-26

Before jumping into this post, you should probably read post Next steps for open-ended questions that describes changes released today that provide a path for integrating open-ended questions on Stack Overflow.

Let’s rip the band-aid off that elephant in the room. Expanding the scope of what’s on Stack Overflow is a big change. There are lots of opinions and ideas on this topic. And it touches on a lot of things, both technical and philosophical. There’s a lot to figure out.

We tried to guide that expansion by building a new area of the site (Discussions) and by creating other branches of the primary experience (what’s been referred to as “Opinion-based content” (OBC)), but neither of those ended up in a good place. Despite their drawbacks, we saw promising usage of those formats — users asking and answering these types of questions, and users organizing them and making sure their quality adheres to reasonable standards. And we have a pretty decent sense of the kinds of things (good, bad, and ugly) that we’ll see.

We are sure that the community can do a better job guiding this than we have, provided you have all the tools you’re accustomed to. And I don’t just mean curation options and moderation features. I also mean the ability to work together and figure out how to sift the pearls from the sand via the policies and guidelines FAQ. I mean the innovative thinking that built and maintains a world-class spam management & control system (Charcoal). I mean the generous efforts that still churn out helpful tools (Stack Apps) from an aging and poorly-documented API. I mean the dedication that keeps you coming back here every day (even if it’s only to tell us what we’re doing wrong). All of that is what Stack Overflow needs right now.

This is the time to seize the moment. We’re all aware of the current challenges with participation. Sure, an earlier start on this would have been great. But it’s not too late, and those who agree can help chart the course ahead. Those who delight in saying it’s too late don’t need to be part of the conversation.

Okay, locker room pep talk over. Let’s get to work.

As the necessary conversations commence, we’re laying out a few guidelines:

  • Expanding the scope of content on Stack Overflow is the direction we, as a company, have decided to move towards. The old status quo is not sustainable. Stack Overflow needs to be guided by rules and policies that make sense for 2026 and the future. Many reading that may feel moved to debate this stance. You are welcome to express that point of view, but take a moment to consider why you feel the way you do. And try to end your posts with a paragraph that starts like “But if it HAS to change, here is how I think it can work best…”

  • We need to use the tools we have. We do not have the resources to reinvent the core curation experience or the reputation system on the public platform at this time. We can make modest changes, purposefully or experimentally. But our capacity to do that is limited, so be thoughtful about the changes you might collectively request. The bigger the change, the longer it will take, and the fewer other changes we can make.

We look to you for guidance on what comes next. We are here to support you. An AMA* with some members of the team supporting you is a quick look at the team that you’ll be hearing from, and a chance to ask them some questions.

What do we need to talk about?

Create some new questions here on Meta and start talking about what’s most important. We’ll be doing some of that, but don’t wait for Community Managers (CMs) to post. We will have ideas and suggestions. And we might nudge you to dig deeper or consider another perspective. But aside from letting you know more about how the “OBC migration” will work, there’s no big sequel to today’s posts waiting to drop.

As we see areas of ideation and focus emerge in discussions on Meta, we might propose establishing working groups to move things along. So keep that in mind as another tool we can leverage.

Want data to guide discussion and decision-making? If you can’t pull it yourselves, let us know what you want to see and we’ll do our best.

Things we should probably tackle first:

  • What makes a high quality open-ended question? Where should that guidance be detailed? We’re hopeful that someone from the community will start this conversation. Highly related topics to discuss:

    • What should the open-ended question type tags be? What should be the type that designates traditional Q&A (currently “How-to / Troubleshooting”)? We will need to stick with the three current open-ended question tags (, , and ) until the OBC migration has been completed, but discussion on this topic can begin anytime.

    • The closure options for “opinion-based” and “seeking recommendations” are now obscured by some guiding text. What should that guiding text say? And where should that text link to for further guidance? CMs can support as needed (for example, creating a new help center page that mods can manage). Developers will need to modify the text and links when it’s time to update those.

  • Do the remaining closure options make sense? Should some be updated, or removed? What’s possible right now:

    • For now, these options must be the same for all questions; no variation based on specific tags. If conversations produce a desire to have different options for traditional and open-ended questions, or different options for each specialized open-ended question tag, we’ll discuss the specific needs with you and evaluate those requests as they arise.

