Welcome to Codidact Meta!
Codidact Meta is the meta-discussion site for the Codidact community network and the Codidact software. Whether you have bug reports or feature requests, support questions or rule discussions that touch the whole network – this is the site for you.
Drafting the Codidact Arbitration & Review Panel [duplicate]
Closed as outdated or superseded by ArtOfCode on Nov 21, 2020 at 21:10
This question has been superseded or is outdated. For more up-to-date information, see the linked post. See: Second Iteration of Drafting the Codidact Arbitration & Review Panel
This question was closed; new answers can no longer be added. Users with the reopen privilege may vote to reopen this question if it has been improved or closed incorrectly.
Since Codidact was founded, we have had one rule leading our path every step of the way. It's the rule that community comes first. That the Codidact "staff" shouldn’t overrule the community, but could be overruled by it.
In any community, acts of moderation should be rare. And even rarer is the need to review these decisions. However, there will inevitably be cases where certain situations need to be reviewed:
- Users think that an action (for example, a suspension) is wrong or even malicious.
- A moderator might misbehave and violate our light Code of Conduct or our Terms of Service.
I want to emphasize that there have been no such cases yet, and there will likely (and hopefully) not be for months, if not years. But we can be sure that there will be one at some point in the future. When this situation does arrive, it's better to have an existing process that can be followed to guarantee the best resolution of the conflict rather than coming up with a brand new process on the spot.
It should be clear that such a process shouldn't involve "us" (the Codidact team), but rather "you" (the community). Hence, at some point, it was decided on the old forum that we'd eventually have some kind of review panel, which would be responsible for these cases.
While there will probably be no "panel elections" for the time being, because the panel members would still be a large percentage of our community members (which wouldn't exactly make sense at this stage), we have made a start on the Panel review process. It is based on these three principles:
- The Panel decisions are binding to moderators and the Codidact team1.
- Every party should be heard before any decision is made.
- The Panel shall be independent and impartial.
This process is currently only a draft. We'd like your feedback, and welcome any suggestions for changes to it. Please leave them in answers to this question.
You can find our current draft here.
-
For obvious reasons, there are some legal limits. However, in these cases, we have tried to strike a balance between legal and community interests. For example, in such a case, the Panel may decide to publish our reasons (with private information redacted). ↩
Article 17.1 disallows successive warnings within two years. I don't think that's best. The second malfeasance may be of …
4y ago
Article 27 requires members to explicitly recuse themselves if they feel they can't be impartial. I suggest explicitly a …
4y ago
It might be worthwhile for this document to make it clear which of its provisions about moderators apply to moderators p …
4y ago
Article 30 discusses that substitutes are sworn in if and only if at least 1/3 of the panel is recused. What happens …
4y ago
I propose that we need a clause governing the criteria and process for removing a panel member in the very rare case tha …
4y ago
Some thoughts: - A clarification of the term "community" might be needed. My take is that this is meant to be a Codid …
4y ago
Article 22 says that all decisions are to be published, exclusive of summary judgments. I think it should specify that t …
4y ago
Note: The comments indented as list items have been added by luap42 as a reply. Articles 7, 13 and 20 could be rep …
4y ago
Article 1 discusses each appointee's vote scores. When trust levels are rolled out, will this clause be affected as well …
4y ago
9 answers
Article 17.1 disallows successive warnings within two years. I don't think that's best. The second malfeasance may be of a different type from the first or a lesser degree than the first, in either of which cases a warning may be warranted. While a second warning within two years may be inappropriate in most cases, we can probably leave it to the panel to decide.
Article 27 requires members to explicitly recuse themselves if they feel they can't be impartial. I suggest explicitly adding a point here that if a member of the panel is the one raising the charge against the moderator in question that said member is automatically recused without having to go through the process discussed in Articles 28-29.
It might be worthwhile for this document to make it clear which of its provisions about moderators apply to moderators pro tempore - those appointed by the Codidact organization to help get a community started, before there's sufficient critical mass to elect permanent moderators.
For example, as mods pro tem are selected by staff rather than by the community, perhaps they're removable by staff as well. (I would assume that this is the case right now, in these wild pre-ratification days, as there are no permanent mods, only fledgling communities, and no panel.)
Article 30 discusses that substitutes are sworn in if and only if at least 1/3 of the panel is recused.
What happens when exactly two members of the panel are absent? 1/3 of 7 is 2 and change, so if exactly two members are missing, that's less than 1/3.
Articles 16 and 17 discuss scenarios requiring a 2/3 majority. If no members are recused, that's 5/7. If exactly one member is recused, that's 4/6. But if two members are recused, you need 4/5. Compared to the 71% majority required for a full panel and the 66% majority if one person is missing, 80% seems unreasonably high, being a 14% range (with respect to the total panel).
I propose one of the following solutions:
- Readjusting Article 30 to kick in if more than one member is recused; this prevents a 5-person panel from existing, limiting the range to just 5%. Alternatively, readjust all fraction to round to the nearest integer, not the next integer, which has the same result.
- Readjusting Articles 16-17 to kick in at 3/5, not 2/3. For higher-member panels that's still 4/6 or 5/7, and it maintains an 11% range, not 14%.
Some thoughts:
-
A clarification of the term "community" might be needed. My take is that this is meant to be a Codidact network-wide panel. But currently we use the term "communities" as in this listing: https://codidact.com/. I take it that having one panel per site ("community") isn't what you wish, at least not until some distant future where there are great amounts of users to justify that.
So the "community" who is responsible for electing the panel is that of Codidact Meta? Where everyone is welcome to participate.
-
Logistics - how to do this in practice? It would be great if the actions of this panel are transparent, yet the nature of the matters discussed calls for privacy, so there's a delicate balance there. Where are the reports and actions to be documented and who have access to it? I think we all agree that it is important that whoever stands accused of something has access to their specific case, to take part of all evidence held against them etc.
