One standards organization takes years to reach consensus, while another standards organization takes months. Although changes are occurring in the most famous and internationally powerful standards organization, the International Organization for Standardization (ISO) to speed its consensus process, the organization that makes Internet standards, the Internet Engineering Task Force, is much faster than ISO. Why?
ISO Tradition
ISO is an independent organization for fostering international agreement on standards with a view to expanding international trade. ISO formally consists of national representatives only. To develop standards, ISO relies on a large number of volunteers who participate in 184 Technical Committees. Each Technical Committee in turn relies on Working Groups. Working Groups develop recommendations as working drafts, then committee drafts, then draft international standards and finally international standards. At each stage, people follow an arduous cycle of corresponding with one another, meeting face-to-face, and formal balloting. On an average day, about a dozen ISO Working Groups are meeting at various places in the world. When each Working Group feels each text is good enough, the draft is forwarded through channels to the national standards bodies. Each member country places one vote at each stage. The ISO process from first correspondence to a published international standard typically takes years.
The scope of ISO standards is not limited to any particular area. ISO covers almost all standardization fields. For instance, the President of ISO is from the heating and air conditioning business where technology and standards change slowly. Furthermore, many people who make ISO standards, such as for air conditioners, may not routinely use electronic networks.
While international consensus seems to be assured by the processes of ISO, the long time required for decisions to be reached has proven a major impediment to the influence of ISO in the rapidly changing information technology arena. In partial recognition of the special character of information technology, ISO joined forces in 1987 with the International Electrotechnical Commission and created a Joint Technical Committee. This Joint Technical Committee on Information Technology must, however, follow the time-consuming steps and strict, copyright rules that other Technical Committees of ISO follow.
While ISO has agreements with the national standards organizations which require those standards organizations to accept ISO's international standards, neither ISO nor the national standards organizations are in a position to mandate that industry uses the ISO standards. For instance, some years ago ISO initiated the Open Systems Interconnection (OSI) standard. OSI covers computer networking and the American government initially felt that following OSI should be mandatory in government procurements of computer networking equipment. However, the success of the Internet, while not specifically following the OSI model, led the government to realize that adherence to the ISO standard approach was not cost-effective.
Internet Standards
The difference between the ISO standards processes and others is particularly pronounced when considering the process in which standards are made for the Internet. The key organization in the Internet standardization process is the `Internet Engineering Task Force' (IETF) which does not operate under the umbrella of ISO.
The IETF is a large, open, international community of network designers, operators, vendors, and researchers concerned with the evolution of the Internet architecture and the smooth operation of the Internet. It is open to any interested individual [http://www.ietf.cnri.reston.va.us/home.html]. The IETF Secretariat is funded by the American government through the Corporation for National Research Initiatives.
The IETF standards process begins with someone expressing an interest and then moves to the creation of a Working Group of volunteers. These volunteers communicate extensively by email and make draft documents publically accessible. While a Working Group may have meetings, no final decisions are made in face-to-face meetings but are made over email.
An IETF Standard is preceded by a Draft Standard which in turn is preceded by a Proposed Standard. Proposed Standard status can be reached arbitrarily quickly. The Proposed Standard must have demonstrated utility and be publically available for six months before it passes to Draft Standard status. A Draft Standard must have multiple, independent implementations. With relatively little bureaucracy, a document may pass from one stage to another and reach permanent archive status as an Internet Standard.
IETF standards are different from ISO standards and are not part of an official, government standardization effort. IETF standards are also different from `de facto' standards. A `de facto' standard is not publically developed but is widely used, such as MS-DOS [see `Standards: the Role of Consensus', CACM, 37, 3, pp 15-16, 1994]. IETF Standards are publically developed and are often widely used.
The success of the World Wide Web (WWW) is related to its close connection to Internet Standards. The technological success of the WWW is based on three information devices: addresses on the Internet, protocols for communicating hypertext on the Internet, and the markup of hypertext files. Each of these devices corresponds to an Internet Standard. The rapid evolution of the WWW is well supported by the fluid IETF standardization process and would be difficult to imagine occurring under the ISO standardization procedures
Where is ISO heading?
To benefit from the positive experiences of the IETF and for other reasons, ISO is changing. The General Assembly of ISO, in extraordinary session on November 4, 1993 in Geneva, endorsed a package of proposals for restructuring the organization in order to meet the challenge of producing standards faster. In that same meeting, ISO fixed as a target the development of consensus on standards in `three years start to finish' [see `Standardization Activities' Computer Standards & Interfaces, 16, pp 375-383, 1994]
ISO is implementing in 1995 a mechanism by which `Publically Available Specifications' can be accepted as ISO standards without needing to go through the normal structures and processes of ISO. Thus ISO could accept from the IETF, the Object Management Group, or any other organization a document and label it an ISO standard.