    • The reasons or their descriptions in the first level menu (primary close reasons) can be modified by developers if needed.

    • As some of you know, CMs and moderators can manage the options in the second level menu (community-based reasons) as needed, which includes adding new options. We’ll need to ask developers to do any placement or changes to obscuring messages. We’d prefer to obscure a reason for a while before removing it, to ease any transitions.

  • Does “on-topic” mean something different for open-ended questions? Let’s not pretend that there’s ever been universal consensus about the boundary of “about programming or software development”. What does it mean to remain true to the spirit of the site’s focus while allowing for more types of questions?

  • In a world with more open-ended questions, should duplicates be handled differently?

  • What automated processes need to be adjusted, reconsidered or stopped?

Spevacus, JNat, and Frog will be starting some conversations on these topics, but we encourage you to start some on your own as well.

These are just some starter ideas. Most of these topics have been written about extensively. Let’s pull from and reference those older conversations, but let’s not rely on years-old words to do the work for us – instead, we invite you to consider them critically in light of the current landscape, and whether they stand to scrutiny given the direction set above. Make new posts that focus on the current challenges and what the future can be.

If someone wants to keep a list of relevant NEW conversations somewhere, feel free to add that to the bottom of this post. Or use other methods to keep those discoverable. Perhaps they could be organized with some new tags.

Please post feedback for this post as answers, to better facilitate discussion.

16
  • 57
    You have resources to create a ChatGPT wrapper that nobody asked for (so.AI) and a moltbook / cq clone that nobody asked for (SOfA). But no resources to make anything but "modest changes" to the platform. OK I guess. Commented Jul 9 at 15:41
  • 14
    My appetite for curating anything goes questions is extremely low. And this post is quite long. Maybe a summary could be added. Commented Jul 9 at 16:02
  • 5
    @NoDataDumpNoContribution The whole point of this post is asking community to help with defining "what goes and what does not". Commented Jul 9 at 17:35
  • 10
    @DalijaPrasnikar Yes but there is always a price tag attached. Somebody will have to do the work. I've seen plenty of absolutely useless open-ended questions in the last months. I don't want to take care of them. Of course you can if you want. Commented Jul 9 at 18:51
  • 1
    @NoDataDumpNoContribution I don't know how the migration process will look like, but I think we will clean up or not migrate useless stuff. Commented Jul 9 at 18:54
  • 2
    @It's not only the migration, it's the whole format. I think it requires a disproportional effort to curate also in the future and while I don't want to do that myself, I'm absolutely fine if others do it. Commented Jul 9 at 18:56
  • I think I know the answer but why is the non-featured AMA so hidden in this post? Commented Jul 10 at 9:18
  • 39
    I already gave lots of constructive feedback on this in November which was completely ignored. Over and over it goes: the company asking for feedback then ignore everything said. Not just some of it but everything. Every time. Happens around 20 times per year. I'm done with wasting my time. If you want feedback for me from now on, I'll charge by the hour. Commented Jul 10 at 12:47
  • 2
    @Lundin - That’s because they don’t agree with your feedback, easier to confirm what the community wants, if you just limit what you listen to only the users that contain within the eye of the storm. Commented Jul 11 at 1:01
  • 2
    @SecurityHound or if you (as in, the company) just listen to the "friendly on-our-side collaborative mods" and ignore everyone else who doesn't support your view or doesn't actively work to protect your posts from negative feedback. Yep, different treatment for mods with some mods being "more equal than others" (just to quote Animal Farm) based on how "supportive" they are is something I have been feeling for quite a bit now. Mods who don't support this will just be ignored and their concern dismissed as "not existing". Commented Jul 13 at 8:09
  • 4
    @Lundin We keep saying we are done with this site. Happens around 20 times per year. Commented 2 days ago
  • 3
    This post should have come like a month before meta.stackoverflow.com/questions/440088/… Commented 2 days ago
  • 1
    Bert, better late than never; but let me pat myself on the back :) Do you remember Can Discussions be saved post? Now we have API, community editing, downvotes, etc. It won't be all hunky-dory, but with the minimums in place, you'll give yourself (and us) a higher chance of success. Commented 2 days ago
  • "making sure their quality adheres to reasonable standards." I guess we can agree to disagree, Commented yesterday
  • 3
    "Expanding the scope of content on Stack Overflow is the direction we, as a company, have decided to move towards." Thanks for saying so directly, even if it took 400 words to get to the lede. Best of luck with that. Commented yesterday

9 Answers 9

42

The cynic in me sees a lot of euphemisms for the fact that the company has gutted the resources it invests in the future of the sites to a minimum.

