Inviting more thoughts: Do these challenges resonate beyond Stack Overflow? (Closed but potentially useful posts)

Question

A few weeks ago, I started a conversation on MSO focused on a topic: A discussion about closed (and potentially useful) posts on Stack Overflow

Specifically, we were talking about "grey area" questions that could potentially be helpful, especially the ones that are more experience-based, exploratory, or practical. These types of questions often don’t fit neatly within the rigid "rules" of traditional SO Q&A, leading to closures despite their potential value. This sparked a broader discussion around how questions are categorized and moderated, the benefits and drawbacks of boundaries between communities, and how to balance strict rules that enable high-quality content in Q&A, with the flexibility needed for valuable questions that don't fit today, while also maintaining quality. It’s a tough challenge to work through!

Last week, I shared a summary of what we learned on MSO to invite further input, but it’s clear these challenges aren’t unique to SO, they likely impact other sites across the network too. That’s why I’m sharing a companion post here on MSE to ask: How do these (and other) challenges show up across the broader network?

Why are we asking now?

We are thinking long-term about the future of the network. As many here know, participation patterns across the network are shifting, more people are relying on AI tools to get answers, and there are ongoing questions about the role humans play in sharing knowledge and what kind of community Stack Overflow should be in the future.

This post is an effort to start a broader conversation about "where to from here?". To be clear: this conversation is not about lowering quality standards or simply allowing more low-quality posts. Instead, we want to understand the tensions that arise when valuable content doesn’t fit existing rules, and how different communities handle that balance between clarity, flexibility, and quality. Once we have a solid understanding of the challenges, we'd also like to explore ideas for how we can solve these challenges.

A few themes from the MSO discussion

To kick things off, here are some of the key areas of tension that came up in the MSO discussion. These were points where people expressed different views on the need and the best path forward, sometimes disagreeing, but also grappling with similar challenges from different angles. Do these sound familiar in your community, or would you frame them differently?

• On the site’s scope (What questions are allowed?): Some view Stack Overflow as a technical Q&A site where precision, strict rules, and reliability make it valuable. Others feel it needs to adapt to stay relevant, embracing a broader scope, such as questions about real-world developer work, including architecture trade-offs, tooling choices, or experience-based insights, even if these topics extend beyond "just code."

• On the site’s purpose (What’s the goal?): Some believe the site should focus only on capturing permanent, reusable knowledge, with anything else considered noise. Others want space for exploration and real-world problem-solving, where questions may not always result in timeless, polished posts.

• On policies and flexibility (How strict or flexible are the rules?): Most agree that moderation needs to be predictable, fair, and understandable. However, there are differences in opinion about how best to achieve this. Some want strict policies with little room for interpretation, while others argue for more flexibility, allowing human judgment to manage edge cases and grey areas.

• On boundaries between sub-communities (How separate should the sites be?): Some value clear boundaries between Stack Overflow and other specialized SE sites, seeing this as essential for maintaining focus and expertise. Others believe that these boundaries may no longer align with how real-world technical work flows, suggesting that more fluidity could help foster cross-topic collaboration. How does this issue play out across the network?

• On "High Quality": What Does It Mean?: There’s broad agreement that clearly low-quality posts (vague, no effort, or off-topic) need to be handled quickly, as too many can overwhelm the site. But there's less agreement on what counts as high quality, especially when it comes to questions based on personal experience or exploration. While technical accuracy is important, some posts might not be perfect but can still be useful. We don't have a clear, consistent way to assess these types of posts.

Here are the key challenges I posed in MSO that I think are most relevant for the broader network, and I’d love to hear if they resonate with you (or if you’d frame them differently):

• How can we maintain high standards of quality and reliability, addressing the flood of low-quality questions, and can we also make the site approachable for experts and non-experts alike?

• How can we accommodate real-world, messy questions, potentially evolving them into Q&A as some suggest, without losing Stack Overflow’s clarity?

• What does "quality" mean, especially for content that doesn't neatly fit in Q&A today, and how can we define it to include meaningful but messy content without inviting low-effort posts?

• How should we think about the boundaries between different spaces and sub-communities on the network? How can we preserve the benefits of focused, specialized communities and a strong sense of belonging, while also making it easier to collaborate and share knowledge across the broader Stack Exchange network?

• How can Stack Overflow maintain a high-quality, impersonal knowledge base for users who prioritize clear, objective answers, while still accommodating the community dynamics that some feel are necessary for the platform’s success?

(Credit to several users on MSO who gave feedback and suggested alternative framing on the above)

We’d love to hear your perspective on these challenges!

Please share your thoughts in an answer below, especially if your experience is different or if you’d reframe these challenges in a way that makes more sense for your community. For now, I'd like to keep the focus on whether these are the right questions to be asking. The next step will be to think together about possible solutions.

For MSO-specific concerns: If you want to keep discussing how these challenges affect Stack Overflow in particular, please continue that conversation over there.

This particular post is focused on network-wide perspectives.

Feel free to bring this discussion to your own site’s meta to gather more local insights too!

• Richard: In ye olde days of yore, SO and SE were successful because people came for the programming, and stayed for the other sites.
• tkruse: A broad survey asking visitors why they come here, what they like here and what they miss here would seem like a useful resource for auch questions.

Answer 1

The responses to this question are indirect signals about the fundamental challenge involved in reshaping these sites.

The business model of SE is that 'experts' will be willing to put in volunteer labor in return for 'magic internet points' (or more likely, personal satisfaction) in order to serve 'non-experts' (scare quotes explained below).

Everything that has happened in the last few years seems to indicate that the technical expert population has some strong feelings about what type of community service they are willing to provide. Roughly, they (we) are most content to help out people who achieved a certain minimal facility with general technology they are working with, and who demonstrated that they are putting in a reasonable level of effort to find their own solutions before coming here. In other words, the questions that do best come from people who are not so different from the answering 'experts' -- they just lack information on some specific problem.

Further, the typical expert is not prepared to start from a very broad set of requirements and write up a comprehensive design, so they (we) prefer questions with focus.

Thus questions in some categories get fairly rough treatment, whether entirely deserved or not:

• students (apparently) dumping raw homework.
• 'script kiddies' who expect to be able to get work done by asking ChatGPT how to do something, and then come here and expect humans to sweep up after the incontinent elephant.
• newcomers to the field who have the best of intentions but truly have no idea where to start.

In addition, a significant population of experts are put off by 'bike shedding' and general pontification, and so dislike questions that solicit them.

On the demand side, this absolutely leaves many eyeballs unsatisfied, frustrated, and sometimes offended. These eyeballs are, also, potential sources of ad revenue. Thus, it's not surprising the company would like to discover some way to extend the range of discourse to serve, at least, some of them.

All questions of principle aside, this will sink or swim on a simple question: are there (enough) experts who are prepared to serve? And why should they leave reddit and come here to do what they can already do there?

History does not lead me to see a positive answer here. However, what do I know? All I want to write is that I'd advise the company to focus on experiments that probe this question without making changes to the existing sites that seem very likely to drive the experts away.

• EmmaBee: Really appreciate this comment, its highlighting an important core tension. I think you're right that a lot of the discomfort comes down to what kinds of help people are actually motivated to give. It's maybe not there aren’t enough people with expertise, but as you say, many have strong preferences around what kinds of questions feel worth engaging with. It makes me think a path forward isn’t about trying to convince the same folks to support a broader range of questions, but maybe asking: Who does enjoy the earlier-stage, or more exploratory kinds of help, and how can we make space for them?
• EmmaBee: This is also reminding me of a recent example, someone shared in chat about a long time user who spends a lot of time in a Physics chatroom helping people work through problems that aren’t a good fit for Q&A. It is interesting that even though this isn’t fully supported, it happens informally. It’s a small sample size, of course, but it makes me wonder: Are there more people who enjoy working through problems like this, and could there be better ways to support these kinds of interactions? This ties back to the concern you raised about the balance between helpers and helpees being important.
• NoDataDumpNoContribution: If it's not a strictly binary sink or swim thing, one could try to improve conditions for everyone (with some balancing) as much as possible in order to get the most out of the interaction here for everyone. Experts want to help. There are people needing help. The company wants to make it work as best as possible.
• Lundin: There's many aspects of this but in the end all users of tech sites will go to where the domain experts hang out, no matter the site culture or design. This has always been the case historically for as long as the Internet has existed. Nobody wants to ask their technical questions to fellow students ("blind leading the blind") if they can get an answer from a domain expert who's worked 10+ years with that specific technology in the field.
• ColleenV: The network isn't just about helping people by answering their questions though. It used to be about building a lasting library of knowledge for the public. I don't care who is asking the question so long as it's a good question for the library, I don't care if the asker was just getting help with their homework. Of course that is frustrating for people who want a tutor, not an entry in a library. The people asking would need to buy into the mission that the community was formed for. The tension is almost entirely caused by the lack of money to support building a library which necessitates ads
• Rosinante: I think that my point here is that there will be no library of content unless experts answer questions. Also, when I look back on my 10-year-old answers, most have been obsoleted by the passage of time and tech.
• kirogasa: >"why should they leave reddit" \n Because there's terrible UI (that's even I use it, use old.reddit); because there users can delete their posts spontaneously, causing a knowledge sharing to be lost.
• EmmaBee: So, I am hearing that what the library REALLY needs is a cafe or a space for up and coming writers, then a space people people to work through messy book drafts, which could end up on a library or maybe they stay in the pile of messy drafts, and then of course the library itself. Did I solve it? :)
• Greg Burghardt: @EmmaBee: "Who does enjoy the earlier-stage, or more exploratory kinds of help, and how can we make space for them?" --- I do! I love teaching people at any stage, especially the early stages! And I think this answer hits a bullseye from a mile away. There is a core tension between the experts that only want to deal with experts, and people like me who really enjoy working with beginners.
• Rosinante: @EmmaBee The term that I didn't mention, but which was a commonplace in ancient discussions of all this, was 'help vampire.' If you create a cafe, you have to figure out how to keep people like Greg Burghardt from being drained by those.
• Matthieu M.: @EmmaBee: This reminds me of one of the top answerers in the C++ tag, who would answer questions that had already been asked dozens or hundreds of times, and that they had personally already answered dozens of times, instead of closing them as duplicate with their gold hammer. They were just happy answering questions, and didn't care for the "meta crowd" and their rules, or the "mission" of SO about building a curated high-quality Q&A. They were just happy answering questions, even simple ones.
• Karl Knechtel: "Everything that has happened in the last few years seems to indicate that the technical expert population has some strong feelings about what type of community service they are willing to provide." - well, the feelings go back decades.
• Karl Knechtel: @EmmaBee The "library" metaphor is very strained, because libraries don't create or publish books. (I don't think it's ever been the right word for the tour to use, but it's hard to find anything better.) But I do think you're on the right track, yes.
• Karl Knechtel: @MatthieuM. yes; but that's harmful to said mission, and also can be done in countless other places already anyway.
• Matthieu M.: @KarlKnechtel: Is it harmful? That's a hard thing to judge. I personally know of no data which measured said harm (or benefit). Perhaps the way SO went about building its Q&A was wrong -- who knows! -- and instead of insisting that each and every question be a gem unto itself, it should have instead embraced the idea of letting all questions in, and letting curators promoting gems out of them. It would have been more welcoming to new users, and there still would have been a curated Q&A emerging out of it. Would it have worked better? I don't know. I doubt we'll ever know.

