Roy Rada, John Berg "Standards: Free or Sold?" Communications of the ACM, 38, 2 pp 23-27, February 1995.

Recently Carl Malamud suggested in his book, "Exploring the Internet" (Interop Publishers, California), that standards should be free! The fees from the sale of standards and the stringent policies on sharing copies of standards may decrease the number of people who access these documents and may thus reduce the impact of the standards.

Consider two scenarios. In one scenario, information technology (IT) standards continue to be available only through "official" sources and the prices of standards documents continue at their present levels. In the other scenario, if information technology standards were available through the Internet on public domain file servers, many users and vendors could find and read the standards, and the efforts of the standards developers may produce greater and quicker results. While the latter possibility is speculative, the former clearly describes the current world. Yet, many lawmakers realize that information technology has become a public good, and IT's importance to society, commerce, and international relationships will only increase.

WILL THIS CURTAIL THE ROLE OF ISO AND ANSI?

What is the role of ISO and ANSI? Formal SDOs provide a venue for establishing consensus support for standards. For example, in the US, ANSI provides protection against accusations of vendors acting in unison against competition in violation of Anti-Trust laws. The value of ISO and ANSI to their respective constituents rests on constituent confidence that the standard approved by the SDO resulted from fair, above-board considerations assuring long-lived consensus support. The actual standard specification development is correctly turned over to volunteer technical experts.

ISO and ANSI also serve coordinating and distributing functions which the paying members of the organization manage--somewhat like the United Nations. ISO and ANSI have "civil servants" whose salaries result from member dues. The annual budget provides the major members with control over member-regulated SDOs. The member organizations may choose to reduce member dues by using revenue from the sale of standards documents for supporting the SDO's member-approved program. Examination of the SDO's budget should indicate the sources of income and the support given to various programs.

The ANSI Annual report for 1993 says $10,808,000 came from selling publications of an income of $16,731,000. On the expenditure side, publications are NOT included. The summary states: "The net income from publication sales continues to provide a significant subsidy of the expenses associated with the Institute's core infrastructure which is not fully funded by membership dues."

Are standards documents a reasonable source of income for SDOs? Standards promoting organizations, such as ISO and ANSI, facilitate the development of standards but the bulk of the work is done by volunteers and no royalties are paid to those who have developed the standards. ANSI might follow their example and provide standards to the ultimate user for the least possible cost. Since the advent of the internet, the costs associated with actually printing the standards could disappear. However, costs associated with providing a "standard documents server," its availability, and its maintenance will replace the paper-related costs. Should these functions be covered by the dues of the members? Could national standards organizations like ANSI choose to make standards electronically available? In particular, the right to print in the US is granted to ANSI by ISO for ANSI's participation. Similar rights are given to other national bodies approved by ISO for participation.

WHAT ROLE DO CONSORTIA PLAY?

Consortia provide an interesting example of alternative sources of standards-related specifications and serious competition to formal SDOs. Consortia may form in response to intense competition among the largest companies, each trying to dominate in volatile market segments such as Unix platform sales. Despite their overtly commercial purposes, these organizations frequently find the development of de facto standards crucial to achieving their ends.

In some cases, de jure standards bodies accept de facto standards or link them to their own standards (from Andrew Updegrove "Forming, Funding, and Operating Standard-Setting Consortia" in IEEE Micro, p 52-61). Should a SDO take the product of a standards consortium and sell that product for a price?

To fund their development activities, consortia may charge fees to members. Funding levels of some consortia is on the order of millions of dollars. These funding needs can be either met by enrolling many members or by requiring very large contributions from individual members. For example, sponsor-level membership in PowerOpen (whose main goal is to foster rapid porting of software to the PowerPC environment) requires $250,000 in annual dues. Many consortia charge different rates for companies with different revenue levels, to permit smaller companies to participate. Again there is no relation between these costs to produce a consortia specification which is offered to a SDO for de jure standardization and the cost of the eventual de jure standards -- the consortium wants to obtain support for the standard, not to make money from the sales of the standards document. Many proprietary and vendor-consortia developed specifications are either on the Internet or are sold at lower prices than their formal cousins. A few examples from graphics are: TIFF (free from Aldus/HP/Microsoft; also at many ftp sites) vs CGM (about $400 from ISO/ANSI) Open GL (on Microsoft's ftp server) vs PHIGS (hundreds of dollars from ISO/ANSI)

The availability of specifications on the internet, facilitates their acceptance as de facto standards and rewards the development efforts of the consortia that made the specifications. The acceptance of the specification is the major objective and is more rewarding to those responsible for developing the content of such documents than the monetary sale of individual copies of the standard.

SHOULD THE GOVERNMENT HAVE A ROLE IN SETTING STANDARDS PRICES?

Recently a group of vendors and users in the United States made a request for free or sharply reduced prices for standard documents and asked the US Federal government to look into this need. Hearings conducted by the US Congress noted the possibility of competition among the US standardization promoting organizations for the returns from standards sales. Did such competition lead to reduced rather than increased cooperation among them? Did the high prices of the standards necessary to support the overhead of these standardization groups inhibit their use? A similar examination of this connection between standards and their prices could be made in Europe. (For mandatory standards the question of usage being inhibited by cost is not a relevant question. Thus, for Japan, which emphasizes mandatory standards, compliance and standards prices would have impacts different from the USA which tends to have voluntary standards compliance.

An obvious example from history -- standardizing weights -- shows a role for government in standards. When it became clear that confidence in commerce required government to step in and enforce adherence to technically established weights, the government did so. Now the fact that the government endorses one standard "pound" and participates in many ways to ensure and certify that consumers actually get a "pound," is a matter of course. We would not want it any other way.

