Partnering with Communities to Modernize Policies & Norms

Honestly, I'm not sure exactly what action you're looking for us to take.

The value that I see is getting outside views on the communities. When things aren't working, we bring them up on a site's Meta site, on Meta Stack Exchange, or to staff via the appropriate channels. Maybe things are different in some communities, but the stuff that isn't working should be out there.

We've been raising concerns about decreasing activity (questions, answers, views, visits, etc.) for a very long time now. I don't think we've gotten the help or support that we need to address these concerns. Maybe this effort can be a start, but I think it requires staff to reach out to the communities with what they're seeing to start the conversation. We've tried to have these conversations, but they don't seem to have gone anywhere.

I'm hoping that you do reach out to one (or more) of the communities that I'm involved with, as a user or as a moderator.

I'm just spit balling this out there right now (this'll likely grow in the future when I'm at a desk), but something I'm supportive of, and I've heard others are, would be migration between sites should be permissible on old and good quality content that's (no longer) on-topic for a site.

You discuss closure being a problem, especially for old content that is getting reviewed, but no longer on topic. Rather than changing what is allowed, why not allow the community to suggest migration? This could be something moderators take on board for actioning for old posts (but the community can flag for the migration), from both sites, but migration to a different site where something is on-topic, rather than closed and can't garner further answers, seems like a win to me.

As example I closed this question as part of my normal usage of the site (I had searched for a problem I was encountering). The post is old, and clearly off-topic for Stack Overflow but I suspect would be a good candidate for Superuser.

I admit, this could be a slow process, but still worthwhile if well implemented. It would, of course, need support of the mods, as bad migrations are also a poor experience for both sites.

I'd probably say, we should do duplicates better. And rather than duplicating a previous post, I'll link it here

While some aspects are mechanical, I'd say being less zealous about duplicates and putting anything vaguely similar in the same bucket - and looking at the helpfulness and quality of new posts as well would be useful.

I think there's a cultural quote from a post from Jeff in the previous post

That's why we actually don't mind having several versions of every question, where there are variations in wording or circumstances. The more chance that someone types a question into Google and finds their exact question already answered, the better a job we've done.

while part of the solution is mechanics, I'd say we can fix/change our culture a bit about duplicates.

This addresses the following dot point:

• Rep thresholds for commenting or other privileges: would a lower/custom threshold better serve your community?

I've gone back and forth on posting this, and majorly trimmed it down, because I get this is kinda tangential to the "policy" discussion but I feel it needs to be said:

Privileges need a rethink/retooling, not just threshold tweaks

Assuming the goal here is to make it easier for people to participate, I believe that simply dropping privilege reputation levels to entice new users is a band-aid. The privilege levels exist for a reason, and that was to keep the more risky tools out of the hands of users who either

1. Don't know what they're doing, or
2. Are not interested in participating in a constructive/positive manner (spammers, R/A, etc)

To give an example of what I mean by a "rethink", let's look at one privilege: commenting, normally locked until 50 rep. In comparison, 1 rep users can post new questions or answers. So why the difference? Because asking and answering is highly visible. Any potential spam or other nonsense post is going to be seen by community members pretty quickly. Comparatively, comments aren't super visible, they do not bump the post, they could go days, weeks, or years before being discovered.

So "just" dropping commenting privileges to 1 rep might be beneficial from the "enticing new users" perspective, but it comes with some risks involved, too.

So how could we solve this? Well, one potential solution would be to make them more visible. What about some sort of temporary elevation system/peer review?

1. A 1-rep user sees a post and wants to comment on it. That comment isn't posted immediately, instead it's held for review.
2. A user in a review queue sees the comment, and can approve, deny, or mark as spam/rude. (this could require two or more reviewers too, just an example)
3. Assuming approval, the comment is then posted proper on the site.
4. That user is now granted comment permission on that post for let's say 48 hours, to allow for follow-up comments.
5. When the 48 hours expire, the user's next comment must also be reviewed.
6. Once they reach the regular reputation level to comment everywhere, they can comment without restriction as per usual.

Granted, this is not perfect either, it's just an example of how you could make it easier for new users while also maintaining the quality control around commenting.

Its this sort of lateral thinking we need, not simply throwing open the gates.

On the Electrical Engineering site, which is currently the only site have got enough reputation to be involved in close / reopen voting I have seen friction around what is considered on-topic.