Answer 2

We are thinking long-term about the future of the network. As many here know, participation patterns across the network are shifting, more people are relying on AI tools to get answers, and there are ongoing questions about the role humans play in sharing knowledge and what kind of community Stack Overflow should be in the future.

For smaller sites, we've often had issues even pre-AI, which weren't really addressed – quite a few were gutted due to unaddressed drama, or never really got the sort of site promotion "early" sites got. It might be worth taking a little focus off "AI as a threat" to "How have historic actions affected community involvement?" Things like staff turnover, the company/community relationships and other events probably have had as large an impact as AI on many communities, if not larger.

I'd also point out one of the impacts of the current AI craze and the post-Covid era is massive downsizing in tech companies. A developer-centric community will lose users if there's less devs – especially if they don't also code in their free time.

Essentially if you want to think long-term, you need to look broader than AI – and both into internal factors and the knock-on effects of companies trying to follow trends blindly.

It's also useful to consider not just raw numbers for engagement. We probably want people who want to be here more than people who want here to be something very different, like a chat or forum. I personally think losing "core" members here contributed more to the decline than AI – but that also means looking back at the last decade and a bit more, and honestly evaluating where things went wrong.

Quite a lot of the scope and quality guidelines are set by the community – for example, I'd personally preferred a slightly laxer view towards product recommendations than Super User does, but the compromise the community worked out – that you can talk about how to use a product to solve a problem – is a "better" solution overall, but a bit esoteric for a newcomer. Early on, I felt there was a lot more enthusiasm and energy towards onboarding new users than there is now, perhaps a little ironically.

If you want to change these things – while Jeff was able to benevolently dictate certain changes, the modern-day reality is that the trust and authority to do that unilaterally no longer exists, and you need to convince the community to do it.

On boundaries between sub-communities (How separate should the sites be?):

I moderate a site which is a superset of others – Super User's scope fairly cleanly covers Ask Ubuntu and Unix and Linux, and between Super User and Server Fault, the two sites and some other sites are covered.

However – Ask Ubuntu has its own, active and thriving community. Unix and Linux has a somewhat more technical bent than SU and might be a better place. For new communities, it might make sense to find a existing community – we didn't really need as many cryptocurrency sites as we do for example, but as much as scope, culture is important.

On "High Quality": What Does It Mean?:

That's a hard question. In theory, we want useful posts and interesting posts. Quality isn't complexity. A good question is clear, provides as much information as needed, and is practically answerable. A good answer is reusable, clear, and teaches things. Useful posts are a backbone, while interesting posts attract experts. We need both.

• EmmaBee: Thanks for the honest take, especially about looking beyond AI. I agree that the issues you mentioned have all played a role in where we are today, with AI just feeling urgent because of the sharper drop in contributions over the last couple of years. One of the goals of these posts is to try to paint a fuller picture by inviting people like you to share different views, because, you're right, the challenges we’re facing aren’t the result of a single cause. Still, if you had to pick ONE big lesson from SO/SE’s history to carry forward as we think about the future, what would it be? And why?
• EmmaBee: Also, really appreciate the way you framed the idea of quality, splitting it into "useful" and "interesting" feels like a helpful lens AND there's likely a lot of subjectivity in interpretation. What’s useful to one person might not be to another, and "interesting" really depends on who’s reading. Some comments on my MSO post got me thinking the idea of "quality." It’s a word we use a lot, but it feels like we don’t have a shared, concrete definition. How important do you think it is for us to have a more shared definition of quality? Or do you think the ambiguity actually serves a purpose?
• Journeyman Geek: Well, one might argue an overstrict definition of 'quality' is precisely what might turn off new users. I can write a useful, single line answer that's good enough, or write a long, well researched post that... is too much. Quality is very much subjective. I try to write the best posts I can (and get embarassed by old ones) but to me - interestingness is a short term quality for attracting readers here, and usefulness is about building things for the long term. That said, getting too wrapped up in quality can be a problem too - I've potential questions that's been backburnered cause
• Journeyman Geek: I wanted to get it perfect - both here, and on SU. My basic level of quality though is "give people what they need to answer/use the post without needing to refer elsewhere", which feels like a low bar - and one that can be workshopped.
• Journeyman Geek: The One lesson would be a long, very controversial post - but I would say trusting and respecting the community would be an executive summary. That said, right now, I've both andrewwegner.com/… and aprilwensel.medium.com/suffering-on-stack-overflow-c46414a34a52 open right now. I don't agree entirely with one of these (Even if its an interesting perspectives) but reading them side by side is very educational. They perfectly encapsulate two different visions for what we need to be, and two different perspectives.
• Journeyman Geek: I'd also consider investment in the platform from the authors, and what their goals are, in reading them, and this particularly interesting analysis by Shog meta.stackoverflow.com/a/320234/674571 linked in the former.
• Jeff Atwood: "while Jeff was able to benevolently dictate certain changes, the modern day reality is the trust and authority to do that unilaterally no longer exists and you need to convince the community to do it" imagine if that changed 🤔
• Journeyman Geek: Maybe some day, but that needs something special , or a lot of work and patience. Maybe both. I do imagine it can, but that needs a very different way of thinking than what came after you and possibly Joel.
• Lundin: It should be noted that the 2019 sXXtstorm happened during the Joel era. Prashanth was brand new as CEO and still way too fresh to reasonably had a chance to fix any of that. It's much too easy to blame the current management and Prosus for "all the bad things" and nostalgically reminiscence about the "good old days", when in fact the network went to sXXX earlier than that.
• Journeyman Geek: Yup, I'm aware - things started going bad towards the tail end of Joel's CEOship but from what I heard there were other factors and people involved too. I would rather blame attitudes and actions than people specifically - I've my own (semi-conspiracy) theories about when and why things went wrong but personally making things go right matter more
• Karl Knechtel: "It might be worth taking a little focus off "AI as a threat" to "How have historic actions affected community involvement?"" - I would go further: worrying about the fact that AI can now answer personalized questions is missing the point. And if it answers a generic question, what it says will represent (or plagiarize) existing material - including from SE.
• EmmaBee: @JeffAtwood thanks for popping into this post! :) I would love to encourage you to share any thoughts in an answer on this post, if you feel so inclined. I'm thinking with respect to historical challenges that might be relevant to today’s discussions, but really, anything you feel inclined to weigh in on. I think there’s a lot we can learn from you about where we started, how things evolved over time, and where things might be heading in the future.
• Jeff Atwood: @EmmaBee like happiness, power is only real when shared blog.codinghorror.com/10-years-of-coding-horror
• EmmaBee: @JeffAtwood Hmmm looks like its about that time for the "21 Years of Coding Horror" installment on the blog? :)
• Jeff Atwood: @EmmaBee you literally read my mind. Not kidding. I just yesterday thought, it's time for that. Now I read this 🙇💛

Answer 3

Half-finished, random showerthoughts:

The posts mentions 'messy questions' a few times. Messy to me sounds like questions that aren't properly refined yet but which hold potential. All sites already have a process for them: Closing, commenting, optionally meta posting, editing, reopening. I know users on the 'receiving end' of that process that have their question closed aren't that happy about the experience of it, but I don't think the premise of the process itself is necessarily wrong: A messy question needs un-messy-ing first. I have not kept up with it, but the staging ground sounded like a promising way to un-messy questions.

What does "quality" mean, especially for content that doesn't neatly fit in Q&A today, and how can we define it to include meaningful but messy content without inviting low-effort posts?

Low-effort does not equal low-quality! There's quite a few posts that are probably included in the 'low-effort' definition: Questions where the question title and body are the same single sentence, or the body is no more than 5 lines of text long to me sound like they never take much effor to write (unless the writer usually struggles with being too verbose and made a real effort to be as short as possible...). And I think these types of questions are scattered across the different network sites. Not all may still be considered 'good quality', but I bet some are. They are just the 'easy pickings' that can be asked on a new site. And sometimes they still make good duplicate targets or stil cover some of the most 'frequently asked questions' about a topic.

As for what quality means, I don't think you'll ever get everyone to agree on a single set of guidelines. It'll always be a case of 'I know it when I see it'. But each community already has guidelines on what quality means, and the 'good' news (when it comes to the questions this post is exploring) is that those guidelines are all different across communities, depending on the 'type' of Q&A a site takes. Sites that take recommendation questions have guidelines on what makes a quality recommendation question. Sites that take identification questions, have guidelines on what a quality identification question needs, etc.

And that's where I get back to grumping about 'what's wrong with the existing process': Because if a community thinks a category of questions can exist on a site as good quality questions, they can already argue for this on site meta's and propose guidelines for what makes it good or not good quality. That seems a perfectly adequate way to determine what a community considers to be 'quality' to me.

How should we think about the boundaries between different spaces and sub-communities on the network? How can we preserve the benefits of focused, specialized communities and a strong sense of belonging, while also making it easier to collaborate and share knowledge across the broader Stack Exchange network?

As for boundaries and communities: One of my pet peeves is seeing site $B recommended as a 'good place' for a question that's off-topic on site $A, by users that have little to no activity on site $B. This often leads to users just copy-cross-posting their post, and now having it closed in 2 places.

How can Stack Overflow maintain a high-quality, impersonal knowledge base for users who prioritize clear, objective answers, while still accommodating the community dynamics that some feel are necessary for the platform’s success?

Keep those two in (mostly) different places, like they already are? High-quality, impersonal knowledge goes in questions and answers. No chit-chat needed there. Community dynamics already take place in comments, on meta (fun posts, topic challenges), and/or chat.