Standards organizations are using the various Copyright laws enforced by government to exploit a monopoly on the sale of standards which, in some cases, were developed by volunteers who paid their own participating expenses--and frequently the expense of circulating tons of paper representing drafts leading to all but the final document. Monopolist or cartel practices can lead to good results, only if the monoplists are regulated. In democracies this regulation must come from the government. Unregulated monopolies or cartels have no need to consider "right-sizing," overhead costs eating into profit margins, or whether technological advances eliminate the need for continuing certain practices and staffs. Organizations such as ISO or ANSI are essentially monopolies or cartels whose regulation is essentially internal as it comes in the form of the guidance from the members of the organization who pay dues to help support the organization and determine its policy. Why should organizations such as ISO and ANSI be allowed to sell standards that were developed for wide public dissemination by volunteers?

If organizations, such as ACM, were to successfully lobby governments to coerce ISO to change its policy and if IT standards were made freely available on the Internet, what would happen to ISO? The members of ISO would have to provide further direct membership funding to ISO. Currently, 50% of ISO's income is from selling IT standards. But there is no clear reason why the selling of IT standards should be in the best interests of the members of ISO.

The United Nations does not charge for its services to those who receive them. Rather the members of the UN have decided to pay dues so that services can be offered on a world-wide basis as needed. Why should ISO be different?

The national members of ISO could pay enough dues to cover the administrative costs of ISO. ISO standards could be made freely available on the Internet. Paper copies could be purchased at a price enough to defray only the direct costs of providing the paper copy. Libraries which stock ISO standards documents could welcome library patrons to borrow or photocopy these standards. At the moment the contrary is typically true, and libraries can not facilitate the easy use of ISO standards.

The analogy to the traditional publishing industry may suggest another model, namely the current model of strict copyright rules and priced products. The publishing advantage is the clear sense of feedback as to which products are wanted by the public and the support of a publishing organization that provides a kind of quality control and provides long-term inventory, marketing, and distribution. The standardization groups have, however, different objectives from the private publishing business, as national and international standards documents are more like government documents rather than documents produced by publishing houses. Among other things, the actual authors of standards do not get royalties.

The Internet is an excellent example of a technology which brings a broader policy issue to the fore. Namely, the issue of free distribution of standards is made particularly prominent by the availability of Internet distribution channels. Furthermore, the Internet itself exemplifies how voluntary, consensus-developed standards can succeed. The Internet is the greatest success in the area of providing world-wide communications, and now it strongly influences through its existence the choice of standards by every organization. And its own standards were always on the net, readily available, and free. So how could standards be made more widely available? "Exploring the Internet" describes Malamud's travel around the world (several times) to promote the placing of standards documents on the Internet. He even conducted a brief experiment using one organization's standards. (Read his book for the details, it is probably the most lively book ever on information technology standards.) Of course, putting the standards on Internet made them available free everywhere in the world. But only for the short time of the experiment.

ALTERNATIVES

The argument that standards should be free assumes that the members of SDOs would pay a higher membership fee to allow the SDO to continue to manage part of the consensus process and to maintain the standards in an electronic library for distribution. One counter argument to this position is that the standards process and products should be part of the free market process. In this case, people or institutions who develop standards would be paid for their work, and the price of the standard would support the effort. SDOs would not have members who subsidize its functioning and volunteers would not do most of the work. Perhaps this model could work.

At the moment SDOs often insist on copyright on documents which are drafts in the path towards a standard. An alternative to providing all documents for free or making the documents so expensive that their sale supports all the work done on them is that SDOs support distribution on the Internet of the intermediate documents. The SDO could continue to sell the final standards and thus help cover its costs, but intermediate documents would not be copyrighted and would be freely available on the Internet. These intermediate or `pre-standards' documents would be freely available to all who wanted them. People to whom the final, official standard was important would have to pay for it. The wide distribution of intermediary forms might stimulate interest in and compliance with standards.

The free distribution of pre-standard documents is one step towards free standards. Another evolutionary approach to the wider distribution of standards focuses on a distinction between types of standards. Some standards are only used by large companies to implement major systems, while other standards are readily understood and are used by a relatively large number of people. SDOs might be asked to begin an experiment by first making freely available on the internet relatively simple standards that have a wide audience.

WHY SPEND MILLIONS ON INFORMATION TECHNOLOGY STANDARDS THEN LIMIT THEIR USE?

The current economic model of standards development is that volunteers write standards and SDO members pay the SDO to collect opinion about the draft standards and to sell the final standard. In the days when paper copies were the only practical way to distribute a standard, charging for its public availability was a different matter than it is today with the free availability of some information on the Internet. We are not advocating that paper copies of standards no longer be sold, but that the current free character of the Internet be further exploited to facilitate the wide exchange of standards.

To our minds, information technology standards or, at least, their intermediate forms should be freely available on the Internet. Member dues for SDOs should be high enough so that such increased accessibility of standards information could be provided on the network. The possibility of free access to standards on the Internet should be extensively investigated. Certainly for the information technology community this would be one way to increase the accessibility and utility of these documents and to create in the long run a more responsive and meaningful library of standards documents.

* John Berg is Editor of `Computer Standards & Interfaces' and Roy Rada is a member of the ACM Technical Standards Committee. The authors would like to thank the ACM Technical Standards Committee, particularly George Carson, Jim Moore, and Irving Montanez for valuable feedback on earlier drafts of this column.