Recently, YouTube removed the public display of dislikes on videos (these will be shown to creators privately). Studies showed that displaying the number of dislikes allowed viewer groups to turn the dislike button into a "game with a visible scoreboard". This impacted the outward influence of the creators, but also (in the case of professional YouTubers, for example) impacted their mental health. Making the dislike number publicly invisible reduced the number of coordinated attacks on users. It is clear that YouTube sought to ease the concerns of content creators with this move. Note that likes are still shown publicly.
However, why were the dislikes shown publicly in the first place? Couldn't it always have been privately shared with the creators only? In an age where space on people's desktops and phone screens matters more than ever for companies (literally, it's the amount of space they're given to advertise themselves!), telling everybody visiting the video the number of dislikes it has must have some purpose, right?
The answer is that viewers decided whether or not they must spend their (precious) time watching that video, by viewing the dislikes. Their public viewing allowed users to stay away from content that others told them not to watch by disliking the video. The display of dislikes served as a public moderation tool.
To summarize briefly, a public show of dislikes generally affect only the content creator (and their ultra well-wishers), but to a relatively large degree. On the other hand, hiding the same affects the public's perception of the video, but it affects each user to a small degree. Both these effects must be acknowledged, and YouTube definitely sided with the creators when they removed the public viewing of dislikes.
The above is a small snippet of what I refer to as the visitor-creator balance.
Websites such as MSE, Quora, Reddit etc. allow users to interact with them as both creators and visitors. For example, if you answer a question or ask a question on such a website, then you're responsible for the creation of content. On the other hand, if you are viewing content that is already present on such a website, then you are a visitor.
Typically, websites model themselves very subtly based on what kind of creators and visitors they can expect. The reason for this is obvious : optimizing your website for a particular kind of audience helps increase the traffic to your website, particularly if such an audience can be found in large numbers.
For example, take Quora. While they are very much an all-purpose website, they prioritize curiosity and helpfulness. Any curious person can pose a question, and any knowledgeable user can provide an answer. The question and answer pair will stay on (unless it violates content policies on Quora, which unlikely to occur if the question is purely mathematical, for example). However, the priority is "write a question, get an answer".
Talk to people or read the posts of popular users, and one sees that there is an air of relaxation and comfort in the way that people contribute to the site. The same applies to Reddit. If you're curious, you ask a question. If you wish to be helpful, go out and answer any question that you can. Discussions, while somewhat hidden on Quora for example, are often the norm as well. Bounded only by some basic social and cultural norms, one is free to say anything they wish.
Mathematics Stack Exchange, on the other hand, is all about content. If I were to use two words to describe MSE that instantly differentiate it from the other websites, they would be reliability and organization. Reliability refers to the accuracy with which questions are answered, and organization refers to the accuracy with which questions are ranked, primarily for the purpose of searchability. Both of these characteristics tilt the visitor-creator balance towards the visitor.
Indeed, think about it. Reliable answers can only be given to questions that admit reliable answers! Therefore, opinion-based questions (comparing handwriting styles of mathematicians or mathematical memes, say) are not allowed. That restricts what creators can ask, or what questions they can answer. On the other hand, visitors are given crisp answers to reliable questions that they can directly apply to their research or work.
Similarly, organization prevents creators from asking and answering duplicates. In some sense, answering duplicates doesn't seem like creation because it is tantamount to repeating what has already been said before by someone else, but that does count as creation in Quora or Reddit. It doesn't in MSE, at least in principle.
The idea of "asking a good question" ensures that creators come to MSE not for help, satisfaction of curiosity, or assessing opinions, but as a "last resort" : they've done their research, isolated a specific problem, have some motivation or effort to show, and can ask their question if it hasn't been answered by going through those preliminary steps already. Alternately, they've seen a question, noted that it's not a duplicate, written an answer, double-checked it to ensure for correctness and on-topicality (occasionally for notation, legibility etc.), and then posted it.
So, creators need to put in much more effort on MSE than they need to on Quora or Reddit. On the other hand, all that work benefits future visitors : searching is much easier for them , answers are reliable, informative and to-the-point, while questions are focused and often have a source/motivation attached, so they're often exactly what you're asking and not something slightly different (at least for most questions at the undergraduate level, this is already the case) and tend to offer more than just an answer : an opportunity for visitors to explore the subject further.
TLDR; Quora and Reddit prioritize creators over visitors. MSE prioritizes visitors over creators. Another way of looking at it is that MSE prioritizes the correctness and structure of content, while Quora/Reddit prioritize the comfort, curiosity, and helpful nature of contributors.
Every platform would love to cater to both creators and visitors : that is a simple consequence of utilitarian thinking. However, as MSE and Quora have found, it's simply not possible, at a large scale, to keep both sides happy.
Here are some basic facets of MSE that illustrate why both cannot be completely happy together(as long as we have just one website for all of them!) :
Solution-verification : Solution-verification questions have askers that write attempted proofs for questions, and ask whether the attempted proof is correct or not. Their benefit to the asker is extremely high, because proof-writing is a critical skill in mathematics. Furthermore, because the asker typically makes an effort to write a proof whose verification is asked for, such questions possess a "honesty" certificate (evidence that the author hasn't just copy-pasted a question onto the website and gone for a vacation, letting someone else do the work they're supposed to), and therefore are often seen as worthy of an answer.