Firing a huge amount of people working on the public platform.

Taking away product resources.

Clinging to product decisions but hacking them into the existing systems instead of making the necessary platform adjustments.

All of that tells me that you've been given the impossible task of doing everything with nothing.

That said, I have a huge amount of trust in the team you've left with you, and in Philippe's leadership. If this truly becomes a collaborative work with the community (and you will excuse me if I wait for a bit until there's evidence that that's truly what happens long-term), then maybe there's a path forward.

But I believe there is one vital thing that absolutely has to come from you all, that you cannot delegate to community discussion.

Let me share with you a sentence I wrote in my response to my exit survey when I left Stack Overflow late last year, in response to the question "Is there something we could have done to keep you with Stack Overflow?":

Provide a coherent and credible vision for the future of the company and our products, and show how we're working towards that.

If you want us to follow you, I think you really have to lead with this.

What is the north star? What do you want Stack Overflow to be when it grows up? What does your Stack Overflow offer to the world?

And to be clear, "Same as the old Stack Overflow, but with subjective questions" is not a vision.

And it can't be the answer that the company usually gives, "something something all the world's technologists blah blah". This always comes down to "Stack Overflow is everything for everyone."

You need to give us something that allows each individual to either say "Yes, I'm on board with that, I'm excited!" or "No, that's not a direction I want to go. Good-bye and good luck." -- and you have to embrace both.

This is something that the company struggles with heavily -- conflict averseness and decision avoidance are so ingrained into everything that it's hard to get anyone to say (let alone write publicly) something that might not please everyone.

But the consequence is that you usually just get "meh" at best, and nobody yearns for "meh".

Tell us about your vast and endless sea, and those of us who yearn for it will help you build that ship.

4
  • 4
    I agree with the sentiment, but asking any company founded on community content right now what the answer is in the age of AI is unrealistic. They don't know. They can't know. Commented Jul 12 at 13:23
  • 1
    @SteveBennett But even if nobody knows, decisions must be taken somehow. There must be some gut feeling about what might be possible, even if it turns out to be unrealistic later. For example, OpenAI clearly thinks that we will only talk with LLMs anymore or so. Commented Jul 12 at 14:47
  • 13
    I'd like the company to say that it intends to make the undisputed best human-curated repository of software-related knowledge in the world. If that was our north star, then necessary but uncomfortable decisions could be made for entirely sound reasons. Commented Jul 12 at 14:58
  • They tried that, with the Documentation experiment. It was too broad, even with the larger active community back then. The shift now seems to be in a more forum-dy direction. Commented yesterday
31

If I understand correctly, opinion-based questions will be converted to use the standard Q&A format. I assume that means the replies will become answers, and we will be able to vote on them, just like normal Q&A answers.

However, some of those replies aren't actually answers. They're more like comments. Some of those comments relate to the question itself, some relate to answer-like replies, and some are in response to other comment-like replies.

How is the migration going to handle that? Will the community be flagging comment-like replies, requesting them to be converted to comments?

3
25

What makes a high quality open-ended question?

We already have guidelines on how to write “constructive subjective” questions in the Help Center. Askers should (always have been /) be linked to that appropriately in the Ask Question flow, or it should just be automatically shown to any participant in Q&A that are “opinion-based” now that the meta tag exists (for better or for worse).

As for asking for tooling recommendations, guidance on MSO is to ask how to do the thing you want to do (and be clear about what you want, which is just generally true of any good question), and let answerers recommend a tool if they think it’s truly an appropriate solution to the problem. In fact, guidance on how to handle questions that ask for resource recommendations is to remove the resource recommendation request if the question already asks how to get what they actually want.