Taking the recent example of Why did Patriot missile battery's tracking performance significantly decrease few hours after each reboot before 90s?:

1. When I first saw the question, left a comment explaining why thought was off-topic, but took no other action.
2. The question got two well-researched answers with positive scores.
3. 3 site members voted to close as off-topic.
4. A moderator re-opened the question, leaving a comment about why they thought the question deserved to remain open.
5. The question became a hot network question.
6. The question got a further well-researched answer which got 12 upvotes.
7. The question has now re-appeared in the close vote queue, with two close votes so far. I selected "skip" in the review as didn't have a strong opinion either way.
8. There is a comment on the question suggesting two possible more appropriate sites for migration to, neither of which are on the "What site does this question belong on?" list available to non-moderators on the close dialog.

I don't yet have any good ideas for how to resolve the above, just noting an issue.

On our end, getting more people to use chat and meta would be nice. Meta lacks visibility. Chat's better than it was, but a lot of it is quiet. I'd like to move conversations back to chat, and see per site metas used more effectively as it was for reopen requests, policy clarifications etc.

The how is tricky. Meta lacking visibility needs a design change and we need to encourage folks to be active on meta, both as a way to get people to review posts and have it seen as an effective way to be heard.

Chat's taken the first steps there - but we need to figure out how to get it back to being a third space. While a good chunk of my chats are not dead we need to incentivise mods and site regulars to use chat in various ways. As a mod, I try to keep the chat going, but between chat users moving elsewhere due to various reasons, and lack of interesting things I am working on, I'm not entirely sure what I can do. Collectively though, we might have better ideas ;)

(This is an expansion on, and based off the equivalent post on the mod stack internal instance)

We’d equally be entirely open to the answer being "actually, our real challenge is Y, and that's what we'd rather focus on"

I'll be taking you up on that :p

(This isn't reflective of what the entire code golf community thinks, but I feel like it might be a fair representation of what might be happening)

Over on Code Golf and Coding Challenges, we've got a pretty good sense of policy - the whole premise of the site kind of relies on having policy to keep things (relatively) fair and to prevent endless rules lawyering. For example, we have a list of loopholes that are banned, a set of allowed methods of receiving input and giving output, and of course our requirement that questions have an objective winning criteria. And that does not even mention the many "case law" type policies (examples: 1, 2, 3, 4). In fact, we even make policy to override older policy (example: old policy, new policy).

If anything, policy is exactly the opposite of what our site should be focusing on; historically, there's been attempts at trying to slow down policy, and points raised that what's more important is giving new (and existing) users a helpful "welcome" style resource. Additionally, when we've tried to update some of our own policy of our own volition, we ended up deciding not to! Needless to say, I think our site has a pretty good handle on scope, culture, policy, etc, and that focusing on updating policy and culture is not a solution for our own declining activity.

Our real challenge is that a lot of people have moved on from Code Golf without new people to replace them. There was a major activity spike around 2021 that gradually fizzled out. Admittedly, some of the activity decline was partly our own fault, but I believe that the greater cause of decline is that there's not enough people visiting the rest of the network to "trickle down" into our site.

Now, it isn't lost on me that Stack Overflow launched Challenges, and that Code Golf (at least on paper) would have been a "better" candidate for the company to focus upon. While it is true that Challenges may be keeping people from Code Golf, I (still) don't think that blindly directing users to our site would have been beneficial - not when there are so many rules and site-specific norms to learn and follow, and not when our site culture requires more than just the "come and have a fun time with popularity contest" vibes of Challenges. In fact, a concern I've always had about Challenges is that its style of questions would not be well recieved on Code Golf. Treating Code Golf as the main site for programming puzzles without careful community input, as well as consideration of our existing policies, is probably not the best idea.

Essentially, in my personal opinion, we'd rather focus on getting more people on the site in a way that does not lead to our extensive catalogue of policy making the experience worse for both new and existing users. I don't have an idea of what that looks like at this stage, but hey, could be worth "reaching out to [our] community through [our] Meta site" :p. (Coming by and saying hi in our chatroom (like y'all previously have) is also something that could be good.)

There are two cultural norm-related items that form a negative feedback loop:

• The community's expectation that newcomers understand our norms, quirks, and idioms before posting, typically by reading a bunch of help center articles and meta posts.
• Newcomers and beginners don't want to read a bunch of help center articles and meta posts to understand our norms before posting.

It's hard to break into a new crowd when you're the new person. You simply need time to be a wallflower and observe your new crowd before joining, and upon engaging in your first interaction, the new crowd needs to be patient and accepting of the newcomer who is still learning the ropes. Too often newcomers and beginners are met with less-than-welcoming down-votes, closed and deleted questions, and links to meta and help center articles. This just drives people away.

