The OASIS Process for Structured Information Standards

April 10, 2000

Jon Bosak

OASIS, the Organization for the Advancement of Structured Information Standards, is a nonprofit corporation founded in 1993 under the name SGML Open. Originally a consortium of small software vendors and large customers devoted to developing guidelines for interoperability among SGML products, OASIS changed its name in 1998 to reflect the inclusion of CGM Open and the growing popularity of the XML subset of SGML.

In 1999 OASIS began two important projects -- the XML registry/repository at XML.org and, in cooperation with UN/CEFACT, the Electronic Business XML (ebXML) initiative. It also changed its membership rules to include individual members as well as corporate sponsors. Today OASIS is an organization in transition, whose future direction depends largely on the extent to which interested individuals decide to participate.

The process by which OASIS develops information standards is also in transition. The overall structure of OASIS is that of an ordinary nonprofit corporation with an elected board of directors. The board is responsible for business and policy decisions, while the membership itself is responsible for deciding substantive technical issues such as the approval of interoperability guidelines. As would be expected in what was historically an industrial consortium, "membership" for purposes of electing the board and voting on technical resolutions includes just the organizational members, with one vote going to each organization. More will be said about the relationship of organizational and individual members below.

Robert's Rules of Order

As a nonprofit corporation, OASIS is bound by law to follow the procedures set forth in its corporate charter and bylaws. The OASIS bylaws, like those of most U.S. corporations, specify Robert's Rules of Order as the procedure to be followed by the OASIS board of directors and by committees created by the board. Thus, adherence to Robert's is a requirement imposed by law. This means, among other things, that members can seek legal relief if their rights are not respected under the traditional parliamentary process specified by Robert's.

While potentially complex and sometimes tedious, the Robert's process has a number of advantages for an organization devoted to the development of industrial standards. As the synthesis of some four centuries of evolution in parliamentary procedure, it is the epitome of majority rule in practice. It has been thoroughly debugged and tested under just about every conceivable set of circumstances; it is built into most of our governmental and corporate institutions; its documentation is very widely available (see below for further information); and perhaps most importantly, it can continue to function even when resolving questions about which there are deeply opposed points of view. It is not surprising, therefore, that Robert's is the process used by ANSI, IEEE, and a number of other highly regarded standards organizations.

Robert's also has some disadvantages, but they are not the kind that some people imagine. The widespread belief that Robert's necessarily imposes an inherently slow and difficult procedure on technical development is simply not true. Robert's rules for committee work are much more relaxed than the rules for full deliberative assemblies that most Americans learn in civics classes, and, in the hands of a knowledgeable chair, a committee running under Robert's can move just as quickly as a less structured group. It's true that the process slows down dramatically when confronted with real disagreements among the participants, but so does any process that recognizes disagreement and attempts to deal with it fairly. The difference between the traditional process and more informal ones is that the traditional process always produces a decision no matter what the level of disagreement, while less structured processes dodge the problem by appealing to a higher authority or, in the worst cases, simply ceasing to function.

The real problem with using the traditional process in standards work has to do with its quorum requirements. A quorum is the minimum number of members that have to be present before a committee can conduct business.

Under Robert's, committees are made up of particular individuals appointed by name. A committee can invite observers to participate, but it cannot add voting members, or remove voting members, or change its charter, or do anything else that might run contrary to the intention of the body that created it. So the composition of an existing committee can be changed only by formal resolution of the body that appointed the committee. There are very good reasons for this system; it ensures that committees include only those people judged qualified to carry out the work, and it prevents committees from being "stacked" by the ad hoc participation of groups with views on a particular question. But frequent personnel changes in modern companies make the staffing of technical committees by direct appointment unwieldy in practice.

One solution to this problem would be to appoint to the committee everyone who expressed interest in it and trust that enough of them would remain involved over the life of the committee to carry out its work. But here the quorum requirement intervenes. Under the default process specified by Robert's, a committee meeting has to be attended by a majority of all the appointed members in order to conduct business.

The quorum requirement is important because it prevents a small subset of members from acting in the name of the committee and thus ensures that decisions actually represent the thinking of the majority of the group. But it is a sad fact of standards work that few of the people who express initial interest in the work of a committee will ever actually attend any given meeting. This problem is exacerbated by the impossibility of substituting one person for another without the approval of the body that created the committee. The result is that a Robert's committee must consist of a relatively small number of specifically named people who are genuinely committed to attending every meeting; otherwise it will fail to achieve a quorum and will be unable to conduct business legally.

