by Allan R. Mears

Considered the sine qua non of standardization for many decades, the role of consensus is now questioned in our fast-paced world. Allan Mears, retired from the British Standards Institution, looks at the necessity of consensus in developing standards that meet users’ needs for speed and reliability.

Consensus is commonly acknowledged to be the essential criterion that distinguishes standards from other normative documents. It is enshrined in the standardizers’ own vocabulary, ISO/IEC Guide 2 [1] (from which the definitions in quotes in this article are taken), where the definition of the term “standard” begins “document established by consensus.”

Is consensus really the principal criterion we think it to be? This article examines this question in the light of the increase in the production of new documents for public use with lower degrees of consensus than traditional standards. These “new deliverables” appeared first as “consortia standards” to specify products, following collaboration among their manufacturers, and have since been developed by some standards bodies. Their purpose is to meet the demand for rapid delivery to keep up with the pace of product development.

Standards
What makes standards special? The full definition of the term “standard” is “document established by consensus and approved by a recognized body, that provides for common and repeated use, rules, guidelines, or characteristics for activities or their results, aimed at the achievement of the optimum degree of order in a given context.”

This definition contains the following four distinct criteria:

• Established by consensus;
• Approved by a recognized body;
• Provides for common and repeated use;
• Aimed at the achievement of the optimum degree of order.

Do all of these criteria matter to the users of standards? Let us examine them and then try to assess their importance for the users of standards.

Criteria for Standards

Consensus
When a document achieves consensus it has gained “general agreement, characterized by the absence of sustained opposition to substantial issues by any important part of the concerned interests.” General agreement does not mean unanimity, only that those holding a minority view are not able to convince the majority that their position is right. Which part is this majority? Can it be all the manufacturers of a product? Can it be all the users who are unhappy with what the producers are willing to provide? The standards bodies do their utmost to balance representation of interests on their technical committees and ensure that the chair is impartial. Appeals procedures exist to solve disputes when opposition to a substantial issue is sustained and consensus cannot be reached.

As a committee cannot be aware of all the factors relevant to a standard, the path to consensus includes “seeking to take into account the views of all parties concerned” by making a draft standard publicly available and inviting comment. Refining the draft in the light of comments to achieve consensus can be a time-consuming process. Standards users have grown increasingly critical of the time this takes and so question the value of consensus. A recent survey, initiated by DIN, [2] found that majority-decision-making was well supported but that a change from consensus was not as important as other ways to speed standards preparation. However, it was also found that a considerable proportion of those in favor of majority decisions wanted to retain veto rights for their own interest groups!

It is implicit in the definition of consensus that a concept of absolute value exists, i.e., that consensus is the same at every level: international, regional, national, and sectoral. However, both regional and national standards bodies modify international standards for use at these lower levels. By “level” here I mean extent of geographical area or, within one nation, an industrial grouping. [3]

Recognized Body
The problem with this criterion is—who does the recognizing? A standardizing body “has recognized activities in standardization” and a standards body is a “standardizing body recognized at national, regional, or international level.” In some instances standards bodies are formally recognized by national governments: some are indeed government agencies. A few standards bodies, ASTM, the British Standards Institution (BSI), and Deutsches Institut für Normung (DIN), for example, have gained international recognition—but why? These bodies have been publishing standards for many years and the users of their standards find them satisfactory in application, though the members of their committees and the expertise of their staff are unknown. All that remains is the impression that standards bodies “do things the right way,” i.e., operate with committees of balanced interests, expose their drafts to the public, and finalize them by consensus. These principles are summed up in the expression “duty of care,” which influences those who prepare standards to act responsibly toward the standards’ users. Recognition must be essentially by the users because standards are voluntary documents so the users are free to choose. They have a strong “brand loyalty.”

Common and Repeated Use
Standards are publicly available for all to use. They can form the technical basis of legal and contractual agreements because they represent what is mutually acceptable to interested parties without bias. This common availability is in contrast to manufacturers’ own specifications or the recommendations for good practice of professional bodies for their members. A standard product is recognized as fit for purpose and safe to use. Standardized products are interchangeable so that parts fit together when bought from different sources and at different times. Standard test methods are available to verify that products conform to their specifications, thus removing a source of dispute between suppliers and purchasers. Provision for repeated use is of value to manufacturers who can tool up for long production runs. It is equally of value to the users who can rely on the ability to obtain the same products over a period of time.