I think these sets of guidance work, for the most part. If anything, the issue I see is that people are not aware of them enough, and that goes equally well for reviewers / close-voters / curators as for askers. I don't see an immediate reason why these sets of guidance should change. What I do think needs changing is their visibility.

I think I've mentioned these things in some shape or form in every single interaction I've had on posts about this subject. It frustrates me because (in my eyes), the way to get the benefits of these existing guidelines is through changes to site UI/UX, and we can't make those changes. Only the company can. So I guess here's to hoping that it happens.


Want data to guide discussion and decision-making? If you can’t pull it yourselves, let us know what you want to see and we’ll do our best.

I’m pretty sure requests have already been made to have data for this content in SEDE (Make Discussions posts available in SEDE ? *waves to myself from over two years ago). If not, and switching the content to fit the general Q&A won’t automatically result in that, please put it in SEDE.

Is this a good time to mention that the usability of SEDE continues to exist in a degraded and sorry state?


Create some new questions here on Meta and start talking about what’s most important.

What tag/tags should we include to make sure staff on this project “approach” see the discussion?

5
  • Re: What tag/tags -- we want everyone to see the discussions about this so they can participate. That's why I also suggested keeping a list of those at the bottom of this post. Then the tags become more for organization than visibility. Perhaps the tags can be created and adjusted as themes emerge. Commented Jul 9 at 15:57
  • The open-ended questions (created from today on) will be in SEDE like any other Q&A posts, and the migrated ones will be as well (perhaps with some rougher edges). SEDE usability is beyond the scope of what we're discussing here on this post, but always good to mention when we're talking about ease of finding data, and keep the bug reports coming. Re: Discussions: Full disclosure, I have not spoken to any devs about this, but I would love to migrate (some of) those into Q&A as well. Commented Jul 9 at 19:14
  • With the tag-based approach to question categorization, I think this opens up more options for presenting (or linking to) guidance in more places via tag wiki content. The Ask page sidebar is one possible venue. And of course tag wiki content can be updated by the community as guidelines evolve. Commented Jul 9 at 19:20
  • 1
    Existing guidance may be good (and you're right, much of it probably does not need to be substantially rethought), but the delivery system may not be the best. Linking to a focused Help Center article is ideal. Linking to an answer in the middle of a big Q&A discussion? Less ideal - scroll up or down a little and it's tough to keep track of where you landed. Linking to a 16 year old blog post? Nice for readability, but it's hard to feel sure you're getting the most current information. So let's also be talking about where all of this should live. Commented Jul 9 at 19:24
  • when I said "I don't see an immediate reason why these sets of guidance should change. What I do think needs changing is their visibility.", I did not mean to imply that I think those MSO posts / blogposts should be day-to-day user-facing. surely we can do better than that. Commented Jul 9 at 21:30
23

We can make modest changes, purposefully or experimentally. But our capacity to do that is limited, so be thoughtful about the changes you might collectively request.

This is something that we've heard a lot over the years. It's something that makes this specific community particularly frustrated because we're predominantly programmers. We look at things and can see what needs to change to improve the user experience. Many of us have even built our own personal browser plugins or userscripts to modify the site to improve usability. Have you considered open sourcing more of your code so that the community can help maintain and improve the site? There are a lot of people who would much rather submit a patch than a complaint.

3
  • 5
    This will never ever happen...ever. Commented Jul 10 at 18:05
  • 10
    Since it won't happen on Stack Overflow, why not try a place like Codidact, where it already happens? Commented Jul 11 at 18:34
  • 1
    @KarlKnechtel maybe not codidact because it is even more dead than SO, and basically empty with less than 1200 sofware development questions after 6 years of existence. Also, from my (external and not in-depth) PoV they also made some rather questionable governance decisions, and language decisions (c# and ruby, ew), not to mention visual-design decisions. Nobody can predict the future, but I'm pretty sure even if all of SE disappears immediately, codidact will not be what will fill the void. Commented Jul 13 at 10:39
17

When using these meta tags, users should be required to add a second non-meta tag, with the purpose of being a hint toward these posts needing to still be about the topics this site is for.