• EmmaBee: I appreciate your half-finished, random showerthoughts. :) That line you wrote hits at the heart of this debate: "the premise of the process itself isn’t necessarily wrong." I agree the intent behind closing/editing/reopening is sound, but I’m curious how we might better handle the experience side for people on the receiving end. What do you think makes that process feel so jarring for askers, even when it’s working as designed? I wonder if part of what we’re grappling with is less about changing the process and more about rethinking how it’s experienced. Curious what you think!
• Tinkeringbell: @EmmaBee I'm probably not the right person to ask about the receiving end. I've not been on it often and I've been around so long, I know the receiving end doesn't equal the end of the world/question. I can guess what makes it so jarring is probably some wrong expectations coming onto these sites, but how to change those? No clue... Mine were set by my programming teacher, who just said SO had all the answers already, and that the stuff we'd be doing in class wouldn't need to be asked as a question again because it would end up closed. Start with teachers? :P
• Karl Knechtel: "Not all may still be considered 'good quality', but I bet some are." Many of them absolutely are. Sometimes a single-sentence "How do I X?" requires surprising effort in figuring out a succinct yet accurate phrasing of X. And many of the all-time top-voted Stack Overflow questions are like this (even if you filter out the questions about using Git).
• Karl Knechtel: As for "setting expectations", the best that can be done from this end - probably - is for staff to change the material on the front page, in the tour etc. to refocus: look, we have a lot of really neat Q&A; in this format, you can see the Q clearly defined at the top, so you only have to keep reading if you found the right one; then the community already chose the top As for you. And if you don't find it, then you can ask it yourself - with a focus on trying to make that experience smooth for the next person.

Answer 4

From the question:

• On the site’s scope (What questions are allowed?): Some view Stack Overflow as a technical Q&A site where precision, strict rules, and reliability make it valuable. Others feel it needs to adapt to stay relevant, embracing a broader scope, such as questions about real-world developer work, including architecture trade-offs, tooling choices, or experience-based insights, even if these topics extend beyond "just code."

The scope of Stack Overflow led to the creation of other sites for "programmers" and "software developers". One of the notable cases is programmers.stackexchange.com, which later changed its domain to softwareengineering.stackexchange.com.

Right now, it's possible to migrate questions from one site to another. Also, a "Stack Exchange Network policy" prohibits cross-posting across the network. Proposals have been made to improve the integration between sites, i.e., to make it easier to relate questions and answers about the same problem.

One reason for having multiple sites is to serve different audiences and allow them to define the site's scope and workings. This leads to sites with overlapping scopes and tags with the same names but different usage guidelines. Also, it is hard to understand the differences between sites' workings.

You should be prepared to handle the effects of the changes in Stack Overflow on the other sites.

• EmmaBee: Thanks for chiming in! I’m familiar with why the boundaries between sites exist - they were meant to help, but in practice they can also introduce friction. A big part of the debate we had on MSO has been that these boundaries aren’t always clear, especially for grey area questions that might not clearly fit in one space or another, and curators might not have the SME to migrate efficiently, which leads to closed questions. Even with awareness of other sites, it’s not always obvious where some questions belong. I’m curious if you see that as a broader network issue or more SO specific?
• Rubén: @EmmaBee That might look like the chicken-egg question about what was first, but in this case, it's clear that SO is the first and flagship site with the most significant participation and the largest number of questions and answers. I hope you learn sooner rather than later that several sites were explicitly created to house IT questions off-topic on SO. Other sites were made to house off-topic questions on Super User, like Web Applications, Ask Different, Android Enthusiasts, etc. However, the problem with migration and integration is not SO exclusive, it's present network-wide
• EmmaBee: Appreciate the helpful historical context about the sites that spun off! Since you mentioned this is a broader network issue, I’d be curious: are there particular examples or types of questions or communities where you’ve seen this show up and create friction on SE? Like unclear boundaries, difficult migrations, or questions that were interesting but didn’t quite fit anywhere? Curious if anything comes to mind.
• Rubén: @EmmaBee One example is spreadsheets. They are on topic on several sites, but not all questions about spreadsheets are on-topic. For example, questions about spreadsheet formulas with certain limitations are on-topic in SO, Super User (SU), Web Applications (WA), among others. Spreadsheet formulas are on topic in SO and SU. If they are about using formulas on a web application, then they are also on-topic in WA. If they are about using formulas in Numbers, they are also on-topic in Ask Different but not on WA. However, questions about using Excel in general are off-topic in SO
• Rubén: If a question is about using a spreadsheet formula, usually referred to as an Excel formula, posted in the context of Google Sheets on SU, it might get closed or migrated to WA but might also be migrated to SO. However, if the question doesn't fit the rules of the WA / SO it will be returned to SU but it will stay closed. Also, closing questions as duplicate causes friction because for some people it's clear that the answers of the "original" question answer the closed question, but for some users it's not clear, like closing a Google Sheets question as a duplicate of an Excel question
• Rubén: Despite the efforts to explain the similarities and differences of spreadsheet formulas in different contexts, some users "blindly" close a question because of the presence of a term like "Google Sheets" or "Excel" and if they aren't closed, then those questions that are kept open are used to "justify" that another question should be reopened even when the case is not the same.
• EmmaBee: This is such a great example that highlights the issue perfectly! Especially when it’s unclear which site a question belongs on or when it's considered a duplicate. I’m curious, what do you think is at the root of the confusion? Do you think this issue is more about the way we categorize questions or more about how askers interpret those categories? Something else?
• Rubén: @EmmaBee the situation that I described is emergent. It was "created" by the system (the Q/A model, reputation system, the tags, the SE network, the users); without the system, this situation would not exist. In other words, yes, the issue is about how we categorize questions, how askers and curators interpret those categories, and how the migration works and its limitations.

Answer 5

• On boundaries between sub-communities (How separate should the sites be?): Some value clear boundaries between Stack Overflow and other specialized SE sites, seeing this as essential for maintaining focus and expertise. Others believe that these boundaries may no longer align with how real-world technical work flows, suggesting that more fluidity could help foster cross-topic collaboration. How does this issue play out across the network?

Newcomers are completely lost, and frankly, even seasoned users are lost. The boundaries are so situational. One recent question got bounced around and down-voted twice.

1. A newcomer asked about how to migrate technologies on StackOverflow: https://stackoverflow.com/q/79619686/3092298 — a user said "The question is too unspecific for SO. Try softwareengineering.stackexchange.com". Of course, the post was downvoted before the author deleted it some time later.
2. Next, this newcomer posted on the SE site: How to properly migrate a large NestJS monolith to microservices with a single API Gateway? — this was met with another down-vote and a comment about cross-posting. Even after deleting the cross-post on SO, the question is still not suitable for Q&A because it asks multiple questions in one. So, strike two for the newcomer.
3. At that point, I suggested they post on StackOverflow Discussions.

I see this so much where questions are bounced between StackOverflow and Software Engineering. Ninety-nine percent of those questions are too broad to begin with and are not suitable for Q&A, and most are posted by newcomers. This was not a pleasant experience for someone new to the network. We have a number of barriers here:

• What is suitable for Q&A?
• What is on-topic for this community?
  • And when people refer the querent to another community, they are mostly wrong about the referral.
• Cross-posting is discouraged, but avoiding cross-posting is an error-prone manual workflow.

How do we direct newcomers in a way that invites them into a community rather than berates them for not knowing the rules? Down-voting is our only tool, and it doesn't feel good. I don't care what anyone says about the intent of downvoting, people feel different about it. Our rules for each community are byzantine, myriad, and inconsistent.

It appears that newcomers have no choice but to flagrantly (if not innocently) violate social norms before they understand where they should post and how. To me, this is the biggest hurdle to getting new members of our communities.

• ColleenV: Migration needs rethinking. It has good intentions: take an off-topic question and get it to a place where it is on-topic, but the implementation is so bad people really have to jump through hoops to make it not a negative experience. I have lots of thoughts about migration, eg english.meta.stackexchange.com/q/11075/80039
• Greg Burghardt: Migration alone won't solve everything, though. There is a social aspect to this whole ecosystem that makes things difficult on members and newcomers. To be fair, this is a hard problem, and it is not a new problem. It's a people problem that's been around as long as there have been people. The perennial challenge is how to use technology to eliminate or reduce those people problems. This is a really, really, really hard problem.
• ColleenV: Yes, there are a lot of connected pieces and a lot of problems that have been ignored. Migration is just one head of the hydra. The trick is getting enough solved to get people re-engaged without focusing on the short-term so much the change isn't sustainable.
• EmmaBee: Thanks for sharing these examples! I really appreciate how you captured a few aspects of the issue. The example of the person asking too many questions it might be both a misunderstanding of norms AND the user's problem might consist of ALL those questions. They could simply be thinking about their issue a bit more holistically vs one question at a time. I'm curious how well do you think the current norms align with how users naturally approach and frame their questions? Or do they sometimes create tension with how users naturally think about their issues?
• Greg Burghardt: @EmmaBee, it's really hard to understand the norms in each community. I don't think you can write a help center article to fix this. And I don't think this particular person did anything wrong. Just as newcomers need a guide, so do us regulars. The challenge is how to be a human, and a volunteer curating content. I feel three things about this: 1) down votes hurt; 2) the community reacts too quickly and harshly when purging low quality content; 3) I really wonder if we are seeing a generational shift in how people want to get help online. (1/2)
• Greg Burghardt: (2/2) I've always really liked how the community reacts in World Building. They get low quality questions but the longtime members still leave helpful comments. Yeah, down votes happen, but there is a personal interaction that is friendly and inviting in nature that helps offset the initial sting of being voted down. I really wish we (as the active users) would extend more patience and understanding in some of our other communities.
• EmmaBee: Love that you’re sharing an example of this working well! It’s a good reminder for me to dig into more of these positive stories we can learn from, instead of just focusing on what’s not working. Maybe we could have some folks from World Building come in and share what they’ve figured out, and see if we can use those learnings more broadly?
• Lundin: These kind of users should have been stopped in the Staging Ground before their post ended up live on the site. Where they might get help by someone who is actually interested in helping them, rather than someone who isn't and just finds the question bad. ->
• Lundin: I've already written several posts before about how hostile interaction and drama is built into the SO system by design, since it uses public shaming as a moderation method. That remains the core problem of the site design and it will never get fixed, since the current user base here is so brainwashed into thinking that the current system makes sense. We'd have to get rid of the voting system and remake the moderation system completely and that probably can't be done without also getting rid of the current user base.
• Lundin: I did some brainstorming about an alternative system over at Codidact some years back: Giving question feedback in private - a moderating system to reduce conflicts. In the end Codidact handles voting and moderation very similar to SE and it inherited the same problems because of it. So it is just yet another SE, rather than a site better SE, which it had the potential to become.
• Greg Burghardt: @EmmaBee - I always forget about the staging ground. It's almost like we need a grace period for questions where we remove the use cases that lead to negative interactions while at the same time communicating to the entire community that the question is a work-in-progress. The staging ground feels like question purgatory with a few gatekeepers because only a few remember to go there.
• Karl Knechtel: @GregBurghardt re "down votes hurt" - one of the greatest strengths of the Staging Ground on Stack Overflow is that voting doesn't start until after the question has gone through a process that, on average, makes it much less downvote-worthy. It's rare in my experience that it makes sense for a question to be voted below zero but not closed; the Staging Ground is effectively a preemptive closure and workshopping. Plus it gets bad questions out of the main feed (the reason we want to downvote them!) and specifically in front of those volunteering to curate rather than answer.
• Karl Knechtel: That is: in principle, the Staging Ground is about as good as a technical solution for "grace period for questions" is ever going to get. It just needs more curators. (Which in turn requires getting people to understand how to curate, and motivating them.)
• EmmaBee: @Lundin As a person who works in UX, I appreciate your comment and agree that a lot of the challenges stem from the design itself. Thanks for raising this! The way moderation and voting are built in definitely shapes how people interact, and that creates some of the issues we see around hostility and community friction. It’s a tough design problem that doesn’t have easy fixes. Are there any design shifts short of a full overhaul that might make a meaningful difference?
• Lundin: @EmmaBee I think a lot of the drama and badwill could be avoided if closed posts are taken offline immediately and moved to an area like the staging ground where only veteran users actually willing to help improving the post are present. And start the post fresh from there, without down votes and comments. The current closing system is designed to maximize drama.