Optimum Degree of Order
Drafting a product specification involves first the choice of those properties relevant to its intended use followed by the choice of values for each property. Optimizing these two choices is the very essence of standardizing products and a lengthy task. For every product there is a balance to be found between too few standard types and too many. Too few leads to the users demanding non-standard products made specially for their needs, and too many can leave the manufacturers with excessive product lines and slow moving stock.

The term “adaptation loss” was coined by Sittig [4] to quantify the loss to the user either of having to accept a (more expensive) product of higher performance than really needed or of having to accept a product of lower performance than required for the task. In general, the user will have to opt for higher performance and, if the product is to be part of a system, this will lead to additional costs as the performance of other parts of the system will have to be raised. Variety reduction is therefore not necessarily a good thing, though often advanced as a benefit of standardization.

The New Deliverables

Origin
Dissatisfaction with the length of time taken to prepare standards led to the publication by non-standards bodies of “normative” documents to meet their own needs. These bodies were initially formed by influential manufacturers with major shares of the global market; their “consortia standards” gained wide recognition. Many users of their products were happy to accept this situation, particularly in the fast-growing information technology sector where rival incompatible systems had become a major headache. Not all manufacturers and users were convinced of the benefits of this non-transparent solution, as it did not enable all parties to contribute to the formulation of product specifications.

In response, some standards bodies, both national and international, have themselves devised new types of normative document prepared by faster routes that do not involve the due process of full consensus required for traditional standards. These new documents are the result of consultation within restricted fora and subject to lower degrees of consensus. Different standards bodies have coined their own names and procedures for preparing these documents, to the confusion of users.

The new ISO deliverables, for example, are the following:

• Technical specifications (TS), prepared within an ISO technical committee and approved for publication by a vote in the committee. There is a degree of transparency involved, leading to what may be termed “committee consensus.”
• Publicly available specifications (PAS), prepared by the experts in a working group of an ISO technical committee and approved for publication by voting in a meeting or by mail. This vote is then confirmed by a vote in the technical committee. Transparency is negligible and leads to “working group consensus” of a lower degree.
• Industry Technical Agreements (ITA), prepared by a workshop outside the committee structure and approved for publication by the workshop members. Transparency is again negligible and leads to “workshop consensus” of a still lower degree.

All three documents are seen as potential standards after a limited number of years. Even these innovations have already been found wanting: at the final workshop of ISO/INFCO [5] (ISO Committee on Information Systems and Services) the need was advanced for “consensus only by the parties affected” in a particular sector.

Conclusions

The criteria of consensus and the reputation (i.e., recognition by the users) of the standards bodies both appear to weigh with the users of their standards. The extent to which this reputation depends on consensus is not clear. In the U.K., the acceptability for use in practice of national standards based on ISO/TSs compared to those based on international standards may offer an indication. However, such a comparison could mislead if the ISO/TSs were originally drafts approved by the U.K. at the formal vote but failing to gain international acceptance. The criteria of common and repeated use and of optimum degree of order do not appear to have the same importance. These are admittedly subjective impressions in an area where hard data are lacking. Only a survey, as thorough as that initiated by DIN, on the relative importance of the four criteria, can provide an objective conclusion.

User experience with the new deliverables of lower degrees of consensus than standards is still limited. ISO has published some 20 TSs that include specifications, tests, and other methods.

Whatever the future holds for consensus, I am convinced that the integrity of the standardizers expressed by the phrase “duty of care” will always remain the underlying, though formally undefined, criterion that enables the users of standards to trust in them. People matter more than processes even when they involve full consensus. //

References

[1] ISO/IEC (International Organization for Standardization/International Electrotechnical Commission) Guide 2, Standardization and related activities–General vocabulary, Geneva, 1996.
[2] Economic Benefits of Standardization, Summary of Results, DIN, Berlin, 2000.
[3] J. Sittig, Economically Optimal Standardization, in Proceedings of the First International Symposium on Applications of Mathematical Methods to Standardization, Wydawnictwa Normalizacyjne, Warsaw, 1972.
[4] A.R. Mears, Standards in Nine Dimensions, ISI Bulletin, Vol. 38, No. 3, pp 90-93.
[5] Standards in a Post-Industrial World. ISO/INFCO Workshop, October 2000. Summary in ISO Bulletin, December 2000, Geneva, 2000.

Copyright 2001, ASTM