12
  • 2
    To avoid doing a conditional requirement for open-ended questions, what if we require at least two tags on all questions? Commented Jul 9 at 19:36
  • 4
    @Berthold you (in the general sense) want the change, you do the work to make it happen. Not every good SO needs two tags, and I suspect many already have only one. Commented Jul 9 at 19:38
  • 1
    I think that'd be fair. The other option i was thinking of would be a lot more... change. (adding a meta tag for how to, for troubleshooting, debugging, etc and using the logic meta uses.) Commented Jul 9 at 19:38
  • 2
    However.... long-term, if we're going to move forward with meta tags we should probably jump all the way in. Commented Jul 9 at 19:44
  • 3
    I wonder if we can not call them meta tags? That term has some baggage (plus it's also used another way) and I don't think this implementation of "categorizing tags" is necessarily the same... or at least it doesn't have to be. For one, they are tied to the list on the Ask page so these can't just be brought into being casually. Supertag? Hypertag? (Credit to Wizzwizz4 for that one) Commented Jul 9 at 20:19
  • 1
    question-type tags would directly describe what we're using them for, that or intent-tags. Commented Jul 9 at 20:23
  • 4
    @Berthold you can put lipstick on that pig but it will stay a pig. "Best practices", for example, is only "super" or "hyper" in the sense of being super-hyper-unsuited for any serious knowledge base, as it's the very definition of a meta tag. To quote Jeff, The very utterance of the words “best practice” should set off warning klaxons, sirens, and flashing red lights. At that moment, you’ve left the realm of opinion and advice. You’ve entered the realm of evangelism and true believers. All the arguments against best-practices still apply and you have not put forward anything refuting them. Commented Jul 10 at 8:19
  • 3
    Re simply requiring all questions to have two tags because of the meta tags: that's not a solution, that's a horrible hacky workaround, only to save a tiny bit of dev work. The meta tags are already special-cased as they're rendered differently; adding a bit of validation logic for asking+editing (what you call 'conditional requirement', aka writing an if statement) should be trivial for a decent dev team. Not doing so would cause many headaches, e.g. when editing old single-tag questions. If you're seriously considering the hacky workaround, the situation at SE must be dire indeed. Commented Jul 10 at 8:27
  • 1
    "Supertag? Hypertag?" Yuck and blargh, @Berthold , I dislike both. It's like when F1 put the word "super" in front of "clipping" to make it sound like "[super-]clipping" is a good thing (it's not, it's terrible for the sport). I'm not saying that these tags are bad but tacking "super" in front on something doesn't make it better™. If Stack Inc is so opposed to "Meta" tag, then why not just call them for what they are "Question type" tags. Commented Jul 10 at 11:12
  • related: Open-ended questions should require at least 2 tags - The type and one other Commented Jul 10 at 18:06
  • @l4mpi - So Spaces vs Tabs (R.I.P Silicon Valley). Commented Jul 11 at 1:05
  • 1
    @Berthold They are meta-tags, in that they describe the question post itself rather than the topic of the question. Though you/folks can always come up with a more specific term to refer to this particular type of tag. Commented Jul 12 at 22:30
16

Skippable premise

I am pretty sure you are quite overestimating the amount of users that consider talking to what to them looks like a deaf wall a worth use of their time. And that is even without considering that the wall (IMHO) is often perceived as condescending ("What if I don't think the scope of content allowed on Stack Overflow should be broadened? But thou must!"), not-so-honest ("Don't worry folks, we heard you on the dump and SOON(tm) we will improve the process"), constantly abusing the same standard reply to avoid any inconvenient question that comes up (the standard "can't talk about this yet", "no finalized plan yet" etc that IMHO aren't simply believable anymore - all clues point to the plan been finalized and set in stone for months if not years) and focused on what most older user see as a destructive plan that you don't look interested in discussing or changing.

Multiple people already pointed out how they already gave feedback about this before and how they already told you it won't work, at least not with the simple implementation you are planning. Some of them by now think they have been ignored before and therefore assume they will be ignored again.. so why would they waste their time again to write the same thing? It is not like reaching extreme numbers on negative feedback really seem to work for sending a message or making the company change its plans, so as an user wasting time to reply to these "announcement" posts often seems an exercise in futility. At least imho.

Skippable premise ends here


That said, let me point out for my own pleasure (because like I said I don't think anything I may say write will make you reconsider) that at the very least you should consider splitting Apples from Oranges instead of mixing everything together in the same crate.