However, these questions are "localized" : they're useful only to the asker because attempted proofs are rarely identical to the T, and therefore it is difficult to convince askers that another proof attempt is identical to theirs. That clogs up search results for future visitors.
Hints versus complete answers : If you are writing an answer, then pedagogically speaking a complete answer is not often the best way to help the asker learn. Hints make them think. Typically, hints as answers will be tailored to the asker's context/ability, and while they're extremely useful if written well, it's also really difficult to show and hide just the right amount of details.
However, it's obvious that hints may not favour future visitors, who will need to reproduce the missing information themselves from the hint before using it in their work. Complete answers clearly favour visitors in that sense.
Voting design: We already spoke about it above, but to summarize : creators would much rather be told what they're doing right and wrong so that they can improve and feel comfortable with their contributions. Visitors would wish to see curated information that they can directly use in their work.
Merely from a scale and software point of view, both being completely satisfied is something that's untenable. The public show of votes, along with the anonymization of such votes clearly shows that creators aren't necessarily going to be told what they're doing wrong, but visitors are going to receive opinions from a larger sample of people : it's easier for them to locate information. Even if this weren't the case, there's way too many questions for each contributor to receive personalized feedback.
Lack of discussion : Discussion is discouraged in MSE : all important information should be present in the post in some form or another, and discussions can occur in chat. This means that creators will have to clarify their question away from its location, so that visitors are able to access answers without needing to look at potentially extraneous comments. It also means, in conjunction with the voting paradigm, that unclear questions may be downvoted and kept out of sight of potential answerers before it can be voted back into attention following edits. Creators will feel the need to get things right pretty much the first time they post.
New users : There's a whole literature on this in MSE, but new users seem to have a lot of complaints regarding their treatment by users of the website. As much as one can say "Read the FAQ", "Here's how to ask a good question" etc. in the comments, this comes down to various factors : a clash of expectations, scale, communication styles etc. However, if visitors are to be favoured, then there simply isn't enough time to talk each newbie into improvement, because the focus isn't on the newbie but on the content they're producing, and unfortunately the average newbie isn't really having anything "new" to contribute to the site in terms of content. Tipping the balance towards visitors and their content will cause strain on new users to product from minute one.
That's just a few things, really. There's so much more beneath the surface, such as the existing moderation system and the avoidance of salutations and congratulations in threads. Also note that performing a quantitative analysis is a possibility only for extremely restrictive measures of content such as votes, close votes, number of views etc., but performing a qualitative analysis, such as analysing attitudes of creators towards closure, or how much visitors are helped by closed questions, is often difficult. That's really what makes this debate so lively despite the fact that, by design, MSE aligns heavily with visitors.
The way that the creator-visitor debate has played out on MSE thus far, it's almost been a combination of everything. Anyone reading discussions on the meta site will see colourful opinions, thunderous back and forth accusations, clashes of vision that are incompatible to the point of no return, descents into compromise, and very informative, statistical debate.
The point is that policy isn't really going to do much. There's that much opinion. Any policy will have many detractors. Therefore, perhaps policy isn't the way to go about it. As usual, the way to go about it is to be holistic (yet again, so bear with me!).
Thinking like a visitor and like a creator at the same time is the way forward. If I am visiting a question, what do I want from its answer, or what is my context? If I am reliably asking a particular question, what is most likely the thing I'm looking for? If I'm reliably answering this question, what's the most likely thing I'll need to do that? An amalgamation of these answers will tell you what to do.
It's really simple : we cannot focus merely on content, because that comes from creators. We cannot merely focus on creators, because the visitors come for the content! So think from both points of view, and suddenly you're doing far better. The only difficulty is scale : what if this means that I'm spending 5 minutes analysing a question when I could be answering another one?
Well, your five minutes spent on curating a question saves five minutes for all the future visitors who decided to interact with that question (or not) based on your assessment of it, rather than waste 10 seconds each to look at the question and make a decision. Creators see such questions and know whether it's fit for the site or not. Anyone taking the time to be careful is actually helping others, but in a manner so silent that it's almost, well, unacknowledged.
If that's not a strong enough argument then nothing is, really. It's difficult to argue that you are doing yourself a service by being careful : after all, there's no reputation to be gained, and you're not getting any mathematical pleasure out of thinking from various point of view, are you? That, basically, is the reason why this practice isn't done enough : people would much rather come for the math than to sort out the visitor-creator balance (one can argue that's actually a good thing, but not in excess!).
Accepting that the creator-visitor balance manifests itself in MSE is a giant step in helping improve the experiences of interactors with the website. Thinking from both points of view, and communicating strongly and effectively via moderation and comment actions can help us bridge divides and foster empathy for the other side.
However, it's obvious that this can only be bridged so much, just like a heat engine that cannot, (even theoretically! Check out Carnot's engine) have 100 percent efficiency. We must listen to our creators and to our visitors : let's help them interact with MSE better. That's how we make the internet a better place, right?