Answer 6

In the list of themes you describe the kind of content that is currently not allowed using a lot of positive words, such as flexible, real-world, experience-based, exploration, collaboration. These words make it sound like a good thing, that you can only be against if you are a rigid rules-for-the-sake-of-rules change-aversive type of person. However, this positive framing is a bit disingenuous, because it distracts from the property that these alternative types of content have in common: they are more open-ended and invite more forum-like posts/threads.

The list in the OP can therefore be rephrased as follows:

• On the site’s scope (What questions are allowed?): Some view Stack Overflow as a technical Q&A site where precision, strict rules, and reliability make it valuable. Others feel it needs to be turned into a forum.

• On the site’s purpose (What’s the goal?): Some believe the site should focus only on capturing permanent, reusable knowledge, with anything else considered noise. Others want to turn it into a forum.

• On policies and flexibility (How strict or flexible are the rules?): Most agree that moderation needs to be predictable, fair, and understandable. However, there are differences in opinion about how best to achieve this. Some want strict policies with little room for interpretation, while others want to turn it into a forum.

• On boundaries between sub-communities (How separate should the sites be?): Some value clear boundaries between Stack Overflow and other specialized SE sites, seeing this as essential for maintaining focus and expertise. Others believe that these boundaries may no longer align with how real-world technical work flows, suggesting that we discuss everything everywhere like a forum.

• On "High Quality": What Does It Mean?: There’s broad agreement that clearly low-quality posts (vague, no effort, or off-topic) need to be handled quickly, as too many can overwhelm the site. But there's less agreement on what counts as high quality, especially when it comes to questions based on personal experience or exploration. While technical accuracy is important, some posts might be a better fit on a forum. We don't have a clear, consistent way to assess these types of posts.

However, Stack Exchange sites are not forums, by design, because that reduces the value of the experience. This argument has been made many times but it is still relevant, see https://meta.stackexchange.com/a/92115/ for a good summary by the former Stack Exchange Director of Community Development Robert Cartaino.

• starball: I don't feel that this is a fair re-framing of the statements in the question post. for one thing, changes to the platform don't need to be changes to Q&A, which I think the wording leaves open, and is consistent with the "three lane highway" idea. granted, I also don't feel like the company is making good on the apparent commitment to keep Q&A the good thing it is and grow engagement through other "lanes".
• Marijn: @starball I see your point about the other types of content going to other lanes like Discussions and Articles (and Chat maybe), and I agree that this does not necessarily need to impact Q&A. However, from reading the company announcements and initiatives in the past, I do get the feeling that staff wants to change the strictness of the Q&A format itself as well (and indeed make it more forum-like), for example with the new comment experiment or any welcoming policies that prioritize quantity over quality.
• zcoop98: I suspect it's on purpose, but complaining about the original framing being disingenuous and then reframing it in an arguably differently disingenuous way is not... compelling. "Forum" is a deeply loaded term as used here– it implies that those who want more discussion or wider scope inherently also want less quality, but I don't think that's a fair assertion.
• zcoop98: "Forum" and "Q&A" are not binary, mutually-exclusive states; there's a whole world of options between them. The design of SE's sites was a purposeful choice to trade away more open-ended content for an emphasis on quality, but it was just that– a choice (and a smart one!), not some reflection of a deeper inherent reality that opinionated content is bad or less valuable overall. The goal was to "optimize for pearls"– the chosen means was barring open-ended interactions. But I don't think it's right to characterize the purist Q&A model as the only possible means to accomplish the goal.
• EmmaBee: Thanks for sharing your thoughts! I appreciate the perspective. When I used words like "flexible, real-world" etc, I wasn’t trying to say those kinds of posts are automatically good or that everyone should want them on Stack Overflow. I just wanted to describe what those types of content look like from some users’ point of view who value this. I agree that these posts tend to be more open-ended and feel more like forum discussions. And I understand why mixing that with the strict Q&A format could cause problems or change what makes Stack Overflow work.
• EmmaBee: I also think it’s important not to assume that all "forum-like" content necessarily means lower quality or less value. Different kinds of interactions can coexist, and there might be ways to support more exploratory or experience-based content without losing the strengths of the core Q&A. I also appreciate the link to Cartaino’s explanation... it’s a good reminder of the original goals. But as a researcher, my job is to highlight complexity and help explore possible paths forward, rather than assume one model fits all.

Answer 7

I think chat is worth exploring as a space for

• people to help each other on questions more casually and interactively, and with less rigid expectations for up-front effort in formulating questions
• the kind of looser discussion or social chit-chat that help people form ties

I think chat as a platform already supports that, but for it to actually happen, we'd need to

1. See if there's room to optimize the tradeoff in how to gate out people we don't want in chat (Ex. people who violate CoC/ToS, spammers, trolls) while lowering the bar so that more people can participate. The current reputation threshold is low, but the vast majority of people don't pass it.

   You can find my swing at that tradeoff-optimization question at Let sites and room owners grant and constrain talk-in-chat privileges for users who don't have enough rep.

2. Promote chat as a place where help and chit-chat can happen and start growing more little communities. Part of this probably means looking at how this is happening on similar platforms and finding what our chat can offer that theirs can't, without trying to "sell" ours as something it's not. Ex. Part of our chat's core identity that that promotion should be faithful to is its association to Q&A sites about knowledge sharing under CC-BY-SA.

3. (Hopefully) address any embarrassing UX issues (including UX for general users and for room owners / mods).

In general, I'm interested in the idea of some sort of space (whether chat or something else) for people to seek help without expectations of either side to deliver something for a lasting knowledgebase, but from which such gems for Q&A can still spawn.

Some advantages of chat are that it's already a thing, and offers room for people to set their own rooms' rules (within certain boundaries, like the CoC & ToS) and be relatively more self-moderating. Blocking, muting, and basic flagging are already implemented. People can (given the privilege) just spawn new rooms, or invite others to join existing rooms.

• starball: some recent related chat transcript: me spitballing ideas on chat gating, me spitballing a helpdesk area idea. pls be gentle.
• Y.A.: I've seen the more reactionary "Chat isn't a more "casual" version of the main site, one with no rules", I've echoed it myself. On reflection, it seems inevitable there'll be a demand for spillover, S/E might as well participate in the supply. Maintaining standards, moderation and curation seem to be something we'll need to address as these venues are derailed into arguments and territoriality. I don't think it's search-indexed (yet) but chat is accessible by external tools and it's just a matter of time before anything placed in chat will need to conform to some sort of behaviour standards.
• starball: @W.O. I know you're not arguing it, but for the record, to people who say that "chat isn't...", if the rules in question are about scope and question quality, I'd ask- by what authority or canonical rule do you say that? I could understand being unhappy with rooms that are wildly unrelated to the associated network site's scope, but if other people want to use chat for relevant, interactive Q&A, and they're civil about it, who are you to stop them? why should we stop them? if we allow chit-chat in chat, why would we disallow something more valuable?
• starball: @W.O. as for general rules, I think the existing general network rules (like CoC and ToS) plus some chat-specific ones like what I loosely recall about using english are enough? also, chat is indexed, but restrictions were added to robots.txt in meta.stackexchange.com/q/403002/997587
• NoDataDumpNoContribution: The problem with chat may be that's it's too unordered. This question here seems to aim for the grey area, questions that are not high quality but still some value. A Reddit style comment system night work better than chat. Not sure though, don't know enough about chat.
• starball: @NoDataDumpNoContribution it's not clear to me what the connection is between threaded comments vs. linear messaging with replies and suitability for "grey area" questions. threaded comments has its strengths as a format/structure, but I don't see clearly why it's better for that.
• EmmaBee: @starball, first: I wanted to say that I really appreciate all the great ideas you've been sharing in chat lately! You've been doing some serious brainstorming with others, its been useful to hear you all talk through these challenges and potential ideas. I think you are definitely on to something about Chat being a core space to lean into, AND I think NoDataDumpNoContribution raises an interesting question: could chat be too unstructured for some problems? I'm not sure. We talk about the 3 lane highway, but is it a 1.5 lane? What do you think and why?
• starball: @EmmaBee as I've said, I don't understand NoData...'s line of thinking. whether chat is unstructured or not is situational- how many people are trying to help someone in chat? did they make a dedicated room or are they using a long-lived, shared, topical room? if conversations start overlapping, as I mention in my post, there's always the option to spawn a new room (if you have the privilege to do so). aside: technically, chat's reply feature gives you the same information needed to construct a tree-view of reply chains. tree vs. linear is a tradeoff; I wager linear works fine for this.
• EmmaBee: @starball ah, I see what you mean about the situational nature of chat’s structure. Agree that it really depends on factors. For when chat might be too unstructured, I was thinking about scenarios where a problem might warrant SOME structure. Like, lets say I am a chef working on a new recipe and needed help from other chefs. Having something like live document open for edits alongside a chat oculd make the collaboration easier. Without SOME structure, chat alone could feel too scattered, especially if multiple parts of the recipe need focused discussion. So likely problem dependent.
• starball: @EmmaBee yeah that's been in the back of my mind too... and yes, the best tool/format to facilitate that could be domain-specific, or depend on what people prefer to use. for software, maybe a repo on github. for documents, maybe something like google docs. but building those things well isn't trivial, and I'd caution against the company trying build thier own github or google docs without first observing use-cases. maybe it'd be best to just let people figure it out on their own with existing stuff.

