PAC: Thoughts about CS 2 (Auto vs. manual)

[Jon Bosak <Jon.Bosak@eng.sun.com>]

These are preparatory notes for next Thursday's PAC phone
conference.

CS 2. Automatic vs. manual committee membership maintenance

Our decision in CS 1 to restrict the role of OASIS to
administrative issues fairly well commits us to some automatic
form of TC membership maintenance, since manual control would
require a management hierarchy of some kind.

We have said that a proposal to create a TC shall conform to
certain criteria.  But we haven't defined the process leading up
to this proposal or the process to be followed in adding or
removing members later.

Note that since we are now using the name "TC" to apply to all the
committees under "C. TECHNICAL COMMITTEE TAXONOMY" on the issues
list, the proposal to create a TC can be a proposal to create a
charter committee whose deliverable is a charter for a technical
specification committee.  But we also seem to be allowing a tech
spec committee to be created without going through this phase.  We
may need to revisit the startup phase later, but considering both
specification and charter committees to be subclasses of TC allows
us to get a lot of mileage out of the maintenance rules we adopt
for TCs in general, as long as we remember that whatever rules we
come up with have to apply equally well to all the different kinds
of TCs.

Our problem is how to ensure that only active participants are
counted as voting members of a committee.

The manual control model has the virtue that you can actually
reach in from above and change the composition of a group if it's
going wrong.  We're abandoning this concept, but we're left with
the problems of how to screen out people who don't look like
energetic participants from the beginning and how to remove
deadwood after a committee has been formed.

Many standards bodies employ the concept of "voting membership" or
"membership in good standing," which is a special status that
members can achieve only through active participation and that is
automatically lost if their commitment flags.
Possible criteria for establishing membership in good standing
include attendance at face-to-face meetings, attendance in
telephone conferences, percentage of votes cast, or some
combination of these.

Of these criteria for judging active participation, attendance of
members at face-to-face meetings is probably the one most widely
used.  An example of a system based on this criterion is the one
used by NCITS, an ANSI organization that provides the American
National Body (NCITS X3 V1) responsible for SGML work in the U.S.
I included this example in a message sent back in November under
the heading "010. Voting membership in a committee", but it bears
repeating here as an existing instance of this method.  The NCITS
rules say:

   A prospective voting member shall attend at least two out of
   three successive meetings of NCITS. A representative shall
   attend the first of these meetings as an observer and reaffirm
   interest in the work of NCITS. Membership becomes effective
   with attendance at one of the next two successive
   meetings. Voting privileges begin with the opening of that
   meeting.

The rule used by NCITS for its subsidiary bodies is significantly
looser:

   A representative of a prospective voting member shall attend
   at least one meeting of the TC, TG, or SG. A representative
   shall attend the first of these meetings as an observer and
   reaffirm interest in the work of the TC, TG, or SG. Membership
   becomes effective after adjournment of that meeting and
   receipt by the Secretariat of applicable fees for the
   membership year, at which time voting privileges begin. For a
   new subgroup's formation meeting, all attendees shall be
   considered voting members...

The termination rules are the same at both levels:

   Voting members of NCITS and its subgroups shall be terminated
   under the following conditions:

      a) The principal and all alternate representative(s) shall
      be warned in writing upon failure of the organization to:

	 (1) attend two out of three successive meetings, in
	 which case the membership shall be terminated if not
	 represented at the next meeting; or

	 (2) return 80% of the total letter ballots
	 (non-accelerated) closing during the present calendar
	 quarter, in which case the membership shall be
	 terminated if the member fails to return at least 80% of
	 the total letter ballots (non-accelerated) closing
	 during the subsequent quarter.

      An organization fails to perform an above action when none
      of the organization's representatives perform the action.

      b) The voting membership shall be canceled by the NCITS
      Secretariat for failure to pay appropriate service fees
      within the time specified by the NCITS Secretariat.

   NCITS or its subgroups may vote to continue the membership
   despite failure of the member to comply with the membership
   criteria in item a) above.

Notice that "member" in the NCITS rules refers (typically) to
organizations, not individuals; in other words, an organization
can meet the requirements for continuing participation in NCITS by
sending alternates.

Several months ago I posted an initial proposal (included for
reference below), a basic assumption of which was that an OASIS TC
must meet in person at least four times a year as I thought was
required by the current OASIS bylaws.  In fact, the bylaws don't
say what kind of meeting, they just say that meetings "shall be
held at a minimum of four (4) times per calendar year" (Article 3
Sect. 7 applied via Art. 5 Sect. 3).  We've since come to an
understanding that phone meetings are to be considered meetings
and also, I think, to an understanding that we don't want to use
physical meetings held all over the world as a way to filter
participation.  So we have to start over with that part.

