I just want to point out that this post worries me because of how disconnected it is from actual CoC enforcement.
To take an example: PyCon’s incident response guidelines are public documents. The official PyCon blog also has CoC transparency reports going back years and listing the types of reports and summaries of resolutions. I think people should look up those things (and to make you do it I’m not going to provide direct links!) and read through them and compare actual CoC enforcement against the sorts of fears and hypotheticals being floated here.
And speaking for myself: I used to be a moderator in a Django community space. 99% or more of the CoC-related stuff I dealt with there was handled by removing the offending post(s), sending a private note to the person explaining why, and potentially issuing a few days’ timeout from posting if it seemed like they needed to cool off a bit. The reality of CoC enforcement is actually pretty banal (though the things that do wind up escalating all the way to a Board-of-Directors-type entity, in my experience, are not, but are also incredibly rare).
A lot of posts in this thread seem to take it as granted that it’s somehow easy or common to get someone fully ejected from the community with a malicious/unfounded CoC report. It very much is not!
I think I’ve already been abundantly clear that I’m talking about worst possible cases, not common cases. Those of seemingly similar mind (like David Mertz) have also been very clear about that.
It’s also rare that an Open Source project finds itself the target of a “hostile takeover” attempt. But - ya - it happens.
Something a bit weird seems to be going on with posts such as these. First say “I don’t have anything against the current board, this is about a hypothetical “worst board””, but then move on to “Are we sure the current board knows what they’re doing? They probably do, but maybe not wink”. This really feels disingenuous, a dodge from discussing the actual issue to discussing whether to even trust the board at all. While I’m sure it’s not what Tim (and others bringing up hypothetical situations) means, it’s how it’s coming off to me. I don’t think focusing on hypothetical worst case scenarios is helping people here, and it seems to be veering off topic from the actual changes being proposed.
I don’t accept that it’s anywhere near out of bounds to ask the Board to clarify whether these changes have been vetted by legal counsel. The initial post said nothing about that, neither any follow-up posts, and it’s been questioned by several people now. Marc-Andre Lemburg’s post contains details for why it doesn’t read to him like it passed legal muster. He has extensive experience with this, due to playing key roles in crafting the bylaws we currently have.
That’s an entirely distinct issue from the content of the proposed changes. When proposing bylaw changes, both the specific changes proposed, and the process by which they were created, are thoroughly on-topic.
If Van Lindberg were still on the Board, I wouldn’t even think once about it. But, far as I know, we no longer have any lawyers on the Board, and legal thinking emphatically does not come naturally to most people.
As a former Director myself, and one who - in the early days - spent thousands of his own dollars on lawyers to craft agreements that would guarantee Python’s independence from employers (think the short-lived PythonLabs, where we all paid the legal bills out of our own pockets), I’m probably acutely sensitive to the legal aspects.
But I’m not qualified to address them. If the Board did, great, all they have to do is say so. I’m not looking for an argument about that, just an affirmation that it was addressed.
A partial response for now, and I hope to get to MAL’s questions on the other thread in due course, along with other clarifying questions: there’s a lot to get through, so thanks for your collective patience .
Yes, we sought and are following legal advice on the text of all three of the proposed amendments.
But you don’t really expect that to last, right? The bylaws of a US 501(c)(3) are public records as a matter of US law, but records of private corporation (non-profit or not) Board meetings and actions are not unless the bylaws say they are. Which, best I can tell, the PSF’s bylaws do not say.
So “privacy!” will surely rule the day, and these votes will be kept secret (or spin it as “private” if you must - but regardless of what it’s called, it’s total absence of transparency).
Indeed, I’m open to keeping such votes secret. But, if so, that’s all the more reason to impose as high a bar as possible. Even requiring Board unanimity is approximately infinitely easier to meet than the current supermajority of all voting members. If a member is such a very bad actor as to merit expulsion, it’s not really asking much for 7 people to agree (minimum quorum for a Board meeting).
“But that would effectively give a single Board member veto power!” Yes. That’s the point. If things “go south”, one honest Director would still have some power to limit the damage.
I’m again fretting over worst cases, and more about ideological purges than about financial shenanigans or general chaos-sowing.
As Carol helpfully pointed out on the PSF-Vote list, the Board actually already has unilateral power to change the bylaws without a vote by the members (only a notification requirement before the change takes effect, and I cannot see any practical way for the members to use that to overturn a change the Board has already voted to adopt).
Which means you have for some time now been a PSF member and Fellow in a world in which your worst-case scenarios were already possible. So I think probably we are at (perhaps quite some way beyond) the point where it’s worth asking whether the continued focus on those worst-case scenarios is still productive or relevant to the question of the proposed bylaw changes the Board has asked the members to consider.
I also think that at this point we can probably knock it off with suggesting that enforcement could lead to “ideological purges”. It’s not constructive in the slightest.
(the PSF’s Code of Conduct has, if anyone’s curious, been in effect for over eleven years, and in fact is exactly as old as the concept of PSF Fellows – the Fellows as a distinct group were created as part of a membership reorg that was on the same ballot as the adoption of the CoC, back in 2013)
I’m not trying to make anything “impossible”. I’m a realist. The most that can be done is to try to stack the odds. You don’t care, but I do, That doesn’t imply either of us is being unconstructive. It’s fine by me if you have no use for my perspective. That’s your preference, and you’re welcome to it.