IMHO is pretty clear that Prosus wants a forum - or at least a "forum-like open discussion model", and it is pretty clear that most users want a structured knowledge base in Q/A format. And the two things... don't really go side by side. Problem is , they don't even need to.
All it takes is creating a new side area and placing the quora-forums-overflow.com there. This way the two factions - forums enjoyers and the old style curators - will no longer meet at midday in the middle of the town and duel to death... because now there are two separate town that will never meet again. You may argue that this was what "Discussion" was about, and that that experiment didn't produce the level of engagement you hoped, in which case imho you should probably ask yourself what went wrong - was the presentation bad or you overestimated how many people would want to switch from reddit or other similar sites to a competitor platform that sadly lately has been gathering a bad reputation in those same ecosystem you are trying to garter to with your changes? Anyway imho you solution here - keeping Q/A and Discussion mixed - will only make thing worse: mixing apples and oranges together in the same box will annoy people MORE, not LESS. And having those old-fashioned Q/A focused users share the same space will go against the easy gathering of the new kind of users that you want now.

Furthermore, by keeping the two sections separate, you get the option to simply shut down one of the two sides in the future, should the opportunity present itself. And forgive me for assuming this, but part of me still thinks that you would gladly shout down the Q/A support should the forum like approach prove sustainable.

1
  • 2
    +1. Thank you for pointing out the obvious path forward of having the discussion forum on a distinct and separate platform from the Q&A. Commented 2 days ago
12

What is on-topic is a natural outcome of accepting and rejecting questions. We cannot say in advance what is going to work and what isn't. What if I ask a question like this:

If my cat were an android, what language would be best suited to write plugins for it?

Is it looking for advice, tooling recommendations, best practices, or is it off-topic? What if I rewrite it into "How to design firmware in order to make it easily extensible by users"?

We cannot predict without any data what tags are needed. We can only learn from past mistakes. In the past, we had "homework" tag, which proved to be a disaster. We agreed that tags should never be meta-tags. We got rid of them because they turned out to be useless and even harmful to the site.

The reason meta-tags are a problem is that they do not describe the content of the question. They describe some other aspect of the question, like the author's skill level, or the author's motivation for asking it, or generally what "kind" of question it is (poll, how-to, etc.).

What you have done now is you introduced meta-tags as first-level citizens. You brought back the garbage we had thrown out because it was stinking. The blog continues to say:

Meta-tags, like [beginner], [subjective], and [best-practices], are useless by themselves -- they tell you nothing at all about the content of the question. [...] Best practices to whom? Beginner by what criteria? These tags are impossible to define by anything remotely resembling an objective metric.

I blame us, for letting these tags take root early in the history of Stack Overflow. We should have eradicated them early on to set the proper precedent. It's particularly encouraging that we can learn from experiments on the nascent Stack Exchange 2.0 sites. There's no reason these sites need to repeat all the mistakes we made with tagging two years ago -- we can do it better each time for each new community, and feed those improvements back into the entire network.

Do you still take the blame?

The new tags will create a lot of confusion and fighting. Who gets to decide what the correct type is? How much on the quality are we willing to compromise to make it work? Who is going to benefit from all of this?

You haven't thought this through enough. You have no plan for success.

Ask yourself this: why have we been closing open-ended questions instead of tagging them as such?