Answer 8

After reading some of the discussion, I started to write this as a comment, but I feel like I should let people downvote it lol. I am aware that a long time ago, people didn't like the idea. Things are different now. Should everyone have the privilege to ask questions?

People need to be integrated into the community before they ask their first question. If we really want to address the problem, asking a question would be an earned privilege like chat. A new user asking a question consumes resources from the community in terms of onboarding, editing the post to bring it up to standards, etc. Why should any random person from the internet be allowed to consume community resources without having contributed anything of value first?

I'm not saying all new users' questions are crap. I want to give the new users who took some nominal amount of time to understand the community and the guidelines before posting more attention by excluding users who just want an answer and don't care about the site.

I think many of the mature sites are at a point now where expecting people to be familiar with what we're doing here before they can contribute their question to the library would help more than it would hurt. Maybe after a certain threshold of questions per day for a site, a token amount of reputation is required to ask a question on that site. Or we leave it up to a site's community to work out how much question friction they want to add. Maybe if you create an account you get to ask one question a month without having to worry about meeting any requirements. There's more than one way to shake a carbuncle.

We shouldn't treat every drive-by as a potential community member. This frees up the community to engage with the people who are more likely to become active members and adds some friction for users that are more likely to be a net-negative in terms of contribution. Guiding people toward some small steps they can take to learn how the site works should reduce the negative experiences and reduce some of work necessary to bring new user questions up to snuff. I have faith we can figure out the details if we agree on the general idea.

I'm not suggesting we exclude beginners, just that we ask people to demonstrate some level of engagement with the site. MusixMatch made me go through some training and pass a test before I could add/edit song lyrics to their systems. SE training could be things like how to find out what is on-topic for a site, how to compose a good question, what comments are for, how voting works, how to get help, how to respond to a close vote, etc.

• user1937198: Theres long been a presumption that questions are the low skill beginners entry-point to the system. But I've been starting to think we should question that. In terms of system understanding, good answers are a lot easier than good questions, and writing a good question is something that takes a lot of skill. Do we need to revisit that path to having people write answers, and then move to writing questions? I know personally I am very comfortable writing answers, but still find writing a good question very difficult.
• user1937198: The problem then being, if the expectation is answers are written by people who are knowledgeable in the subject, and questions are not an entryway, that how should we allow less knowledgeable participants to learn the system. I'm not sure, but I can't help but wonder if revisiting the problem of outdated questions/answers might be worth looking at by looking into the possibility of allowing beginners to suggest extensions/updates to existing questions and answers.
• user1937198: So I may not be able to write a good question from scratch, but I can identify a good old question that needs updating the latest language version.
• Karl Knechtel: " If we really want to address the problem, asking a question would be an earned privilege like chat." The problem is how to earn that privilege - something else has to be the bootstrap, and expecting everyone to e.g. contribute a useful edit (or even an answer) seems unrealistic and doesn't solve the drain-on-curation-resources problem. Simply seeing role-model questions probably isn't sufficient teaching, either. The Staging Ground is a great idea fundamentally, and I suspect nothing else really substitutes for it.
• Karl Knechtel: @user1937198 In principle, there's nothing wrong with "we effectively don't allow participants below a certain level of knowledge". In principle, lots of experts are not only eager to present material digestible to beginners, but willing to hammer their education into the shape of a good Q&A. But in practice, that doesn't happen nearly enough to satisfy demand.
• ColleenV: @KarlKnechtel The staging ground is fine, but I'd rather all that effort be spent on people who would be most likely to make high quality contributions in the future. I want a bigger speed bump before the community spends precious volunteer time curating a question.
• Karl Knechtel: @ColleenV Actually I do have one idea for implementing that sort of speed bump: have new users take a quiz first.
• ColleenV: @KarlKnechtel I think MusixMatch's quiz is a great template to start from. I found the whole process pretty engaging and they give you a badge for taking it which might incentivize more veteran users to take it if SE did something similar.
• Karl Knechtel: Do you have a link?
• V2Blast: Relevant page about MusixMatch's Lyrics Curator process: community.musixmatch.com/become-a-curator?lng=en
• EmmaBee: @ColleenV love the example from Musix Match. Thanks for sharing! Had not seen this before. You raise a really valid point about the benefits of requiring some learning before allowing new users to ask questions. I am sure its true that many of the true low-quality questions probably come from users who may not be familiar with the site’s expectations, and introducing a "speed bump" process could definitely help improve overall question quality. I've seen attempts at this, Staging Ground being the bigger effort, but there is probably a lighter weight V2 to explore.
• EmmaBee: That said, I think the challenge of grey-area questions goes beyond just new users not knowing the rules. Even experienced users often disagree on what kinds of borderline questions should be allowed. So while onboarding can help with improving question quality and educating askers, I think it doesn’t fully address the deeper disagreements about what types of questions fit within the site’s scope. Or would you see it differently?
• ColleenV: @EmmaBee I think that people get frustrated with low quality questions and that leads them to be harsher on all questions. When I was a mod on ELL, a lot of the angst was over a perceived lack of effort, which was particularly hard to overcome because native speakers can have a blindspot for certain aspects of the language that are really difficult for learners. I don’t think our disagreements about content would be such a big deal if go rid of the most egregiously low quality posts and increased the chances that when the community spent time helping someone with a post, they improved.
• ColleenV: A lot of the perception that we’re closing good questions is from people who care more about the individual author of a question getting help than whether the site benefits from the Q&A. I think trying to filter the people asking questions will make helping them more rewarding for the people that focus on that, and less frustrating for people focused on maintaining quality. We will never be completely correlated on quality judgements. That level of agreement on subjective assessments takes a bunch of training. At least that was my experience when I ran image quality assessments with analysts.
• Karl Knechtel: @EmmaBee "I've seen attempts at this, Staging Ground being the bigger effort, but there is probably a lighter weight V2 to explore." The light-weight version is: questions start closed and with up/downvoting disabled; comments are automatically purged when the question is opened (per what is currently called "reopening"); the purpose of this initial state is clearly explained to the OP up front as part of the question-asking process; at some threshold the privilege is granted to ask questions directly. Oh, and maybe a queue for such questions.

Answer 9

I had another random thought during a discussion where the tension between people wanting humans to answer their questions, people wanting to help them, and people focused on curating a library of knowledge came up again.

Maybe Stack Exchange needs both a library and a tutoring network. People can explicitly sign up to tutor people instead of having to fit what they enjoy doing into rules designed for curating a library of knowledge. The people more interested in curating the library can refer people to the tutoring side of the network. The tutoring side of the network can possibly create a Q&A from a particularly fruitful tutoring session as well as use the library side of the network as a reference. The UX of each side can be refined to support their different missions. For example, the new comment interface really belongs on the tutoring side, not the library side.

This will give us an option besides closing an interesting question that doesn't fit in the library scope, as well as get some people interested in the library side of the site to support their interests in tutoring.

Like I said, this is just a random thought so I haven't refined it. I was working on giving our users a way to customize their workflows as a way to resolve conflicting change requests to a feature and thought it might work for SE too.

• Y.A.: Best answer so far. A new direction for the company that will last as long as humanity's desire to strive and achieve lasts.
• NoDataDumpNoContribution: Great idea. I had hopes that the three lanes can achieve something like that, with discussions and chat the tutoring part and Q&A or articles the library part. However in the latest experiment, the company mixes everything again citing problems with visibility. It probably would depend on how many people want to volunteer for tutor. What's in it for the tutors? Why should they do it? Human tutoring doesn't scale well.
• ColleenV: @NoDataDumpNoContribution I think there are people that enjoy tutoring and don't mind volunteering their time to do it. Maybe there are more on ELL than SO. I assume the tutoring side could gave some sort of reputation ladder. Maybe you can earn the privilege to hold mini-classes? I don't have the patience to tutor so I'm not sure what would incentivize them.
• NoDataDumpNoContribution: I haven't either. At least not for free, money might incentivize me. The library thing has at least the scaling going for it. I can imagine that tutoring really will be an AI thing rather. But who knows. In any case separating tutoring and knowledge building more looks like the right way, mixing it together as the wrong way however.
• NoDataDumpNoContribution: I thought more about it and part of the value proposition at the beginning was also speed. The size of the place was so big that you would get answers super fast. This is important for the tutoring, but much, much less to the library part. One could imagine the tutoring section taking over triages for new questions to semi-automatically tutor people as quickly as possible and either direct people to the library or request a new addition to the library. The library could take more time, hide comments, reorganize the knowledge.
• starball: I was floating some related discussion in the tavern in the past few days. circa chat.meta.stackexchange.com/transcript/89?m=10131763#10131763
• EmmaBee: @ColleenV Separate space that's more tutoring oriented is definitely an angle worth exploring. Thanks for sharing! Also, this is really interesting "Maybe you can earn the privilege to hold mini-classes" Out of interest, would you personally be interested in being a "helper" in a space like this? What would need to be true for something like this to be something YOU might participate in?
• ColleenV: @EmmaBee I have no patience for tutoring. If y'all wanted to run a test, I think ELL would be a great place to start. Lots of learners there would appreciate tutoring and there are plenty of community members that are actual teachers participating there that would definitely be interested in a more flexible space. They might have better ideas of how it should work as well.
• EmmaBee: @ColleenV Fair! I totally get that tutoring isn’t for everyone. I really appreciated your idea though... I think a separate tutoring space could help address some of SE’s current tensions. That said, more broadly I’ve been thinking about what kinds of spaces or interactions feel worthwhile for folks who aren’t beginners. Not necessarily to help others, but to stay engaged themselves. Many of our conversations often center on what new askers/beginners need, but I’d also be curious to hear your perspective on what might be missing or needed for non-beginners.
• ColleenV: @EmmaBee I'm guessing by "beginner" you mean someone new to the network? I think the best path forward is to stop focusing on what people might like and start focusing on what the company wants to accomplish (other than making money). There's this library mission that the current leadership inherited and it seems to have lost its shine for them. What do you want to build? Express that in a way that people will want to help you build it. My time on SE wasn't just entertaining myself like a video game. I thought we were building something worthwhile and I was happy to volunteer.
• ColleenV: I believe there is a global need for helping people in places that don't have a lot of native English speakers learn English from native speakers. Being fluent in English opens up a lot of opportunities. Maybe SE can start there. Maybe instead of a Q&A site, genAI should be tutoring on how to use these tools for all skill levels (good luck getting people to give that expertise away for free though). For me personally, I have so many demands on my time, I don't want to spend it on something with no impact.