In early 1995 to facilitate information exchange, ISO started a WWW site [http:/www.iso.ch/]. This site includes the full catalog of ISO Standards and drafts and a meeting calendar. However, this ISO information service does NOT include the actual standards or their drafts.
Some of the volunteers who write ISO standards are using new communication channels to support faster and more effective decision-making. For example, in March 1995 a set of ISO Working Groups used the WWW to support their meeting in Sydney, Australia. Materials from each daily session were placed onto the meeting WWW server [http://yarrow.wt.com.au/~realize/]. Standards experts who were not able to attend the face-to-face meetings in Sydney, Australia could then actively participate in the technical dialog via email overnight.
Is this enough?
While the evidence of ISO and its volunteers moving to faster decision-making are encouraging, are they enough? The `Publically Available Standards' fast track may bring good documents quickly to ISO standards status, but this fast track will short-cut the normal consensus process in a radical way and may have various, unanticipated impacts. What organizations deserve to provide standards on the fast track? If a consortium of software vendors succeeds in fast tracking a new standard which is NOT compatible with a `de facto' standard, will those vendors have cheated the public?
What difference will the ISO WWW server make? It contains only administrative information. Where are the standards and the draft standards? The increased used of information technology in some ISO Technical Committees is not balanced by a change in rules of ISO to accommodate the new means of information sharing.
ISO standard development results in a copyrighted document to which relatively few people have immediate access. ISO standards are not freely available on the Internet or anywhere. In contrast, Internet Standards and draft standards are freely available on the Internet.
Proposal
A variety of approaches might be taken to cross-fertilize the methods of ISO and of IETF. For instance, the Joint Technical Committee on Information Technology (also called JTC1) might experiment with IETF-like procedures by sharing some standards on the Internet. Or the IETF might borrow from the ISO process and send its standards to the national bodies which constitute the membership of ISO.
JTC1 might experiment with a new method on a 5-year trial basis. New `Experimental Standards' would be authorized by ISO and could be cited by industry. These Experimental Standards would be developed in an entirely open process, such as that used by IETF. Mailing lists and news-groups would be primary vehicles of communication. And the Experimental Standards could be freely copied.
In further detail, the developers of JTC1 `Experimental Standards' would first produce a `proposal', and when they were satisfied, they would send it to the national bodies of ISO for comment. After a short period, the developers would reply to the comments and the national bodies would then decide whether or not to accept the proposal as a Draft Experimental Standard. After another time period during which the Draft was tested in public, it might leave Draft status and be simply an Experimental Standard.
Another alternative is for IETF to begin discussions directly with national bodies that also belong to ISO. IETF could ask these national bodies to review and endorse the IETF standards without ISO involvement!
Conclusion
What other issues might this column have raised? IETF and ISO are different in more ways than sketched here. IETF expects implemented standards, whereas ISO accepts hypothetical standards. The conflict between speed and consensus is not the only conflict. It's actually a three way battle, speed versus consensus versus quality. Both speed and consensus are enemies of quality, in different ways, of course! However, no matter how many other issues one might rightly raise, the conclusion remains that a slow consensus process does not provide needed information technology standards quickly enough.
The ISO standards process has traditionally been based on the circulation of paper drafts of standards and face-to-face meetings to discuss revisions to these drafts. This cycle of circulating paper drafts and traveling to meetings to discuss changes often takes years. The IETF on the other hand has reduced the need for many, long, face-to-face meetings through the frequent exchange of drafts via the Internet and through extensive email discussion of the content.
The open, digital libraries that the IETF maintains help volunteers speed their standards development and still maintain consensus. The success of the IETF approach speaks for itself.
Acknowledgements:
The author accepts full responsibility for any unwelcome biases or errors but thanks the ACM Technical Standards Committee, Tony Rutkowski, and several others for help. John Berg started the earlier published CACM column on `Standards for Free' on which this column builds. Brian Meek raised the importance of quality versus consensus and speed. A letter from Roger Price to the `Communications of the ACM' which appears in the August CACM stimulated the inclusion in this column of the pointers to the Joint Technical Committee on Information Technology. Roger's help has proven invaluable in various other ways, and the `Proposal' presented in this column comes from Roger.