doc: discourage private TSC meeting and document process

[joyeecheung]

Establish the principal that TSC meeting should be public and written communication is preferred for private topics. Also, propose a process for recording private meetings, if possible.

The goal of this proposal is to make sure that private sections, especially the one without recordings, remain rare exceptions that need to be justified and double checked, and to prevent them from being a habit or a norm. The norm should be publicly streamed meetings and written communications.

[nodejs-github-bot]

Review requested:

• @nodejs/tsc

[jasnell]

I disagree with this last bit. For many kind of conversations that require private discussion, it's most often better to talk in person that in writing. It's just the reality that talking face to face (zoom to zoom?) is often more productive to resolving issues

[joyeecheung]

For many kind of conversations that require private discussion, it's most often better to talk in person that in writing. It's just the reality that talking face to face (zoom to zoom?) is often more productive to resolving issues

I think my question would be, how many, and are they in the majority of our private sections? Considering the turnout of the TSC meetings is usually <50%, it generally already limits what we can achieve with these meetings other than bringing a topic to the attention of <50% of the TSC. Even when discussions happen in the meeting, written communication as followups generally are still needed to involve those who are not at the meeting.

Note that this is just the principle, "it would be more productive if we discuss it now with the people present at the meeting rather than using another written channel" can still be used as justification, on an exceptional basis.

[joyeecheung]

Specifically, for the "bringing it to people's attention" case, we can give an exception like this:

If discussions must be private, written communication is preferred. If private topics need to be brought to the attention of the attendees, it can be brought up either at the beginning or at the end of the meeting with links or other references provided if no further discussions need to happen at the meeting. For actual discussions of the topic, a private section must be requested.

[jasnell]

I don't think it really matters how many because it's not about a quantity but about what is specifically the most effective and most appropriate for any given case. We've always limited private TSC sessions to matters that are sensitive in nature, it's not like we're just off regularly conducting regular business in private all the time. I do agree that votes and binding decisions should be made via email where all TSC participants have the opportunity to weigh in but that does not mean that we should actively avoid private real-time discussions when they are necessary.

Perhaps a better way of approaching this would be:

In principal, TSC meetings should be public, streamed, and documented. However, there
are times when private real-time discussions or private written discussions are necessary.
When binding decisions are to be made, those should never be made within private real-time
discussions as it is often the case that the full TSC cannot be present given our global
geographic distribution and volunteer model.

[joyeecheung]

We've always limited private TSC sessions to matters that are sensitive in nature, it's not like we're just off regularly conducting regular business in private all the time.

What I've observed so far, I think, is that since there is no quantifying measurement of how sensitive things are that it must be made private, we have developed a habit to make it private even if it's just a tiny bit more sensitive than usual and does not absolutely have to be private, or we just forgot to make it public even though everyone agreed that it could be public (this not only happened to TSC but also loader team meetings as far as I remember).

The suggestion partly LGTM, though I think we still need to emphasize the importance of written communications. https://community.apache.org/contributors/mailing-lists.html#inclusion-and-transparency put it quite well and I will borrow some of that.

[RaisinTen]

Suggested change

	switch to writing communication instead, before proceeding with the private
	switch to written communication instead, before proceeding with the private

[RaisinTen]

It would be a good idea to mandate the retention of clear records for a set period

Suggested change

	section locally and make it available to other TSC members after the meeting.
	section locally and make it available to other TSC members after the meeting.
	An archive of the recorded private meetings must be retained for at least
	2 weeks.

[jasnell]

I'm generally -1 on this part starting from this paragraph down. Sometimes it is simply necessary for something to be discussed in a private session even if an individual TSC member disagrees... otherwise, it would be possible for a single TSC member to unilaterally stall or block necessary discussion on an important matter.

Private TSC sessions should never be recorded in my opinion. We need these to be safe places for TSC members and invited guests to feel safe to express what they need to express.

And just in general, an individual should never feel the need to justify not wanting to be recorded and we should never make them feel pressured to agree to being recorded. There are privacy laws in place that strictly protect this right and, frankly, it's just the right thing to do. Public meeting session are an exception here because it is established up front, before individuals join, that the sessions will be recorded and they can choose to participate or not.

[joyeecheung]

Sometimes it is simply necessary for something to be discussed in a private session even if an individual TSC member disagrees... otherwise, it would be possible for a single TSC member to unilaterally stall or block necessary discussion on an important matter.

I think the proposal still leaves room for that to be requested if necessary, and provides a hatch to prevent stalling/blocking - switching to written communications. Considering followups with written communications is a must for any important discussions due to low turnouts of the meetings, it seems questionable how much not being able to use a private meeting for this is effective in stalling/blocking a discussion. It should also be noted that the TSC meeting only happens once a week, so if a bad actor did exist, it seems the more effective way for them to stall or block necessary discussions is to insist that this must discussed in private meetings, and if the meetings never collect enough attendance and we only get one retry per week, that can stall the discussions much more effectively than e.g. on a private mailing list or private GitHub.

[joyeecheung]

Private TSC sessions should never be recorded in my opinion. We need these to be safe places for TSC members and invited guests to feel safe to express what they need to express.

I think the safety concern comes in two facets: some might feel unsafe to participate when there's no recording going on (personally, as a woman who suffered from verbal harassment in a professional meeting before - completely unrelated to Node.js FTR - this would be my preference. While I trust the TSC would not behave in that manner, being recorded would no doubt make me feel even safer). Also the rule does not forbid unrecorded private meetings, it just discourages them. We can start a vote on whether the recorded or the unrecorded route is preferred, if necessary.

Public meeting session are an exception here because it is established up front, before individuals join, that the sessions will be recorded and they can choose to participate or not.