Answer 10

I'm not sure if this is on point for the question or tangential to it, but I feel that outside of SO there's an absolute plague of weird and arbitrary rules, which get applied even though they were never formally decided on and not everyone agrees with them.

One of the most widespread is that idea that you can only use one ? character per post, and any question that contains more than one ? character must be closed, regardless of the context or meaning behind the sentences they terminate.

Of course there is a long standing and perfectly good rule that you can't ask multiple unrelated questions per post. But this has been interpreted more and more narrowly over time, to the point where nowadays someone can ask something like "Is there a widget that does X? If there is such a widget, what's it called?", and it will be closed as "lacks focus". (I will not be surprised if someone pops up in the comments to argue that "Is there a widget that does X?" and "If there is such a widget, what's it called?" are actually different questions and should be posted separately, notwithstanding that if they were they would immediately be closed as duplicates.)

This isn't the only such weird "folk rule" - there are lots of others and a lot of them are site-specific. It's generally easy to get questions reopened when they fall foul of such rules while otherwise being perfectly good questions, because there's usually no shortage of people with the appropriate privileges who see how silly it is. But the point is that even if we as a community want to demand that "If there's a widget that does X, what's it called?" is acceptable while the above two-? version isn't, we shouldn't be doing it without reaching that agreement explicitly first and setting its boundaries.

The current situation is that certain kinds of question are constantly closed and opened again by people enacting and retracting rules that have never been agreed on or even written down, and which the OP consequently can't possibly know about. This isn't good for anyone, neither the person trying to get an answer to their question, nor the hypothetical future visitor with the same question, nor for anyone who just wants to hang out and answer questions without the constant stress of arbitrary bureaucracy.

To tie this back to the question, I think it's really good that this question is being asked, and I hope it will lead to some kind of resolution. If we can at least agree on what the site's rules are for, it would be a step towards making some agreements on what the rules are, instead of just leaving it up to the close/reopen voters.

• Y.A.: That would be down to the good sense of those who apply the rules and naturally people disagree - hashing these thing out on the per-site metas. Your example of: "Is there a widget that does X? If there is such a widget, what's it called?" - on many sites the first would be closed as a silly question as it invites a yes/no answer, the second closed because it asks for product recommendations which would be off-topic on all but two sites.
• N. Virgo: @W.O. (in regards to the on/off topicness of product recommendations, that's irrelevant to my point so I won't address it, but the rest of your comment deserves a response, give me a minute or so.)
• N. Virgo: @W.O. the point is that these decisions are not being hashed out on the per-site metas. The rules shift because people's perceptions shift, not because any agreement is being made to change them. If you look at the original discussion about multiple questions you find quite reasonable positions in which it's seen as self-evident that some related questions in the same post should be allowed. I don't think you'll find any meta discussion on any site where people agreed to change that.
• N. Virgo: But in practice it has changed anyway, despite a lack of discussion, and that's what I'm trying to highlight in this answer.
• Y.A.: You say: "the point is that these decisions are not being hashed out on the per-site metas" - They most certainly are on the sites that I frequent (Worldbuilding, biology, medicalsciences, psychology, Sci-Fi&Fantasy etc..). Some sites are more strict on this than others. Is there a particular site that's a target of your comments? I'd also disagree with your conclusion drawn from the link you give. To me the consensus remains one question per post.
• N. Virgo: @W.O. you have to read the second half of the answer I linked to, not just the first half. People argue about stuff in those metas, but I repeat that I don't think you'll find a discussion where any specific change has been agreed to on this specific point (a strict rule against multiple ?'s per question as opposed to multiple insufficiently related questions), on any site's meta. If you can, great - post it here.
• N. Virgo: Just in case it gets lost, the point here isn't about whether one or other choice is a good rule vs. a bad one. The point is about not having a systematic mechanism for determining what the rule is, and therefore constantly subjecting the site's visitors to arbitrary bureaucracy and pointless arguments. If there was a clear agreement about the rule, even if I disagreed with it, I'd be much happier than I am with the current situation.
• Y.A.: The second half of the linked post seems to be about rate limits to multiple posted questions each in their own thread, I don't get what you mean. For an example on W/B, this question seems to do a great job of clarifying (as an aside to the actual question being asked there): Risk Factor Definition: Multiple Questions. Continued.
• Y.A.: Cont.: On SF&F: When are we okay with multiple-questions-in-one? - "when their presence clarifies, pins down or narrows the scope of the original question" which seems clear to me. I can find others if you wish.
• N. Virgo: @W.O. oh, I agree that's clear, and it matches my understanding of the long-established rule. The issue is that a lot of people vote to close questions with multiple ?'s, regardless of whether they pin down or narrow the scope of the question, because that nuance is gradually being forgotten.
• N. Virgo: @KarlKnechtel I should take these things to meta more often. But it's a hard job. People can get really attached to these folk rules - they're used to it being the rule because it's why questions get closed, and they close questions themselves because it's the rule, and things can get pretty passionate if you try to question that. If it's not addressed carefully the folk rule will end up getting codified as an actual rule, and I don't feel I'm good enough at politics to reliably avoid that.
• Karl Knechtel: You've already done a service by identifying the problem of folk rules being a thing, and raising awareness on the network meta.
• EmmaBee: @N.Virgo This is a great addition to the conversation! I haven't really thought about how "folk rules" can form quietly over time. I think a few questions this raises for me: Are we all following written, agreed-upon rules? Or are we just enforcing what we think the rules are based on what others do? And then... IF the rules are evolving, are they evolving through consensus, or more through... momentum? I’d love to know: what do you think could help here? Are there ways we could better distinguish between some of the rules without leaving it so open?
• EmmaBee: +1 to what @KarlKnechtel says: You've already done a service by identifying the problem of folk rules being a thing" Thanks for your contribution to this important conversation, N. Virgo!
• N. Virgo: @EmmaBee I think it would help a lot if something came down "from above" making people aware that rules can drift over time in this way and that they're not supposed to, at least not without proper discussion. That would hopefully mean that when someone points out that a folk rule isn't actually the agreed rule it would have a bit more impact. I feel that in the past there was a social agreement that the review queues were meant to implement the agreed rules, and that close/reopen voting against the meta consensus was bad behaviour - that seems to have been lost a bit in recent years.

Answer 11

Personal opinion, not covering all aspects of the question.

Often, I found that an otherwise good question find no fit on any of our sites, simply because it's open-ended; yet being open ended, does not necessarily mean it doesn't have a determined answer.

Take a radical example, an elementary school student in some country is learning about quadratic functions in math and python in IT, he was told by his teacher that 5-degree polynomials and higher have no solution formula, so he go on to research about this. He came to Math.SE and asks about his confusion, but cannot formulate his idea. A good answer (in my opinion) may explain that Newton's method is a general-purpose algorithm that can find root in any real-valued continuous (analytic? is that a necessary condition? I'm not well-versed here) function, and was told about Sage. He return to his IDE and search engine to practice more. This is just a hypothetical example I came up with, it's not real though.

Real experience I've had on Computer Science.

Recently I'm interested in compiler design. While doing thought experiments on optimization before reading books to verify my ideas, I realized that general program semantics cannot be reliably analyzed - many of us may have heard about the halting problem, it's one case of program semantic.

I asked about the impossibility of reliably optimizaing away all "redundant" computations in general, but was faced with requests to define what I mean by redundant. I understand what they were concerned with - an ambiguous definition may deviate me away from my goal. But the one answer I received made its own definition of "redundant", probably because my definition wasn't satisfactory.

Oh, nearly forgot. I once gave an determined answer to an otherwise poorly-focused question: https://crypto.stackexchange.com/a/114687/36960

• starball: "But the one answer I received made its own definition of "redundant"," is there an implied good/badness judgement here on your part, or is this just an observation? I can't tell what the point of this story is. do you consider that Q&A pair good?
• DannyNiu: I'm trying to argue my point that "an open-ended Q can have a determined answer" when I introduced the example of the "own definition of redundant in answer". @starball . These aren't directly but I believe are somewhatly related.
• starball: related: second half of stackoverflow.com/help/dont-ask. but I don't think your example is a good example of that class of constructive questions.
• DannyNiu: @starball Appreciated. My SE participation experience really don't make me a good presentor here.
• NoDataDumpNoContribution: I feel like you touch on the only real space that is left for Exchanges to improve. Opinion based or too broach questions. They could exist in a different setting and would be useful if done right. All the other "messy" questions just don't resonate because they don't add anything.
• DannyNiu: @NoDataDumpNoContribution You summarize well!

Answer 12

The problem is ultimately intractable because Stack Overflow's fundamental concept - the thing that makes it tick - is community involvement over and above just asking questions and receiving answers to them. That means curating: working the review queues, editing content to be objectively better, and/or participating in initiatives like Charcoal to proactively fight spam.

Without curation, the site becomes an unusable mess and Stack Exchange Inc. has nothing; yet nothing that SE Inc. has done, or tried to do, over the past decade has been focused on improving curation. In contrast, in that time the company has been responsible for numerous actions that are actively hostile to established curators, resulting in a continuing exodus of the latter. (Personally I was in the top 1,000 of curators at one time but have not actioned a review queue in months, except for some activity in Staging Ground.)

The end result is that the curation burden has increased and the site has become worse overall. SE Inc. has responded to that by requiring questions to be more narrowly focused, thus giving curators leeway to close a higher percentage of them - but that's not a fix for the problem, because being able to review and close 10 questions in the time I would previously have taken to review and close 1, does not help when there are 100 questions that need to be reviewed.

It also creates a new problem in that questions that previously would have been allowed to remain on the main site, and potentially attracted useful answers - i.e. exactly the type of questions you are talking about - are now being closed, and that pisses off the people asking those questions. The unreasonable hostility towards SO from the general programming populace who does not curate, used to be bad; it's many times worse now as a result of the question closure changes. If you don't believe me, read any reddit thread involving SO and count the number of complaints about its "moderators" being "toxic".

The answer to all of SO's problems, then, is "get more users curating". The problem there is that SE Inc. has spent the past decade refocusing the site towards users who don't want to curate, and any attempt to pivot away from that aim to onboarding users who do is unlikely to succeed, because those users have already found other communities to do that in. On top of that is the fact that the tooling for curation is simply abysmal and has been ever since the site's creation, despite the repeated requests of established curators for its improvement. New users aren't going to curate if doing so is an annoying and painful experience, which is most certainly is.