8
  • 5
    I think this misses the fact that meta has successfully used meta tags for its entire existence. That's a smaller sample size than what's being proposed, for sure, but I wouldn't equate 3-4 category-type tags as equivalent to the [homework] tag. To rephrase the quoted guidance, meta tags are bad if their designation isn't meaningful. [homework] tells the reader nothing whatsoever about what the question contains– [discussion], [support], [feature-request], and [bug], however, very much do. As long as we figure out how to make the new type tags well defined, I think we sidestep most problems. Commented Jul 9 at 16:32
  • 8
    You're quoting something from 2010. I'll quote something from 4 hours ago: "Let’s pull from and reference those older conversations, but let’s not rely on years-old words to do the work for us – instead, we invite you to consider them critically in light of the current landscape, and whether they stand to scrutiny given the direction set above. Make new posts that focus on the current challenges and what the future can be." Commented Jul 9 at 19:30
  • 10
    @Berthold they're both just things someone wrote: there's no particular reason yours should be given more weight because you did so today (quite the opposite, in fact). Commented Jul 9 at 20:06
  • 9
    @jonrsharpe It's not about when I wrote it, the point is what I wrote. Things from 2010 don't need to be treated as holy words or eternal truths if they no longer apply. We learn from the past but must acknowledge that we are in a different world today. Commented Jul 9 at 20:24
  • 8
    @Berthold Well, a lot of folks think these words do apply still. Folks do consider them critically in light of the current landscape, and whether they stand to scrutiny given the direction set above – and find that, yes, they do. Commented Jul 9 at 21:01
  • 12
    @Berthold Think of it like a scientific paper. The quoted text describes the result of an experiment: we tried X, and it resulted in problems Y and Z. It's a statement of fact based on real data. There's nothing wrong with arguing against that statement, but the burden of proof lies with the person doing so. You're going up against years of hard data and personal experience. At a minimum, you'll have to explain how this "different world" has somehow made these tags no longer confusing and arbitrary. Facts are "eternal truths", by definition. Commented Jul 10 at 2:22
  • 2
    If a tag can actually change the UI and/or moderation and answering criteria for a question, it is special. Unlike in the past with tags like [homework]. Think of it as something other than a normal tag, just stuffed into the tag UI slot. I'm cautiously optimistic that this is the best way forward, or at least a reasonable way. (Especially if SO corporate insists that scrapping OBQs entirely is not an option, which I guess is what this answer is really arguing for.) Commented Jul 13 at 1:28
  • @bta To take a crack at defense, the "facts" at issue here are not quantitative data– they're qualitative study observations. Still empirical, still valuable, but what's being disputed isn't numerical measurement, it's community impact. And at its very core, community management is surely not a hard science, nor is it static or unchanging, especially over the 15 years since those tomes were etched... to say "policy changes should be data-driven" is surely good, full agree– but to treat 15 year old experimental choices (Stack was in early, early days then) as gospel is also folly, I think. Commented 2 days ago
12

Whenever I use the word "you", I mean the company of course.

Right now, Stack Overflow is mocked. Hated. Obsolete in the eyes of many. People have no reason to go there when they can go to Reddit and ChatGPT. It's kind of crazy how quickly that shift happened, but I think it can be explained away by there essentially being two camps that really helped old Stack Overflow thrive and grow. Those that want to do their due diligence... and those that were forced to do so because there was Stack Overflow and not much else. The latter is no longer a target audience and the former turns out to be a frightfully small group of people.

So I wonder what makes Reddit so different. As it is popular as ever even though it has its own fair share of controversy.

It is not a knowledge base; pretty much the Achillis heel I would say. A knowledge base is a very closed ecosystem; many checks need to pass before content is allowed in. That is in stark contrast to a goal of having more engagement. You want an open ecosystem. Reddit open... Stack Overflow closed.

Subreddits exist and each subreddit has its own rules and... rulers. This is a very powerful aspect of Reddit I would say. Yes there are plenty of subreddits that are absolutely unpleasant to say the least and they are being run by people with a power fetish, but there are many more subreddits which are wonderful because they are run by and visited by good people. Each subreddit having its own rules is strong as it allows them to be tailored to the content and the visitors.

Let's pretend for a moment that language tags are subreddits. There is just the one C++ tag. The C++ tag is very mature; well-established. People frequenting it set the bar quite high for new questions. You had better do your homework well and not get your tags messed up. In Reddit, there would probably be two subreddits. One for hardcore C++ programmers with strict rules and one with more lax rules aimed at casual C++ questions, and the occasional C question pretending to be a C++ question. And the odd question masquerading to be a generic C++ question but it's actually very specific to Microsoft. And the occasional shitpost about how object-oriented programming is for morons. It's all good.

On Stack Overflow... good luck soldier, it's not all good. The one C++ tag is all you get. And so you have to face the one quality standard.

Reddit is more geared towards allowing fun. Subreddits have their set topics for sure, but off-topic things are not shunned unless the rules strictly prohibit it. It can't be all work and no play. Certainly not in this day and age where there is so much misery going around, people flock to places that are more lighthearted in nature. Stack Overflow isn't that, at all. Even though that is by design and for good reasons, it can't be denied that it makes engagement suffer.

