Note: I understand the reluctance to use a word like "policy" here. But it appears to me to be a fact that, when any sort of in-the-moment dispute arises, there is an inevitable appeal to authority of a meta question. Even if we don't want to call things "policies", pointing to a meta question's top-voted answer as the last word of a dispute is functionally identical to actually having and enforcing a policy.

So I don't want to get too hung up on the specific words. If another word which describes how things play out here is a better fit, I'd be happy to edit it into this question. If "policy" is the most accurate one, that's information worth being aware of. Especially if that's not the way users would like the stack to operate. And it really does seem to me that when some users want to see A, and someone does B instead, it turns out that those users find A is policy-like enough that B is formally unacceptable.

This all might be better than edit wars and proxy disputes. But whatever the stack at large might prefer, if it looks like a policy, walks like a policy, and quacks like a policy, it's probably a policy (or a duck, but it's the policy part that is relevant here). I'm open to being better educated on how this all plays out here, but I feel that the issues I'm trying to describe exist identically whether we use the word "policy" to describe them or not. And it's those issues which seem to me to be causing the problems that I am observing.

I've observed some disagreements in and around the stack lately regarding specific actions in specific circumstances. Some of those seem to have become heated and unpleasant for those involved. I don't think it would be helpful to this question to link to, or precisely describe, specific examples. But some explanations offered in many of those situations are very explicitly "this is stack policy, I'm just applying it".

Outside of a handful of specific rules, I'm not clear on how stack policy is defined. We discuss things on meta fairly often, but frequently questions and answers only draw the attention of a small number of members. Worse, the vote totals can change over time. I recall a specific case of trying to determine stack policy which was decided (as best as I can tell) by an answer gaining a plurality of votes in a certain window of time. But now, some time later, the balance of votes indicates the community favoring a different approach than the one described in the previously top-voted answer.

I'm not saying I have a specific, superior suggestion for how to decide on policies. I am, however, unsure if we already have a standard which says that if 41 votes are cast for suggestion A, and 40 votes are cast for suggestion B, suggestion A becomes, officially, the only permitted approach and that the 41st vote represents an absolute justification for A in all subsequent disputes.

Do we have a set of criteria which define official stack policies, especially for questions discussed on meta? Is the balance of votes on any situation raised in meta the final word on that situation? Should we be periodically reviewing specific, key meta posts to see if community consensus has changed (automatically changing enforceable policy with it)?


Policy isn't a really useful term

Don't get me wrong, I see the appeal and I certainly have used it myself (and will probably do again). It is a nice hard edge by which to act or judge actions and that makes it easier. It's also an easy explanation to give, and we do find ourselves explaining certain things over an over again, so especially for those we land on them being easy. I get that. But hard edges make it a blunt instrument.

The Stack (and all the Stacks) actually operate on a best practices. But those are really hard to write down and communicate and are best taught by talking over specific cases as they come up. That process happens in comments, meta, and chat.

So how does the process work?

As mentioned, most of the time we just talk about a singular case. Or to put another way, the default for our meta discussions is that it is talking out what's best to do in a specific case. That discussion will still be useful to link down the line when talking about similar cases.

However, we sometimes get problems that are recurring and end up not being handled well. That's when having a discussion to... create direction for our practices comes in. That direction can include an actual change in practice, such as making a question type off-topic, or just push a harder line on something, such as closing rapid iterations of homebrew reviews as dupes, or push a softer line on something, such as handling questions using phrases like "does it break anything" or "should". But how those come about are also a matter of best practice, so there's no hard line for what they are.

Ideally, and for most of, these directed practices then solve the problem and thus don't need to be re-examined. They either become integrated in our normal pool of praxis, or the problem goes away and the become forgotten. Then, there's the tiny few other cases and that's where the policy term comes in.

Because its only really useful, or specifying its nature, is only really useful once a meta which directed practice becomes contentious. So let's take a (newcomer's, hindsight) analysis of the elephant in the room; Don't Guess the System. It tried to solve two problems (or a split problem, but whatever): A number of users were playing similar or overlapping games, and at the same time questions were asked which didn't specify which system they were asking about. Answers coming in from the wrong system is still a problem, so let's not spend any time on the problems that arise from such answers not being obvious. The solution: require all questions to explicitly state their system. The cost is inconvenience for some number of questions which needs to go through extra steps (the fullness of these arguments can be found on the appropriate metas).