But finally, and most importantly, is that the landscape has changed. Software development as a career is no longer dominated by thoughtful autodidacts who desire to be taught how to fish, in order to improve their skills; it is being overrun by those who expect - not want - to have the fish handed directly to them, on a silver platter. A site like Stack Overflow is an invaluable resource for the former group; to the latter it's actively hostile because it expects them to comply with rules before it maybe gives them their fish. And why would you bother with rules when you can just vibe code with an LLM?

In short, I don't believe that the existential crisis facing Stack Overflow is one that can be overcome: it has become Experts Exchange, the thing it hates, because its owners focused on making money rather than building a community large enough to ensure its survival. And now it's simply too late to do the latter, because enough of the new generation just isn't interested in SO's mission of being a knowledge repository.

• NoDataDumpNoContribution: I agree with everything except the end. "now it's simply too late to do the latter" Why should it be too late for anything? There is no better alternative to StackOverflow existing. It might have been easier, would we have discussed this and changed it 10 years ago, but as long as there is no other such service, how can it truly be too late?
• Ian Kemp - SE killed by LLMs: @NoDataDumpNoContribution the current generation of "programmers" has made quite clear that they do not place any value on the existence of a knowledge repository like SO (or Wikipedia, or ...) and therefore have no interest in curating it.
• ColleenV: SO will survive, at least for a while, it just won't look like the SO that you and I signed up for. It will have to adapt, probably in ways that won't appeal to us, but we're already actively disengaged. It would be a waste of energy and resources trying to get us back.
• Sam Dean: @IanKemp-SEkilledbyLLMs any source on that comment? I'd consider myself a fairly young programmer and I contribute, and in my friendship group people certainly place value on opensource and collaborative endeavours.
• NoDataDumpNoContribution: "they do not place any value on the existence of a knowledge repository like SO" But then it doesn't really matter what they try here, even getting us back won't help because the current programmer generation decided they won't need us to answer their questions. It wasn't the fault of the company really. The company didn't make the programmers. Knowledge building is just not fashionable at the moment?
• Ian Kemp - SE killed by LLMs: @SamDean I'm speaking in the aggregate, of course, and going by the sentiment expressed in the mentioned reddit threads; any attempt to explain to a commenter there as to exactly why their question was closed, is invariably met with deeper hostility. Unfortunately it seems that by and large today's newcomers are simply incapable of responding rationally to being told "no".
• Ian Kemp - SE killed by LLMs: @NoDataDumpNoContribution I strongly disagree; SE Inc had a responsibility to understand that community is the foundation of any successful community-driven website, and they utterly failed at this not just once but time and time and time again. They had all the tools, all the opportunities to make SO a success - and chose to ignore them to chase ad revenue and user counts and moronic experiments to boost those.
• NoDataDumpNoContribution: But you say that by and large newcomers are simply incapable of responding rationally to a no. Even if the company would have done everything as you desire, how would newcomers be different? Do you mean that a different company would have educated newcomers so much they would behave quite differently from how they behave now?
• Karl Knechtel: "SE Inc. has responded to that by requiring questions to be more narrowly focused, thus giving curators leeway to close a higher percentage of them - but that's not a fix for the problem, because being able to review and close 10 questions in the time I would previously have taken to review and close 1, does not help when there are 100 questions that need to be reviewed." - In practice, the requirement for focus rarely makes people ask more questions - they just give up when they can't get a complete step-by-step solution to a problem. And focus is really necessary for a searchable resource.
• zcoop98: I'll also voice dissent against (only) the final conclusion, for two reasons. First, because it's simply bleak and defeatist, and leaves no room for survival, which means I don't like it and refuse to submit to its implications (lol). Second, I think it gives too much weight and representation to the culture of Reddit "SO-is-toxic" vibe coders, which I just don't think is as representative of the industry as the conclusion implies, especially long term. It's far too early to declare that widespread vibe coding is the (bleak, dark, empty) future we're stuck with.
• starball: how can you say "nothing that SE Inc. has done, or tried to do, over the past decade has been focused on improving curation" and mention staging ground in adjacent sentences? SG essentially frontlines the closure state for a temporary period- it makes feedback go before answering. it's not ideal, but it's not nothing.
• starball: aside: think people who don't vibe code will always exist. it's like how in gaming, some people are innovators, and some people just want to play the meta game.
• Ian Kemp - SE killed by LLMs: @NoDataDumpNoContribution Because of the way SE Inc. positioned the site they created a false expectation of what it's about, which permeated into programming culture as a whole. The end result is that we became overwhelmed by people are aren't interested in community-building and aren't interested in being said "no" to, and the harsher question closure policy necessitated by that influx then disincentivised the type of people who we'd like to see as members of the SO community from joining.
• Ian Kemp - SE killed by LLMs: @zcoop98 1. It's not just the number of comments that are negative towards SO, it's the number of upvotes on those comments. Yes, drawing conclusions from the cesspool that is reddit data is questionable, but there isn't really a better way that I'm aware of to garner sentiment on how developers who aren't part of the SO community, view it.
• Ian Kemp - SE killed by LLMs: @starball Staging Ground is an idea that might have saved the site if it had been implemented 5 years ago and for all review queues, when the load was low enough and quality high enough that it would have served its purpose of helping to educate new users. Nowadays SG is populated primarily by irredeemable trash and the favourite activity of users whose questions are reviewed there appears to be to entirely ignore feedback, pad their question with nonsense so it becomes eligible for reevaluation, and submit it for such.

Answer 13

Some view Stack Overflow as a technical Q&A site where precision, strict rules, and reliability make it valuable. Others feel it needs to adapt to stay relevant, embracing a broader scope, such as questions about real-world developer work, including architecture trade-offs, tooling choices, or experience-based insights, even if these topics extend beyond "just code."

I need to offer a bit of a frame challenge here. I'm not entirely sure what you mean by "precision", but I guess you're talking about the expectation to avoid subjectivity in questions. But this, along with the idea of closing questions according to a strict checklist of reasons, are not in conflict with broadening a site's scope. It just requires suitable alterations to the checklist.

Highlighting from my own answer on the original MSO post:

...comparison questions... could be fine [by current standards] if the OP provides a single, objective metric for comparison

But when people ask about alternatives like this, I think what they're usually really trying to figure out is a process for choosing between the alternatives.... The problems occur because people want someone else to decide for them what factors are most important...

So a technically-oriented Stack Exchange community should never accept a question which is "which way should I implement this?", because that addresses only the OP (or a unique set of circumstances) and doesn't help future readers. But it could be fine to accept "what do I need to consider in order to choose between these implementations?". (Although some care is still needed.)

• starball: related to comparison questions: meta.stackoverflow.com/a/426216/11107541

Answer 14

Specifically, we were talking about "grey area" questions that could potentially be helpful, especially the ones that are more experience-based, exploratory, or practical. These types of questions often don’t fit neatly within the rigid "rules" of traditional SO Q&A, leading to closures despite their potential value.

Stack Overflow gets lots of "gray" questions that are off-topic for the same old reasons: asking for software recommendations, asking about configuring/managing some piece of software and not programming, etc. These questions typically get closed (as they should).

One of the best things Stack Exchange has done is to create new stack sites to address these needs. SuperUser and ServerFault have been around for a long time, but more recently we got sites like Software Recommendations, Code Review, Programming Language Design, User Experience, etc. This ensures these types of questions have a home and aren't just discarded. In my personal opinion, it took far too long for these stacks to be created. We closed a decade's worth of software recommendation questions whose authors would have loved to have an appropriate place to ask them. I understand that the infrastructure to do that wasn't there in the beginning, but we've had the capability to do better for a long time.

Trying to change the scope of an existing site to incorporate new types of questions will always be an uphill battle. Creating new sites for those questions is a much smoother approach. You can set guidelines and expectations from the beginning, and potential contributors know what to expect before they decide to interact there. You can easily judge the success of the experiment by recording stats for the new site instead of taking data from a larger site like SO and trying to extract data for only a certain type of question.

I'll recommend a couple of things for going forward:

• Periodically review the most common reasons questions get closed, and consider whether setting up a new stack site for them would be an appropriate response.
• Make it easier to migrate a question from one community to another. This is currently done via the "Close" button (which has always seemed to me like an unexpected place to put it). There are a few buttons for pre-selected migration targets, but it's much more difficult if you need to migrate it somewhere else. Users should be able to easily migrate a question from anywhere to anywhere.
• It would be helpful if questions could be flagged as potentially off-topic while they are being written. A gentle warning like "This looks like a request for a software recommendation. Would you like to post it instead to the Software Recommendation site where it is more likely to get a response?" would reduce the number of off-topic posts without reducing the number of questions. Perhaps even a "which site should I submit this question to?" wizard for new users.
• Remember that when questions don't "fit neatly within the rigid rules of traditional SO Q&A", that means they likely also don't fit within a web application framework that was designed for that rigid Q&A environment. Be open to more significant differences between sites. I have included some examples here, I'm sure long-time users of the various stack sites would have their own suggestions. These sorts of changes wouldn't make sense for all sites, but anything outside the traditional Q&A realm will likely benefit from some sort of use-case-specific customization.
  • The existing system of selecting a "correct" answer doesn't work as well for the Code Review stack. If I have two answers that each give good feedback on different things, choosing which one to mark as "right" feels arbitrary. When posting code to review, I sometimes wish there was an option that would display line numbers for larger blocks of code so that they could more easily be referenced.
  • You could enhance the Electrical Engineering site so users have a CircuitLab-like schematic editor when asking questions. When answering the question, I can "quote" the schematic to load it up as-is, then make a few tweaks and display it in my answer. The current workflow involves posting screenshots of circuit diagrams created using external software, which makes answering questions more of a burden than it needs to be and thus makes users less likely to make the effort to contribute.
  • Questions on the Home Improvement stack often involve photos of a problem. When answering a question, it would frequently be helpful if I could "quote" an image from the question, crop/zoom it to a section of interest, and add my freehand red circles directly from the answer editor.

Answer 15

In the real-world I'm a solution architect, and in the Stack Exchange world I focus on questions tagged with [architecture], specifically in the Stack Overflow (SO) and Software Engineering (SE) sites.

My perception is that there's a natural tension between Architecture as a discipline and SO/SE guidelines, particularly those concerning "Opinion-Based" questions:

• Architecture is not binary in the way that a programming language like C# or Java is, which are defined by specifications - code can be taken and tested to see if it adheres (true or false) to the relevant language specification. On that basis the SO/SE concept / guidelines are well suited to programming questions.