“Both optimists and pessimists contribute to society. The optimist invents the aeroplane, the pessimist the parachute.” - George Bernard Shaw
I’ll add that pessimists also invented the vast majority of sorting algorithms in widespread use by serious applications - while I expect you’d advise “just use quicksort - bad cases are so unlikely it’s counterproductive to worry about them” .
Like some others in the discussion here, I tend to be leery of proposals like Change 3 that concentrate more power in the hands of fewer people (irrespective of how much I trust the current individuals holding those positions).
This isn’t an absolute, though; it has to be balanced against other considerations. But in reading this discussion I realize it’s hard for me to even evaluate the costs and benefits. The main justification I see for the proposed change is this:
This seems a bit vague. Can we get any more detail about why this is believed, what sort of distress this is meant to avoid, and what kinds of scenarios are envisioned that would cause such distress? We’re talking a lot about scenarios where the board misuses its power, but what would be the scenarios where it uses that power appropriately in a way that’s better than what can be done under the current rules? Are there examples of situations in the past where having this rule in place would have led to a better outcome?
Suppose there’s a case of unwelcome sexual advances to a person B that person A just won’t stop. To prevent “the evidence” to the public could well be traumatic to B, especially given the dynamics of public “debate”.
But I don’t think it’s really much better under the proposal unless the Board also changes its rules to make “boot the member” votes secret too (see my reply to Thomas a few msgs up).
So I expect that’s an inevitable change that’s not yet been acknowledged.
After that, the situation is radically different: it’s possible that nobody outside of A, B, the CoC WG, and the Board will ever hear about it. Person A will just vanish.
But that’s just my speculation. It certainly merits a reply from the proposers.
I don’t think the concern being raised about “ideological purges” was realistic, nor particularly close to realistic. It’s more in the category of what in the security world is sometimes called a movie-plot threat.
Sure, someone could put together a complex plan to secretly install their evil henchpeople into enough seats on the PSF Board of Directors to be able to win majority votes and then for some reason only use that power to file CoC cases and end the memberships of some Fellows. But is this something that’s actually realistic enough, here in the real world, to be worth rearchitecting the bylaws around? From the amount of time and attention being spent on this concern you’d think it was something that happens on a weekly basis, if not oftener.
We are, as I mentioned previously, eleven years in to the PSF CoC era, and not only have there been zero “ideological purges” of the membership, I can’t even find evidence that it’s ever been attempted.
So, yes, let us be realistic and focus time and energy on things that realistically might happen. A sorting algorithm which only accepts worst-case data probably would not achieve very widespread use, would it? And I’ve heard it said that practicality beats purity. So maybe we shouldn’t spend so much time worrying about this special case and whether it’s special enough to need handling in the rules.
I know this is one of those threads that everyone feels they must weigh in on, and doing so usually subtracts value, so first of all apologies. But I actually understand what was meant by the “ideological purge” comment that Tim referenced. It was a real set of events, an account of which is at Report on NixOS Governance Discussions - by Chris McDonough .
The situations might not even rhyme: that account leads off with a power vacuum, decisions in censure and booting cases did need to be unanimous within a six-person moderation team, not everyone was booted, and many folks left under their own power, only under threat of a booting. TL;DR: mod team goes rogue and ideological. The Python community is currently much, much, much healthier, but for anyone wondering how bad it can get, well, there y’are.
The PSF has a long established pattern of relying on the Open Source Initiative to define what is and is not an open source license. It is in our expanded mission statement, here: Mission | Python Software Foundation
“Works with the Open Source Initiative to ensure that PSF licenses conform to the Open Source Definition.”
The bylaws change allows us to start making things clearer and then we can work on administrative processes that could make the implementation better. (I won’t ask you to join another workgroup, but I will make sure the Membership Working Group sees this suggestion.)
PSF staff typically does not work on the weekends (and I personally spent the last three days at PyData in London) so the responses may seem slow, but we absolutely sought and are following legal counsel on the text of all three bylaws changes.
The question of a practical way through this discussion is subtle. Chris McDonough’s comment seems to suggest a way that group dynamics play out in the real world. I didn’t follow all the details, but it was certainly messy.
If I follow the discussion correctly, then the PSF is already empowered to take many actions with regard to a PSF member who is violating the CoC. That seems like a practical solution to CoC problems, because it addresses the problematic behavior directly. It’s not clear if revoking membership is a symbolic gesture, or if it directly addresses some real problem.
I think group dynamics are complex, and I’d be worried about the mechanism for membership removal being misused at some future date. While it is largely symbolic, it does seem like the most drastic step the PSF could take. I’d prefer to have a higher standard than majority of the board. Perhaps an appeal mechanism for that board vote would make me more comfortable.
I don’t understand the particular urgency of this change, and I think the design space feels under-explored. What other options were considered? In what way were they determined to be inferior to the current proposal?
As a comparison, we already have a codified procedure to remove core developers by two-thirds majority [1] of the Steering Council. This is a power the SC cannot delegate, but in practice the recommendations of the Conduct WG are highly valued guidance.
The Steering Council also effectively has the power to ban toxic contributors, either permanently or temporarily, from all Python/PSF controlled forums, including Discourse, Discord, GitHub, and the mailing lists. It’s definitely the least fun part of the job, but also during my time on the SC, a responsibility we have taken very seriously. It doesn’t happen often, but it has happened.
Can you tell us more about the nature of that advice? For example, we used to invest a lot of attention and money on legal advice about the Python license and it’s GPL compatibility. You could certainly have asked, Does any of these changes jeopardize the status of Python’s license? But that seems unlikely! What kind of advice did you ask for? What legal risks were addressed?
.
.