There is a difference in why people come to Stack Exchange sites — at least the technology-related ones (maybe just the programming-related ones).

Based on a myriad open-ended questions at Stack Overflow, I see question askers coming here for these reasons:

• They want to connect with fellow professionals.
• They want to learn the ropes from professionals instead of professors.
• They are completely lost because they don't know the buzzwords to search for.
• They are curious to know how professionals got their careers bootstrapped.
• They need help with design or problem solving.

Moreover, they come here to get help.

The community is looking for interesting and well-thought-out, well-researched questions from other professionals. A majority of people aren't rude, but moderation happens fast enough that content curation gets misinterpreted as "the new crowd doesn't like me, so I'll go someplace else."

The four stages of Stack Overflow disillusionment:

1. You start as an innocent pink square with a question

2. You naively decide "let's ask Stack Overflow!" (still smiling, poor thing)

3. Your question gets flagged as "DUPLICATE OF SLIGHTLY RELATED QUESTION FROM 2006" that uses deprecated libraries and doesn't actually solve your problem

4. You return to being a square, but with PTSD and a solemn vow: "NEVER AGAIN."

And that's how developers learn to debug by staring at their code for 8 hours instead of asking for help!

(source)

(and we'll also pretend that Stack Overflow was around in 2006 instead of 2008)

So, we have content curators who want to build a wiki of programming problems, and a population of question askers who want help (without building a wiki), and we have a company that owns these internet properties who wants to drive engagement as hard as they can, insisting that these two populations live together peaceably.

This disconnect in goals and incentives is the problem to solve.

Gimby posted a comment on one of my SO meta questions, which finally breathed life into these feelings after many years:

[...] there is an us (meta), and there are two thems; one them is the general populace [the company wants] to attract to the site at all costs, and the other them is the company itself. A rather conflicting triforce.

How do we as a community accommodate people who want help, but lack the knowledge of our cultural norms, and lack the technical knowledge to phrase a question that grizzled professionals want to read and answer in the 5 to 10 minutes in between meetings, builds, random questions from coworkers who should know better (but they know enough to not ask on Stack Overflow), and whatnot?

I would be interested in what changes we can make to policies and norms to address this in the Software Engineering and Stack Overflow communities, but I'm sure this applies to more.

I don't see any evidence that AI assist on SO's home page is in any way benefiting the SO community. Is there any data that actually shows otherwise? How often are the "useful" answers it provides sourced and cited directly from SO? How often does interacting with AI assist result in further interactions within the community, such as asking questions, commenting, and answering questions?

So far this has just been a black box with more or less promises that it's doing well and a boon for the community... but I don't think we've ever actually seen any public evidence of that.

Why bother, you never have before. Normal practice is just to force half-baked agenda-driven irrelevant rubbish on everyone, and then wonder why no one wants to engage with the sites and you have alienated 90% of your core users. And then praise yourselves while blaming outside unavoidable influences such as sunspots or high tides while doubling down on the rubbish.

A mild suggestion would be to treat your users & mods as valuable volunteers to be looked after, not low ranking third World employees who can be bullied into whatever you want. If you look after people they tend to reciprocate, if you don't, then in the immortal words of D.R. & QUINCH "Eat Plutonium Death, You Disgusting Alien Weirdos!"

No offense intended.

When moderators never leave, progress never arrives.

I've long argued that one of the most frustrating aspects of Stack Exchange is its system of moderation. Users progress from inexperienced user > experienced user > high-rep user, with clear benefits at each stage in terms of access to new menus, tools, and privileges.

As you become more senior, you become aware that moderators are working behind the scenes, but you’re effectively blocked from becoming one yourself because elections are held so infrequently. High-rep users often become less active after a few years of intense participation, and we're failing to capture them as site leaders during this period of peak engagement. It’s possible — indeed, likely — that a user could go through their entire 'active period' without ever seeing a single election that they could realistically win.

Fresh moderators are essential for engaging with the community, encouraging the development of new policy and topical direction, and generally shaking up the status quo. By contrast, a permanent overclass of long-standing “moderators for life” who may have been elected when the site was in its infancy can end up acting as a brake on the development of new policy, or effectively exercise a ‘pocket veto’ on changes they personally dislike simply by ignoring requests that things be looked at or altered.

My suggestion is that we should have a system of regular elections (say yearly) and long-standing moderators should be encouraged to step aside when they've become less active in the meta-community (say, when they've not posted or answered a single question on Meta for a six months).