The problem is that we can't let people log into phone meetings at
random the way we can let them observe face-to-face meetings at
random; that just doesn't work.  Phone meetings substitute lamely
for face-to-face meetings *only* if everyone knows the other
participants and everyone knows who is present.  But if we're not
going to require face-to-face meetings, we can't simply substitute
phone meetings to perform the same screening function.

I have thought of a way out of this, but it's pretty radical: to
encourage fairly large TCs (20-30 members), which is the direction
things like this tend to go anyway if important in an industry,
and then have them form local subcommittees, perhaps using rules
like those used for labor unions and fraternal organizations (I
don't know anything about the details of these processes, but they
prove that you can do this).  Most large committees do seem to
contain clusters of members from particular places and, of course,
larger clusters of people who can get to within a one-hour plane
flight of each other.  If we formed local chapters based on
nothing but geographic distribution, we could probably reduce the
cost and trouble of meeting at a local level to practically
nothing.

I believe that the random level of talent in a randomly assembled
local is no less, on average, than the average level of talent
in any random group drawn from the same larger community (though
there is of course a much greater likelihood that the membership
of a local will be stacked by a nearby large member company), so
there's really no reason, on average, to expect an inherently
lower level of achievement from groups formed this way, and
assigning specific work items to specific geographically
co-located participants would give them the bandwidth of
face-to-face communication and greatly improve their
effectiveness.  In many cases, the local could meet weekly at a
particular coffee shop.  Enormous amounts of work can be perfomed
this way, and the practical advantages if work groups could meet
locally would be considerable.

A simpler alternative form of this plan, which I think would
amount to much the same thing in practice, would be to encourage
the formation of TCs whose initial meeting schedules consisted of
frequent face-to-face meetings in one particular place and thus
would attract the participation of people living near that place,
while not, of course, prohibiting the attendance of people who
could commit to traveling there.  This would retain f2f bandwidth
while not requiring a mechanism for federating the locals into a
larger steering committee.  The tendency of a group living at a
particular location to be dominated by a single company would, I
think, be compensated by the ability of any interested party to
join the committee.

If anyone wants to go further with this thought, let me know.  If
not, please suggest specific alternatives.  I can't think of any.

Jon

==================================================================

[This former straw proposal is from a posting made back in
November 1999.  It assumes frequent face-to-face meetings, which
works, I think, if the meetings are held in one place, but not if
meetings are held in multiple international settings.  I include
this here for reference so that we can easily recycle any parts of
it that still work.]

   Any OASIS member or employee of an OASIS member organization
   can attend the face-to-face meetings of any OASIS TC as an
   observer provided notice is given the chair of the TC [insert
   some procedure about notice here].  Observers cannot vote or
   participate in committee deliberations, though they may answer
   direct informational questions put to them by the chair.

   A prospective voting member of an OASIS technical committee
   shall attend at least one face-to-face meeting of the committee
   as an observer and shall reaffirm interest in the work of the
   committee at the end of that meeting.  The prospective voting
   member then becomes entitled to participate in the committee's
   email list [this needs to be changed] and telephone conferences
   (if any), but cannot cast votes until voting status is
   confirmed through attendance at the next regularly scheduled
   face-to-face meeting.

   Voting membership is conferred upon an individual after
   physical attendance at two consecutive regularly scheduled
   face-to-face meetings of the committee and becomes effective
   after adjournment of the second meeting so attended.  A
   prospective member who fails to attend the second regularly
   scheduled meeting returns to observer status at the next
   regularly scheduled meeting and loses the right to participate
   in the committee's mailing lists and telephone conferences upon
   adjournment of the missed meeting.  The requirements for
   certification as a voting member of a technical committee apply
   to individuals and cannot be met through the substitution of
   individuals representing the same OASIS member organization.
   [Note that this assumes a particular resolution of issue CS 3,
   "Individuals vs. organizations".]

   To maintain voting membership in an OASIS technical committee,
   each member must attend at least two out of every three
   successive face-to-face meetings and telephone conferences (if
   any) and must also respond in a timely manner to at least 80
   percent of all ballots conducted through regular and electronic
   mail closing during any calendar quarter (with the exception of
   properly arranged leaves of absence as specified [somewhere
   below]).  In the case of individuals representing OASIS member
   organizations, this requirement can be met through the action
   of another individual representing the same organization, but
   only if that individual has separately qualified as a voting
   member according to the procedure specified above.

   [Procedure for notifying members that they are in danger of
   losing voting status goes here.]

   [Procedure for actually terminating a voting membership goes
   here.]

   [Procedure for transferring an institutional membership goes
   here.]

   [Procedure for requesting a leave of absense goes here.]