Reddit is a bubble site. People live in bubbles nowadays. They enter one, do their thing, get out again and then forget everything that has happened inside of it to do exactly the same things, make exactly the same arguments, have exactly the same fights, set exactly the same fires and make exactly the same mistakes again the next day. Stack Overflow on the other hand, is one large perpetual bubble. It wants you to enter and stay there forever, remembering everything, searching for everything, keeping everything up to date. Not what people are looking for. The bubble must pop.

That's kind of the thing isn't it? The goal is to be able to have the micro bubbles exist next to that one large perpetual bubble, without them sticking together like gum in hair. I have to be honest... I don't see that ever working. It needs to be a different site. That's in your benefit as well, as "Stack Overflow" is not exactly going to be selling tickets anymore. You need DidNotReddit Overflow.

4
  • 6
    "It's kind of crazy how quickly that shift happened" — I don't think it did happen quickly, actually. I think that as the post-COVID bump disappeared and activity continued to fall of as it has since 2014, it became easier and easier for content creators to find something to mock; then this reached a tipping point, where resurrecting long-standing patterns of hatred for SO became memetic. Commented Jul 11 at 18:30
  • 9
    As for Reddit, it continues to be popular because it's the same kind of algorithmic hellhole as all the other social media. Commented Jul 11 at 18:32
  • 8
    And the problem with Reddit is that if I search for something on Google and it gives me links to Reddit, everything older than 6 months is probably no longer relevant. Reddit thrives on new posts and quickly forgets the old. It is not a knowledge base, it's a forum, a kind of social media where people go to chat about recent events. Commented Jul 13 at 10:59
  • 3
    I think something else worth noting is just simple network effects– Reddit keeps users because it has users; everyone "hates" Reddit, yet everyone uses/ references it; the network effect flywheel in action. Stack used to have that too... we had enough experts that people were willing to endure an often abrasive onboarding experience to join the community. Now that we don't, some of those problems that have been around for a very long time are more prominent, even if they didn't "get worse", simply because the reward side of the risk/reward ratio for engaging has diminished. Commented 2 days ago
11

We can make modest changes, purposefully or experimentally. But our capacity to do that is limited, so be thoughtful about the changes you might collectively request.

You (the company) had the resources to implement a terrible Chat-GPT-Clone to "answer" question.
You had the resources to create and tinker with the awful new "beta" site which EVERYBODY told you - in advance - was awful.
You had the resources to create the open-ended questions, with the thread-style comment system.
You had the resources for all kinds of things the community - loudly and repeatedly - told you were bad ideas.

But for changes the community would suggest... there suddenly are not much resources available. But then again... you'll have to convince me that this time you'll actually listen to the community, because all the goodwill and good-faith we had has been long depleted.

The closure options for “opinion-based” and “seeking recommendations” are now obscured by some guiding text.
Do the remaining closure options make sense?
Does “on-topic” mean something different for open-ended questions?

Well, here we are back with "closure is bad, let's remove closure". If you insist that “opinion-based” and “seeking recommendations” are now here to stay and to be considered on topic... There are still a lot of things that are definitely not on-topic and not a question that should be asked here, and are now steadily flooding in disguised as "open ended questions" This once more feels like the prelude to removing moderation for the sake of "increasing activity".
Let me be frank here - as with the LLM fiasco a a few years back (which took a moderator strike to get you to see sense) I can't help but feel Deja-vu and feel that the goals of the company and the community have utterly diverged. It feels like, while the community wants to have a archive of help and knowledge, the company primarily desires high activity / engagement, short term gains even at the cost of the long-term (i.e. eating the seed corn).
I've said it before, and I'll repeat it again - the faith and goodwill the community had for StackOverflow has been drained and eroded by your actions, by not listening to the community, by acting inversely to the community's desires. If you want to rebuild and restore StackOverflow, and turn around this trend of decline, you need to show the community that StackOverflow is worth preserving - with deeds, not words.

1
  • How can you be sure that doing something differently would turn around the trend of decline? This might be a bit speculative. Does it just mean you think it's a bad idea? Commented 8 hours ago

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.