To address these and other issues related to the creation of technical committees (TCs), the OASIS board of directors has created a Process Advisory Committee (PAC) and chartered it to develop a set of changes to the OASIS bylaws that will provide for easier committee maintenance. The PAC is expected to deliver a package of proposed bylaw amendments in time for distribution at the OASIS meetings to be held in conjunction with the Paris XML Europe conference in June 2000.

Interim Procedures

In the meantime, the OASIS board, with the advice of the PAC, has created a temporary set of procedures for TCs based on the default Robert's Rules process specified by the current corporate bylaws. The complete interim procedure for the creation of an OASIS TC can be found at http://www.xml.org/archives/workprocess/msg00123.html

This complete procedure includes forms to be used at various stages of the process, which can be summarized as follows.

1. To handle the creation and maintenance of TCs, the board has instituted the Advisory Committee on Technical Committees (ACTC). The ACTC is made up of one chair from each technical committee and is itself chaired by the OASIS Chief Technical Officer, a position that is at this writing held by Norbert Mikula of DataChannel.
2. A group of OASIS members (individual members or employees of member organizations) wishing to create a new OASIS TC draws up a statement of purpose, proposes one or more discussion leaders, and submits to the ACTC a request for the creation of a mailing list. Under normal circumstances, the ACTC approves the creation of the list and appoints the discussion leaders named in the proposal.



3. When the list is operational, the discussion leaders submit an announcement to the membership (all of the individual members and the designated representatives of all of the organizational members) stating the purpose and initial membership of the list and inviting the participation of all interested OASIS members and employees of member organizations. The announcement is forwarded to the general membership list. Discussion on the new list begins a week or so after the announcement.
4. After everyone interested has had a chance to join the list, the discussion leaders post to the list a statement of purpose for the proposed TC, a proposed TC meeting schedule, a list of proposed deliverables, and the names of the proposed chair. The decision of whether to determine these parameters in advance or work them out on the mail list is up to the discussion leaders. Then they call upon those members of the list who agree with its proposed purpose, deliverables, and chair and who can commit to attending the meetings to nominate themselves to be the voting members of the committee.
5. After another week or so to give people a chance to respond, a list of the names of those who committed to attending TC meetings according to the proposed schedule is submitted to the ACTC. The ACTC can, in theory, create TCs with any membership it likes, but typically it will create a TC composed of the list as submitted and will appoint the chair named in the proposal.
6. Coincident with its creation of the new TC, the ACTC requests the OASIS board of directors to add the chair of the new TC to the ACTC so that every OASIS TC is represented on the ACTC.
7. Once created, a TC runs according to the default process specified in the OASIS bylaws, which of course is Robert's. New members can be added to the TC, but only by resolution of the ACTC.

It should be obvious that the process just described won't scale gracefully to a very large number of TCs. The PAC chose the simplest procedure it could devise that would work for a few months and still be completely legal according to the OASIS bylaws.

Important Features of the Process

Even though it is explicitly a placeholder, the interim process has some interesting features that will almost certainly carry over into the long-term version.

First, all the TC mail lists are publicly visible. This is substantially different from the policies regarding the visibility of working group mail lists in some other organizations.

Second, while the agreement to attend meetings necessarily restricts the voting membership of TCs, any OASIS member or employee of an OASIS member organization can subscribe to any OASIS TC mail list at any time and can contribute to discussions on that list. So while in form a TC consists of a small number of members and a much larger number of observers, it can just as easily be looked at as a large mail list whose discussions are collected and ratified by a group of appointed voters. Which view is more appropriate depends on the level of meaningful contributions on the list by the nonvoting members. This structure allows a large number of "designated lurkers" to track and comment upon the work of a TC while guaranteeing the dedicated involvement of a group of appointees who can meet the quorum requirement and thus provide a legal foundation for the process.

Since every TC mail list is publicly visible, there is, of course, nothing to hinder the open discussion of OASIS TC work in public forums open to everyone.

Another feature of the OASIS process is that voting membership within committees is by the individual and not by the organization. This gives OASIS an interesting bicameral quality, not unlike the relationship between an upper and a lower house of parliament or the Senate and House of Representatives in the U.S. Congress. Basic OASIS corporate policy decisions and elections of the board of directors are by vote of the organizational members of OASIS, but the committee votes that determine the design of the standards produced by the TCs are made by individuals. This means that individual OASIS members have exactly the same status within TCs as employees of large vendors and big XML customers, while the big players retain just enough influence over the overall direction of the organization to keep them paying substantial organizational dues, thus subsidizing the lower rates for individuals.

