remain areas where amendment, or at least clarification, would be of material benefit to the effective advancement of standards development through consortia.
Finally, the industry is waiting to see to what degree the predominantly laissez-faire antitrust policies of the two preceding administrations will abate, and to what extent consortia will be under increased scrutiny. It was rumored some three years ago that the federal antitrust agencies were undertaking a broad investigation of high-technology consortia at the time that OSF came under examination, but no public announcement or other enforcement action followed. As a result, although industry hopes that the pro-competitive activities of consortia will continue to receive at least some measure of government sympathy, if not outright legislative encouragement, the potential for suit by private litigants (such as competitors) remains a concern, especially where NCRPA protection does not extend, or where registration was not deemed to be feasible. In that regard, it is worthy of note that OSF was sued by private parties concurrently with the commencement of the government investigation, and it is popularly believed that the government inquiry was inspired by private party complaints.
Participation by government in consortium standard setting activities can (and should) occur at more than one level, such as: coordination of consortia standard-setting activities, funding of standards development and conformance testing, direct membership in consortia as a consumer of goods and services, and active promotion and participation in the founding of new consortia, where necessary.
One method whereby government could more dramatically facilitate the evolution and coordination of standards to support the development of the NII might be to promote and help fund the formation of a national clearing house, or "Consortium of Consortia." This would bring together consortia with an interest in the technological areas that comprise necessary components of the information superhighway. Such a body could:
As a first step, such an organization could catalogue the areas of expertise and interest of existing consortia with an eye toward assessing their utility in developing standards for the NII. More intelligent government decisions could then be made based on this foundation.
Although the appropriateness of the more proprietary of the strategic consortia for government partnership may vary, the utility of formal standards bodies, specification consortia and the "less strategic" consortia such as the X Consortium for establishment of useful standards can scarcely be seriously questioned (otherwise, why would so many of their standards and specifications be recognized by the de jure standard-setting organizations?). To ignore this ready, available, and (somewhat) coordinated standard-setting infrastructure of responsive organizations would be to neglect a resource of importance. Those responsible for the formation and implementation of NII policy should work to thoroughly involve existing consortia in the standard-setting process, and should promote the formation of new consortia, where appropriate, to develop and maintain those standards components of the NII for which no existing organization is an appropriate choice.
For two different treatments by this author of some of the same material contained in the above article, see:
"Forming, Funding, and Operating Standard setting Consortia," IEEE MICRO, Dec. 1993, pp. 52-61, which is part of a special issue dedicated to the analysis of standards setting issues in high technology industries. This article does not address government issues, and focuses more heavily on organizational issues and concerns.
"Forming and Representing High-Technology Consortia: Legal and Strategic Issues," Computer Lawyer, March 1994, pp. 8-17, which focuses more heavily on intellectual property, organizational, antitrust and other legal issues important to the formation and operation of consortia.
The strategic consortia discussed in this article are of recent enough vintage that there has been (as yet) little academic or other thorough analysis of them. Consequently, most sources are to be found in the trade press and in information disseminated by the consortia themselves. However, the following are useful sources for data:
The World of Standards. 88open Consortium Ltd. (San Jose, 1991, 2nd ed.: 1991). 243 pages, plus introduction. This is an excellent source book, containing detailed descriptions of over 100 important recognized standards and the organizations which sponsor them, as well as various products which are significant in the industry, such as WINDOWS, Motif, PostScript, SVR4, etc.
Softwars: The Legal Battles for Control of the Global Software Industry. Anthony Lawrence Clapes (Quorum Books, Westport CT: 1993; pp. 261 - 274). This interesting and opinionated book has an excellent chapter on the Open Systems debate and the way that it has played out in the consortia trenches.
The Evolution of Open Systems. 88open Consortium Ltd. (San Jose, CA: 1992; 16 pages). Good, concise review of the history and evolution of open systems, with some emphasis on the current role of binary compatibility standards.