The private recording is formatted in the same way - before individuals started private discussions, notify that the sessions will be recorded and they can choose to participate or not. If they insist it must not be recorded, give a reason and move on. If there's disagreement about whether an unrecorded private meeting should happen, then just switch to written communication. The difference between the two is just that private recordings are not shared to the public by default. If we have no issues with fully public meetings, it seems strange to argue that TSC-wide-public meetings are breaking even more rules. This provides a third practice that acts as a preferred buffer between fully public and fully private meetings.

I understand that this makes private, unrecorded discussions more difficult, but I think that is the whole point - fully private meetings should always be deprioritized.

[jasnell]

Placing the red X on this since there's another approval so that this does not get merged as is.

[marco-ippolito]

I found myself having to dm another TSC member to get a recap of the private meeting. Having private meetings recorded is very helpful.

[ruyadorno]

IMO a recorded private session does not meet my personal expectation of privacy since that record exists and may be leaked (intentionally or not) or shared privately between other parties without my consent.

The way I see it, the current wording here this PR is removing the concept of "private meetings" (again, in my personal view) which is a valuable tool to have in the project governance.

That said, I fully agree with the idea expressed in the other changes, that this tool (private meetings) should be a last resort and actively avoided until it's really necessary.

[marco-ippolito]

Recording doesnt mean that the recording is shared publicly, so the privacy is respected

[ruyadorno]

well, like I said, if a recording exists that means it may be leaked (intentionally or not) or shared (privately or publicly) without my consent and that does not meet my expectation of privacy

[joyeecheung]

I think there are multiple use cases of private meetings:

1. It's premature to disclose the discussions publicly, but the result of the discussion would eventually be public and would be things we want to let more people know. The impact of it being released would be fairly minimal, and if we are meeting our member expectations, we should not be saying things that we don't feel comfortable sharing externally in the first place. This includes e.g. security discussions, selection of summit locations & hosts, .etc.
2. Private because of publicity concerns, the impact of it being released is still somewhat debatable as long as we are meeting our member expectations - one can argue the community deserves more transparency from a community-driven open source project, others can argue it's sensitive and we don't want drama. The result of the discussion would eventually be public so people would notice that we are having discussions behind close doors. For example, legal and administrative matters, governance discussions.
3. Private because the discussions are about individuals. For example, personnel issues. This is where we'd want privacy to protect individuals. And the result of the discussion would generally remain private.

I think in the case of 1, if we are truly saying things that we don't feel comfortable saying if it gets leaked, we probably need to first reconsider if what we are saying are meeting member expectations. In my very vague estimate that probably is >50% of the private discussions we have. 2 is probably 20-30%. 3 is probably 10% or less.

Personally other than 3, I generally only say things in 1 or 2 that I personally wouldn't mind leaving traces of them. I think it's debatable whether in an open source project, we are even supposed to say that many things that we'd only feel comfortable saying when there will be no trace of it later, especially if the outcome of the discussion would need to be public - the more we are discussing these things behind close doors the more those who are not at the meeting (both in and outside TSC) would question about it. I do agree 3 should be fully private especially when the outcome is also meant to be private, but they show up in the TSC meetings at most once a couple of months from my impression.

[joyeecheung]

Or I think we can divide the discussions into two categories:

1. Does the outcome of the meeting need to be surfaced to the public eventually (e.g. released, documented, announced)? If so, we should leave a trace as much as we can, to avoid keeping the process behind close doors even though we are obliged to disclose the result publicly and may need to answer questions afterwards.
2. If the outcome of the meeting need to remain private, then it automatically gets a exemption from recording in order to limit the exposure and protect individuals.

[jasnell]

Fully agree with @ruyadorno . You can consider this to be me explicitly revoking my individual consent to be recorded during any private meetings I am a part of except only on a case-by-case basis (at which point I would voluntarily indicate my consent only for those specific cases)

[joyeecheung]

It seems somewhat difficult to document this, I'm not very comfortable with bringing up your handle in the document as an explicit exception. Do you want to draft that exception yourself so that I don't misinterpret it? Or would you prefer not to be mentioned in the document?

[jasnell]

There's no reason to include this in the document. I'm making a statement here that I do not give default consent to being recorded in any private meetings except only on a case-by-case basis in which I will express my explicit consent. That is, there won't be any reason to ask me if I object to it each time. My objection is standing unless I explicitly say otherwise.

sorry I missed this last paragraph in my review:

• If there is an objection, the TSC chair should ask for a justification to not
  switch to writing communication instead, before proceeding with the private
  meeting without recording it.

as long as there is still a path to private meetings without recording I'm not -1 to the proposed changes.

[RafaelGSS]

I'm +1, but I'd prefer if we could set a time span for the recordings. For instance, store only the last X months of private meetings.

[RaisinTen]

I'm +1, we should do this. It's very much possible for people to misquote and twist the words of participants in private unrecorded meetings and trouble them unnecessarily. That makes participation unsafe. Recordings are essential for effective assessment by trusted groups like Moderation and TSC. Safety comes first, so this should be done.

[jasnell]

Firmly -1 on this as it enables any single TSC member to cause any topic to be deferred. Some topics are time sensitive and cannot be deferred to async written communication. It must be possible to override the objection and keep the item on the agenda for a private session but this proposed change does not detail that such objection is possible or how it would work.

[jasnell]

This is not sufficient as it doesn't cover how such justification is evaluated. Who decides if the justification is sufficient? What exceptions are there? What if there are objections, etc. This raises more questions than problems it's trying to solve and is not a good policy to have.

[jasnell]

I would drop the vast majority of this proposal and just say that private sessions are discouraged but sometimes necessary.