The Process Advisory Committee

As already noted, the PAC is working between now and the XML Europe conference in June 2000 to define a more sophisticated committee structure that will scale to an indefinitely large number of TCs and will provide for the automatic addition and removal of members. While the PAC is not a TC (it lives one level up, a sibling of the ACTC), its mail list is nevertheless open to public view at http://www.xml.org/archives/workprocess/maillist.html and the mail list can be joined by any OASIS member or employee of an OASIS member organization.

The current membership of the PAC is as follows:

   Terry Allen, Commerce One

   Jon Bosak, Individual member, PAC Chair

   Robin Cover, Isogen

   Eduardo Gutentag, Sun Microsystems

   Debbie Lapeyre, Mulberry Technologies, PAC Secretary

   Tommie Usdin, Mulberry Technologies

   Lauren Wood, SoftQuad

Though appointed by the board of directors rather than the ACTC, the PAC was formed by basically the same process sketched out above. These are the people from among the two dozen members of the workprocess mail list who were willing to commit to the meeting schedule.

At this writing, the PAC is about a fourth of the way into the design of the long-term process. A recent issues list can be found at http://www.xml.org/archives/workprocess/msg00094.html.

A copy of the first set of issues to be formally decided by the PAC appears as "Design Principles for OASIS Committee Process" in the minutes of the PAC meeting of 2000.01.20, which can be found at http://www.xml.org/archives/workprocess/msg00073.html.

If you want to track the progress of work on the long-term process for TCs, you are invited to monitor the workprocess archive or, if you are an OASIS member or employee of a member organization, to join the workprocess list.

About Robert's Rules of Order

General Henry M. Robert wrote four editions of the work that bears his name. The last of these, a complete overhaul of the 1876 original, was published in 1915 as Robert's Rules of Order Revised, or RROR for short. RROR went through two very minor revisions (so minor that they kept the same page numbering), the first appearing in 1943 and the second or "Seventy-fifth Anniversary Edition" in 1951. RROR was printed in a format small enough to fit in a pocket, so it was truly the "Pocket Manual of Rules of Order" that Robert originally intended.

To maintain copyright in what had become a family industry, Robert's heirs completely revised the work in the seventh edition, published in 1970 as Robert's Rules of Order Newly Revised (RRONR). This edition could no longer fit in a pocket, but it was still reasonably compact. The family continued to add to the book, however, and the current edition -- the ninth -- is over 700 pages of lawyerly, rambling prose. To be fair, I should note that some of this bulk is value added in the form of language that specifies useful things like procedures for writing corporate bylaws and rules for running political conventions, but I have to confess that I cannot bring myself to read the current edition, and refer to it only to consult sections pointed to from the commendably thorough index.

Like most corporate constitutions that use Robert's as their parliamentary authority, the OASIS bylaws contain a formula that automatically makes each new edition of Robert's the normative one, so in legal fact, it is the ninth edition that governs OASIS business. But while the ninth edition is the one that serves as our final authority, it's not where I go to answer most practical questions about procedure. When I have a question, I go to my copy of the 1951 edition. Aided by its shorter scope and clearer organization, I can find the answer to most ordinary procedural questions in under a minute.

In addition to being shorter than RRONR, RROR is written in an old-fashioned, no-nonsense, soldierly style that I personally find a lot easier to take than the bland corporate prose of the later editions. This is, of course, a matter of taste, but for committee use I would recommend ordering a copy of the 1915 RROR (now in the public domain), which can be obtained as a paperback reprint from Quill, ISBN 0-688-05306-8. Unfortunately, the text of the Quill reprint has been enlarged and printed on cheaper paper than in the original, and as a result, even this old version no longer fits in a pocket, but it is otherwise as good an authority on parliamentary procedure as you will find. Paperback reprints of editions published before the major revision of 1915 are not faithful guides to current practice and should be avoided.

The 1915 RROR is also available over the Internet beginning at http://www.constitution.org/rror/rror--00.htm.

This version has been reformatted to replace all the references to page numbers with links to sections instead, which makes it much easier to use online. For convenience, I have downloaded the entire book and put it in a single zip file at http://metalab.unc.edu/bosak/rror/rror.zip.

Just unzip this into a directory and start with rror--00.htm, taking care to preserve and abide by the copyright notice at the top of the file.