Then time happened. And with it one of the problems went away, specifically users playing mixed or hard to distinguish systems and taking it as the default. Users posting questions without stating the system didn't though, and hasn't despite specific attempts to mitigate it. So the problems which surround it change, the cost seems to be the only effect, and it is coming up against the urge users have to help. And so it gets talked about, discretized from a directing meta to a Policy, revisited (twice, so far) and contentious.

That's probably simplifying it a lot, but my point is that it's just a process happening, and sort of a by-product. Or having to talk about something as a Policy is a by-product of other problems and certainly not a goal. You might particularly see the not-goalness rear its head when a meta discussion is started trying to create a policy.

So, to try to answer the question itself, we try to do what's best for the stack. We look at what's sustainable, both specifically and in general. That means if the practice direction given by a meta is giving problems, it may need to be revisited. But constantly revisiting isn't useful either, and if the community is hard split on an issue, changing practice whenever votes tip past one another isn't sustainable either.

†: Whether accepting the tag alone as such a statement help fuel the issue is a discussion to have elsewhere, but probably not something we stand a chance to solve now.

‡: This is again in reference to the Don't Guess the System issue and to the "5e" subpart. For the latter votes changed with time, and because the change side would be more motivated to cast votes after the fact, I made the call to open a revisit on it. While that has now also happened to the re-revisit of DGtS, but a further revisit hasn't happened. How to handle that kind of case probably warrant its own discussion.


Policies are rightly blunt to direct utility and support meta-discussion.

Policies are for in-the-moment decision making on the main-site. We defer to the law of policy to redirect the discussion to a place where we have a more open environment which is the meta-discussion.

We defer to the authority of policy on main-site because we want to prevent escalation of the conflict. Policies are the tools that enable us to make a decisive incision to avert an escalation of a disagreement before it spirals out of order. They exemplify the line that we drew in the past, we discuss it in meta, but abide by it until it changes.

That is stiff and rigid, but also pacifying. Policies work as an equaliser. A moderator has to adhere to the policy and is held accountable to the community decision - it is a check to their power and the tool that a 300 reputation user and a 150.000 reputation user can wield equally.

A policy is a blunt weapon by design, and that is right. We work out the details later on meta, the fine tool - the peaceful adjudicator who lets us give things the proper care.

  • 1
    \$\begingroup\$ Interesting answer; though I will point out that policy has no value beyond that what it accrues to the people whom it is intended to serve. For a site based on a hobby, guidelines are more useful than policies. \$\endgroup\$ – KorvinStarmast Feb 1 at 18:57
  • 1
    \$\begingroup\$ @KorvinStarmast what is your understanding of their functional difference beyond the term? :) \$\endgroup\$ – Akixkisu Feb 3 at 19:28
  • \$\begingroup\$ I've written policy in real life; doing it well is rather difficult even when that is what you are paid to do. "It's policy" is the tool of petty bureaucrats the world over, and of disempowered and indifferent employees in business all over the place. "That's policy" is an excuse to be lazy and unhelpful in far too many cases. Guidelines for an all-volunteer effort, and our ongoing attempts to capture Best Practices is to me a far better place to start from. Beyond that, see @Someone_Evil's answer. \$\endgroup\$ – KorvinStarmast Feb 3 at 19:31
  • 1
    \$\begingroup\$ @KorvinStarmast what would calling them guidelines change about our current process? Frankly, I would just see: "See our guidelines" and act according to guidelines, but I see a good case for arguing guidlines in comments. So all I can see is more discussion in a place that I then have to flag which already takes a lot of time. \$\endgroup\$ – Akixkisu Feb 3 at 21:11
  • 2
    \$\begingroup\$ It's what people infer when the term policy is used that I have found to be negative, on this stack in particular. Used as a bludgeon. \$\endgroup\$ – KorvinStarmast Feb 3 at 21:42
  • \$\begingroup\$ @KorvinStarmast Would that experience improve if the process stays the same, but we make a guideline about calling guidelines "guidelines"? \$\endgroup\$ – Akixkisu Feb 3 at 21:48
  • 2
    \$\begingroup\$ With this community as it stands right now, I highly doubt it. Maybe with a different mix of people, but there are some overarching SE based things that mitigate against it. And you are so very right about the blunt instrument. I am still annoyed at the tool rec decision, how it was reached, how many perfectly good answers from this sites early days are under seige, etc. \$\endgroup\$ – KorvinStarmast Feb 3 at 21:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .