Hi all, I'd like to start a discussion on the way we select working group chairs. I brought this up on the wg-chairs list some time ago, and saw dns-wg is discussing it[1] as well, so I thought I'd bring it to the wider list. The latest wg chair job description draft[2] says:
Each WG selects their chairs with a clearly defined process on a regular basis. The process should be announced in advance and clearly described to the WG members in advance in order to avoid surprise and to encourage as many volunteers as possible. ... - A selection process that also allows for online participation
My initial questions to the community, with elaboration below: 1. Who should be eligible to participate in a chair selection, either by vote or sharing views? Mailing list subscribers? People who have actively contributed in the past N months? Anyone present at the session (in person? also remote?)? Anyone signed up to the mailing list before a certain time? Checked in for a certain number of RIPE meetings? Something else? I've seen no objections to my quote from the new draft job description, that online participation needs to be possible. (Explicitly not talking about who is eligible to be a chair candidate here, that's out of scope for now, I'm only talking about who should have a voice in the decision.) 2. Should "selection" and "election" be more clearly distinguished? What distinguishes contributing to the consensus from a vote? If we're effectively running an election, should we design it properly rather than calling it something else? Should there be a minimum length for something to be "a statement of support", vs e.g. 15x "+1 for Sasha"? To be clear, I am not in favour of electing chairs, but even less in favour of pretending we do one thing while doing another. 3. Do we actually need 11 slightly different processes to select a chair, one for each working group? I am aware there have historically been some strong opinions on having individual processes, but I'm not convinced this serves our community. We could also still distinguish "special interest groups" from "deciding what the RIPE NCC should do" and have two, depending on the harm a "malicious" chair could do. Though I don't feel that latter factor is reflected in current complexity. I bring this up, because over the years I've been involved in chair selection from several angles: as a candidate, as someone trying to encourage others to put themselves forward, seeing it in other working groups. I've had a look at all the existing WG chair selection procedures, details below. They vary quite a bit: in-person vs. remote participation, mailing list vs. meeting session, explicit ballots vs. consensus. They share a common structure: candidates are presented, WG members express views, chairs declare a consensus, kind of. With some exceptions: open source currently claims to do voting, but we have not actually done that. What "expressing views" and "WG members" mean is vague. One way this is visible: some selections I've watched start to look more like an election, on the mailing list. The chairs will ask for "send your statements of support to the list", and then the mailing list fills with "+1 to <name>". In more than one case, candidates saw the +1's on the list, got their (otherwise not RIPE involved) personal network to sign up to the list, and send +1's too. That creates an awkward situation for the chairs, and I can say from personal experience it is also awkward as a candidate. Some of the current procedures explicitly tie participation to physical presence at a RIPE meeting. DNS explicitly says "mailing list". Most leave it vague. The process for disputes varies wildly too: delegate to RIPE chair, have some voting process run by the RIPE NCC, or undocumented. Many selections are reasonably smooth, I feel we struggle more with finding candidates to begin with, which means some of our "consenus" has an element of "well, glad someone volunteered, sounds fine" (without me meaning to downplay the qualities of all those volunteers). But vague processes certainly can hurt signups for chair position too. Comparison: RIPE PC selection is a straightforward election, open to in-person and remote meeting attendees. I am not advocating we do this for chairs. More similar is the RIPE chair selection, but that process is far too heavy for WGs. I do think it is a good example of a process with similar underlying goals though. I am interested to hear what the community thinks. Is there a real problem to solve? And if so, how should we move towards a solution? I think Mirjam was planning to bring this up in the RIPE community plenary as well. Sasha (without my hat as opensource wg co-chair, but fully acknowledging we are part of this problem, though we hope to bring an improved version to the list soon) [1] https://mailman.ripe.net/archives/list/dns-wg@ripe.net/thread/B4WK4MHXJVCLBV... [2] https://www.ripe.net/publications/docs/ripe-documents/other-documents/draft-... Thanks to Raymond Jetten, Brian Nisbet, Leo Vegoda, and our RIPE chairs for their input in earlier threads. == Current policies: Connect and Database say:
WG members express their approval or otherwise of the presented candidates … the Chair(s) declare a decision
MAT (the shortest process of all) and RIPE NCC services are similar but not identical to Connect and Database. IOT says:
The incoming co-chair will be chosen by the WG using consensus and the remaining co-chair will make the consensus determination.
IPv6 and Security say:
WG Chairs are elected at WG sessions at RIPE meetings. Anyone physically present at the WG session is eligible to participate in the selection process. The candidate does not need to be physically present at the WG session. If possible the Chair should be elected by acclamation by the WG or by consensus after discussion. If the result is unclear, then a secret ballot should be held. In the case of a ballot, votes will be counted by RIPE NCC Staff and/or Chairs of other WGs. The result will be determined by simple majority.
Address policy is similar to IPv6, except with slightly different in-person/remote requirements:
The working group will select new chair(s) at the RIPE Address Policy Working Group session. Those present at the session, either in person or remotely, will determine by consensus among themselves who takes the available position(s). The remaining chair will determine whether consensus has been reached. If the working group finds itself without a chair the RIPE chair will determine consensus. If no consensus can be reached then a secret ballot to elect the new chair(s) will be held at the working group session. Everyone physically present at the session can participate in the secret ballot. Votes will be counted by RIPE NCC Staff, and the result will be determined using proportional representation through the single transferable vote, otherwise known as PR-STV. The winner(s) of the secret ballot will become the new chair(s).
DNS, where consensus can explicitly also be a tie:
At the end of a co-chair's term, the WG will decide by consensus who is appointed to the available co-chair position. In the event of a tie, the consensus tied candidates will draw lots. Consensus will be determined on the DNS WG mailing list
Routing:
these are generally appointed by consensus of those present at a Working Group meeting.
Open source:
If there is one (or more) application, the current co-chair and the new candidate start to discuss about among themselves who will be serving next terms. If no consensus is reached until 14 days before RIPE, then we'll have a voting at the RIPE meeting.