• A key aspect of architecture (as a discipline) is making decisions that form an architecture and guide its implementation. The collective experience, judgement and opinions of those involved is central to how that happens. Whilst there are established principles and method in architecture, their relevance and application is often subject to interpretation and judgement - i.e. opinions, which is not well suited to the SO/SE concept / guidelines.

Specifically, we were talking about "grey area" questions that could potentially be helpful, especially the ones that are more experience-based, exploratory, or practical.

My interpretation of what you mean by "grey area" questions is where the substance of the question and the SO/SE guidelines, don't always neatly align. I'd argue architecture / [architecture] questions often fall into this area, for the reasons outlined above; architecture definitely skews toward being 'experience-based, exploratory' as you put it.

Relevant Reputation

Anecdotal research indicates that those closing [architecture] questions often have low or no [architecture] reputation; I totally acknowledge that they may have valuable real-world architecture experience despite this.

But this does raise the question: to what extent should "grey zone" questions be exposed to closure by closers with tag-relevant reputation below a certain level, given that subject matter knowledge is a key aspect of "grey zone" questions. In many cases I have seen, the justification to close [architecture] questions appears to be based solely on perceived guideline violations and without any consideration for the inherent nature of architecture, or the value that the question and it's answers might bring.

Solution Ideas

• Of the quorum needed to close a question, X% of closers need to have tag-relevant reputation above a certain score. This rule could be applied to all tags, or selectively.

• Grey zone questions can be flagged but left open. The flag could take any form, including the description of the guideline at risk of being violated. The goal would be 2-fold: to encourage improvement to the question; warn readers.

Data Of Note

As of September 2024, of the top 50 [architecture] questions, ranked by upvotes:

• 43 (86%) were answered.
• 18 (36%) were closed, with only 2 of those not being answered.
• These 18 closed questions had an average 304 upvotes.
• Those not closed had an average of 439 upvotes. The top-voted question is an outlier, by taking that out the average upvote is 318 upvotes.

My interpretation of that data is that [architecture] questions:

• Can still be answered even where they are judged to not meet guidelines, and
• seem to be of use based on upvotes awarded.

I have not gathered data on the awarded Answers to these questions but I would assume they also attract significant upvotes.

Update - Regarding: 'we allow "constructive subjective" questions'

You raise a great point. My observation is that the interpretation of those specific guidelines isn't always consistently applied (specifically within [architecture], hence my hypothesis that this is partially due to the inherently subjective nature of architecture).

Update 24 May - Regarding: 'the guidelines for formulating "constructive subjecting" questions' and 'I'm talking about clarifying them'

Sounds like a good idea to me.

Has there been any study into how moderators actually moderate? E.g. How do they actually remember and apply guidelines? Understanding that would help clarification efforts.

Some specifics...

However, if your motivation is "I would like others to explain ______ to me", then you are probably OK.

I see a lot of closed architecture questions that I think would meet this statement, although in practice the word 'probably' is an issue: questioners probably default to expansive interpretation, where as moderators are required to make judgement calls.

Looking at the current guidelines, there's an array of specific points which are applicable, but I can see how a moderator would be hard pressed to apply them all equally.

...What does that mean? Constructive subjective questions:

1. inspire answers that explain "why" and "how"
2. tend to have long, not short, answers
3. have a constructive, fair, and impartial tone
4. invite sharing experiences over opinions
5. insist that opinion be backed up with facts and references
6. are more than just mindless social fun

Regarding the specific guidance above and Architecture questions:

Re #4 - as I sit here and write, I get the gist of what I think this means, but I'd be hard-pressed to define it, let alone as a rule. My impression is that in giving an answer to an architecture question, it would be very easy for someone to write of their experience and for a moderator to see it as an opinion.

As to how the questioner frames their question, to elicit experiences and not opinions - that sounds non-trivial, and even harder to quantify into guidelines.

Is there any guidance on how to distinguish between experiences and opinions?

Re #5 - there is a body of architecture knowledge out there, so it's often not hard to provide references. (I'm less certain about 'facts' as far as architecture goes). But often connecting the reference (and your answer) to the question falls back on experience, which in turn is often interpreted by moderators as opinion.

Again, as to how the questioner frames their question... There's probably scope to improve this guideline in a helpful way. I don't have the ideas to mind right now.

One thing I have observed a lot of: architecture questions that I know are answerable with referenceable material, but which someone without architectural domain knowledge could assume is only answerable with an opinion.

It reminds me of some job interview situations where the interviewer is looking for a few key words in the answer, and if they don't include those then the answer isn't as "good". We don't want 'how to ask' subjective questions to turn into some arcane dark art.

Re #1 & 2 - I might be guilty of conflating here, or poor judgement, but I suspect that questions seeking "long" and "why" answers will often look like ones seeking "opinions". If someone is trying to moderate a bunch of questions, I can understand the tendency to skim read, and conclude in mental short-hand that something is/requires an opinion, and close - with mitigating guidelines hence being overlooked.

Re #3 - I'm unsure how much weight this carries in most moderators minds - my impression (totally lacking in any evidence) is that if a moderator thinks your question requires an opinion, the best tone in the world is not going to save it from closure.

• starball: "i.e. opinions, which is not well suited to the SO/SE concept / guidelines." see second half of meta.stackexchange.com/help/dont-ask it really irks me when I see people who don't seem aware that we allow "constructive subjective" questions.
• starball: "X% of closers need to have tag-relevant reputation above a certain score" consider that not all close reasons are the same, and not all may require or benefit from subject mater expertise as much to judge.
• Adrian K: Agreed - 'not all may require or benefit from subject mater expertise'.
• EmmaBee: Thanks for this thoughtful response! This is a really good example that highlights the issue well! Your idea about requiring tag-relevant reputation is an interesting one. Could probably help reduce uninformed or knee-jerk closure votes, especially from people who aren't familiar with the topic. Still, that might cover half (or some % of) the challenge...
• EmmaBee: Starball’s comment above (perhaps unintentionally) highlights what stand out to me as deeper issue: there’s still a lot of disagreement or confusion about what kinds of questions actually belong here. Even when the guidelines allow certain types, like constructive subjective ones, people don't know, interpret or apply them very differently. So maybe the challenge isn’t just uninformed votes, but also a lack of shared understanding about the boundaries themselves. What do you think?
• Adrian K: RE 'lack of shared understanding' I 100% agree. People (the community) are individuals, so perspectives are different and shared perspective will only ever be ... fuzzy / partial. Therefore, for a solution to be effective, it would have to work with that in mind.
• mbaird: @EmmaBee I see lots of disagreement and confusion over what questions are allowed on SO for questions that fall under the AWS collective. I constantly see close votes on AWS questions from people with low SO reputation, and low or 0 reputation in the relevant tags/subjects of the question. With questions that cross the boundaries of software development and cloud infrastructure, there is often no clear home on the SE network, but the existence of the AWS collective on SO draws most of those questions to there, where people then vote to close because it isn't a basic programming question.
• mbaird: @EmmaBee It is my opinion that there are too many SE sites like devops.stackexchange.com and softwareengineering.stackexchange.com with massive "grey area" overlap with StackOverflow, due to the tendency we software engineers have to want to categorize things and place them into separate buckets. I believe that we focus too much on this and close too many, otherwise valid questions, because we don't like what site they were asked on. I would prefer that we instead focus on assisting people that are asking for help, instead of focusing in closing their questions.
• Karl Knechtel: @starball The concept of whether a subjective question is being "constructive" really needs to be hashed out on meta. That actually has better visibility than the Help Center for dedicated curators, and would be a place where interpretation can be done and consensus can be reached.
• starball: @KarlKnechtel I don't follow the train of through between my comment and yours. are you talking about loosening the guidelines for formulating "constructive subjecting" questions?
• Karl Knechtel: I'm talking about clarifying them, and about making sure people are on the same page about them. I suspect a lot of people try to infer the entire policy from the close reason text.

Answer 16

My experience here is that this site used to be very useful. I have started asking questions in new areas where I am less expert, but nonetheless am an experienced programmer, with many questions that have been upvoted a lot, so the fact that many of my recent questions have been downvoted, valid answers actually DELETED? The implication is that either:

1. I have gotten lazy and my questions have gotten significantly worse
2. I am asking questions in areas where it is more difficult to ask a perfect question and people answering are impatient with any imperfections
3. The site is being overrun by people who are just playing gotcha and enjoy downgrading questions.

I think to be fair, it's a mix of 2 and 3. There are questions I have asked where someone actually answered,and I learned something, but sometimes the question was downvoted after, before, and (I don't understand how this is possible) answers were actually deleted (which were correct, and I would have given them the credit).

I know management at StackOverflow feels like you are the keepers of the flame, but as a user, it is coming off like you are rewarding arrogance, antisocial behavior, and that you don't want me as a user. This place has become hostile, and not very useful. AI has very little to do with it. I posted on meta before, and gave up after seeing the usual nonsense. It's always this way in an organization that is spiraling toward the toilet. Everything is fine, the problems are due to external circumstances, no one is to blame.

Yes, if traffic is 60% down, that's because they are more beginner than they used to be. That makes sense, because 60% of the people with experience who used to ask questions come to this site died last week?

I really wish I could have the site back in its original form. It used to be that if someone had good intent, was puzzled by something and had tried to put in some work, then a kind soul would take a look and try to answer. And if they couldn't they would move on, and not feel obligated to vote it down, make derisive comments.

There's a fairly large subreddit where people have shared their experiences here. So I know I'm not alone

• starball: answer deletion typically indicates something is really wrong. do you have post IDs for examples of your community-deleted answers?
• Karl Knechtel: I can assure you that the people closing questions don't enjoy it. We'd usually be far happier if the question weren't there in the first place. Except for good duplicates; then we're happy that the target became easier to find.
• EmmaBee: Thanks for sharing this. I can hear how frustrating and discouraging your recent experiences have been, especially after being part of the site for so long. I can imagine its not a good feeling to FEEL like your questions are being shut down or ignored when you're asking in good faith and genuinely trying to learn. That tension you named between "good intent" and "getting everything perfect" has come up a lot, and I think it's something the site is really struggling with right now.
• EmmaBee: These challenges don't have easy fixes, but just wanted to say that your perspective matters. And if you're ever up for it, I'd be curious to hear more about what kinds of experiences do feel worthwhile to you these days, or what you think has been lost that you'd want to get back. Thanks again for taking the time!
• Dov: @EmmaBee I would love to talk privately to anyone on stackexchange who will listen, because I would like to use this site. I don't know how to do this, not interested in posting my email here, but you should be able to contact me if you are on staff (I think?)