Article 22 says that all decisions are to be published, exclusive of summary judgments. I think it should specify that the decisions be published with reasons included, and inclusive of summary judgments (though those may not have reasons stated). The more transparency into the workings of the panel, the more trust people will have in it. For a summary judgment in particular, I don't like that it can have been noticed by only two panel members, has no appeal, and has no publication.
Note:
The comments indented as list items have been added by luap42 as a reply.
Articles 7, 13 and 20 could be replaced by a single article stating that each case which is handled by the panel is assigned a reporter.
- This will be considered. One possible problem is, that the Title 4 proceedings reporter doesn't need admin privileges in most cases, but we can find a formulation for that.
The first sentence of the second paragraph of the preamble should be adapted:
The Panel’s duties are reviewing moderator actions and reviewing whether a moderator is suitable for the role or should be removed if this is questioned.
Removing primary eliminates possible interpretational issues like "Oh, and what are its secondary duties?". By defining clearly what the panel has to do and what not also eliminates the need of Article 33. Removing "in more serious cases" makes clear that the panel can theoretically always intervene but also removes the uncertainty what a more serious case is and what not.
- I'd not do this for two reasons: First of all, the panel is also responsible for other cases (Title 4). Secondly, I'd rather have talks than punishments. Removing "in serious cases" goes against that principle.
Can someone please explain the reasoning behind "In the first election, the three members with the lowest vote scores shall be elected for one-year terms only." (Article 1). Why doesn't this simply apply for everyone in the first election?
- See manassehkatz's comment. The idea is, that we won't change the whole panel at a time, but about the half (once 3, once 4).
Article 11 states that only the Codidact team and other moderators are eligible to initiate a Moderator Review Proceeding. Why can't regular users directly initiate one? I think that this opportunity must be there in a community-first approach. For the concern that this could lead to unwarranted and many appeals at once, there can be threshold in form of providing evidence. Evidence can be reviewed and according to Article 8 be rejected if "obviously unreasonable".
- Warranted moderator review will (and should) be rare. I'd rather not have many moderator review complaints because of possibly unfair question closure. We can revisit this if there is a need. Furthermore, communities can choose for themselves, how mods are selected and hence logically also "deselected". If a community mistrusts a mod, they can remove them without initiating a review procedure, which is for egregious Codidact rule violations.
Speaking about Article 8, this should be a general clause as the reporter articles should be as this applies to every case presented.
- I'll review this.
For Article 12, why does someone on the team need a moderator status to carry out the designated duties? In my understanding (especially as @luap42 answered me about how these terms are defined), a team member is "responsible for the administrative day-to-day operations" - is moderating considered an administrative day-to-day operations? I thought day-to-day operations are concerned with running Codidact which not only evolves about moderating but rather organising the project, the volunteers and preparing documents for boards and governing bodies. This sentence needs some clarification, especially with a legal entity in mind as the removal of a moderator status of a team member shouldn't impair the project. (See the bus factor.) It also helps to show that we all are in it and that the panel views everyone as equals in their operation.
- You are correct, that some dev won't need moderator access in most cases. However, they might get admin access (access to error logs and some site settings for example), which includes all privileges moderators enjoy. In these cases, the special provision applies.
How is option (b) in Article 17 reasonable? "Yes, you did something wrong - but hey, we know you, it's alright." Option (c) should be the main action taken but there are others problems as noted by @msh210. However, I think that (b) should be removed - personally, I don't get it. (I favour more warnings instead of acknowledging something while doing nothing.)
- This is for comparatively minor offenses, where the mod did something wrong, but it isn't severe enough to threaten removal (=formal warning) and we also don't want to dismiss the review as "unfounded"/"invalid", because it's technically correct.
Temporary replacements for recused members (Article 30 and 31) should be randomly drawn from a pool of candidates for which every member of the community can register. This reduces the probability of bias in choosing replacements as well as situations where the same individuals are repeatedly taken as replacements therefore acting as unofficial panel members even though they were never elected. The panel should be as free from bias as possible - randomising replacements keeps this relatively low.
- This is worth looking into.
Article 32 is in my opinion highly questionable. Either the board itself has the right to change the rules or the community as a whole. (I favour the community as a whole.) The Codidact team (whoever this might be after the creation of a legal entity) should have no say in this as this basically allows to dismantle the panel whenever the team feels like this. It doesn't help that "discussions with the community about the merit of the change" are held as no option is indicated for the community to stop an unreasonable change.
- Who exactly is able to change these policies will be defined by the constitution/by-laws of our organization. For now, the "Codidact team" is used as a placeholder for both all actions coming from "us" and not coming "from the community". Requiring public discussion should eliminate most possible cases of abuse. Furthermore, the community will send representatives to the board, which can have direct influence. This should be enough for now.
Besides these specific points, terms should be defined more precisely. As answered in the comments of @Lundin's answer, the panel is supposed to be network-wide, yet this word doesn't appear once in the rules (expect for removing a moderator network-wide). Similarly, as soon as there are help pages or documents explaining how the Codidact team or boards are created, these should be linked. (My last sentence only applies as soon as there is a legal entity - for a private venture I don't think that it's necessary.)
- This will be reviewed for the next draft version. However, many things with regards to legal formation are yet open and need to be discussed with a lawyer in the future and can't be decided yet.
Article 1 discusses each appointee's vote scores. When trust levels are rolled out, will this clause be affected as well to instead reference the appointees with the lowest trust levels? Additionally, until TLs are rolled out, what happens in the case of a tie? What about if appointees differ by only a handful of points? Because of a single upvote, should one appointee be given another year on the committee?
1 comment thread