4
$\begingroup$

Prior note

While in most cases a proposal's score is a clear indicator of how much it is accepted, it will be hardly the case here. Indeed, we're talking about new user engagment here, and newcomers -who are the main target of this proposal- barely go to meta, learning how SE works and all. Disengaged users because of this trouble won't vote, either. This proposal unfortunately has to be voted by people who are not the target of this system, so you'll likely have to put yourself in the mindset of a new user to review this proposal effectively.

It does not mean your vote is not important. It means that you should perhaps not use this statistic alone as your reason to think the proposal could be useful or not. Alternative elements to judge this proposal could be checking how many new users come back and try to edit after having asked a bad question, their reaction, their ability to improve, etc.. Alas, I don't have access to such statistics, so it's impossible for me to know the current state of user engagment. We also cannot know the true effectiveness of the proposal until some tests are done, for instance with A/B group testing. So please be careful with that!

For a clearer differentiation in how the proposal is accepted vs its innate qualities, I have given a simple "yes, I agree with your proposal in general" answer you can also upvote/downvote as you wish. Downvoting it means that in overall you explicitly disagree with this proposal. Upvoting means you explicitly agree with it. The meaning of votes on the question itself here stays as usual and can also indicate your (dis)agreement. Remember to add your own answers/comments if you want to discuss some specific points!

What is the proposal?

In one sentence : When newcomers receive a series of downvotes on a question or answer, send them a "tutorial" notification to help them find the reasons they have been downvoted.

How does it look like?

Here comes my excellent drawing skills when making mockups, behold!

Mock-up of dubious quality showing how the proposal would look like for questions.

For people having trouble reading the picture, the text appears before and reads :

Help : You have received downvotes recently on "this question/answer" (hyperlink to the question/answer that received downvotes). You can check the comments under your question or the "tour" (link to the tour) and "help-center" (link to the help-center, page "how to ask a good question", or how to write a good answer for answers) page in order to know why it has been downvoted and how to improve it.

The text is a placeholder and can (I'd say even should!) be refined. I'm not a native English speaker, so wording is likely to be different, at the very least. But that should give you the gist of what to expect.

Who does it apply to?

It applies only to users who just registered an account and received their first X downvotes on the same question. The amount "X" of downvotes required to trigger this notification can be debated, but I'm setting for a big 3 downvotes for now. Also, the question should have a negative score for this to happen.

It happens only once per user. Once a new user receives this notification, it will never reappear again.

This is not retroactive : Existing users are not affected by this change. This also does not apply to established users, ie. high-score users on one site joining another SE website who are already starting at 100 score. By extension we could also consider not sending it if the user has managed to score enough points before witnessing this event (e.g. : 1 000 points like the established user privilege, I think? The page doesn't seem very up-to-date...).

What are the strengths of this proposal? Why do I make it?

  • It helps solving an issue some new users have : Having access to resources as to why their question have been downvoted.
    • There's currently no direct link between the score you lose and any help-center explanation. There's no dynamic, tutorial-like feedback explaining why you lost points, and it can lead to feeling frustration. It is currently a task for the reviewers to do, and as seen before, people are not exactly too kind to force that behavior.
    • So instead of forcing the voter to do something they don't want to, we use the notification's red color to catch the attention of the user and bring some basic information. This on-top of being given when they really need it give an actual dynamic tutorial explaining the reasons behind downvotes.
    • Plus, redundancy in teaching the basics to someone having troubles doesn't really hurt!
  • This is entirely invisible to the reviewer. One of the key (and most protected) point of voting is that it is frictionless : A click and you're done. This system doesn't add any click to the voter, and in fact, doesn't impact them at all.
  • It is anonymous : Another keypoint of not changing votes, the downvoters' identity is not revealed.
  • This is also very lightweight to the user receiving the notification, since it only happens once and is not a screaming popup. It also (usually) happens when they receive comments, so the notification mark would often also be here anyway.
  • My intuition tells me it's not too hard to make technically. We add a new message on specific, well defined triggers using an existing, normally well-tested system. The graphics don't change at all, so it's mostly code-stuff.

What are the potential weakness of this proposal as you see it?

Let's be thorough and transparent, there are some potential disadvantages to this system. I'm revealing you what my skills intuitively tell me, so you have all information to make an informed decision. Still, I wouldn't have made this proposal if the strengths didn't overcome the drawbacks from my viewpoint :

  • Tutorial "sensory overload" : If there are too many elements displayed to the user at once, they risk getting confused. The proposal is supposed to be lightweight... But we never know!
  • At most it tells "why" their question is downvoted. It doesn't directly tell "how" to solve it.
  • This will not solve every problem : Some users will care and use the notification to improve themselves, others won't and keep asking/answering poorly.
  • It might annoy... How do we call them already? Ah yes, sockpuppet accounts, ie. experienced users who create new accounts on top of their main one. It is indeed more information they don't really need if they are legit, allowed sockpuppets.
  • I don't think there is already a system like this in place... But I have the memory of a goldfish and I don't recall having ever been in that specific, not so good situation (even for my admittedly could-be-improved first question). Remaking the same system could be a no-no.

Why are you not proposing this to upvotes too?

I'm raising this point because some (a lot of?) people, both on worldbuilding SE and meta's meta think that downvoting is "just" the opposite of upvoting, and therefore are close-enough that they should be treated the same way. As a more or less experienced... User Experience designer I heavily, but really heavily disagree with this.

I will try to keep it short as not to bother you too much : In game design (recall the voting system takes heavily from video-game systems), the process of giving rewards -like giving points- and punishments -like removing points- are treated differently : Rewards are usually self-explanatory in what the player (here user) should do next time. Indeed, they are informed they did well, so if they want to perform well next time, they have to do about the same kind of thing. In contrast, punishments happen when you did something wrong : There are improvements to be made. A good, well-designed game system tells you not only when you did something wrong, but also why, and better yet, how to improve. That's a key difference with rewards, which only need explanation if what you did to get rewarded is not obvious. Moreover, the process of telling you failed (to meet the reviewer's expectation here) is a double-edged sword : it can help someone improve as well as it can be discouraging, especially when you're new at something. Therefore it often has to be made more carefully, unlike rewards which can be given more freely with generally less negative consequences.

Besides the above point, it is to stay focused on one small change at a time. The more focused it is, the less likely people will refuse it just because they disagree with half of the proposal (namely the upvoting or downvoting part). If you believe it could see some purpose to reward new users with well-voted questions/answers, you could perhaps tell in another post how you envision it (more medals? Specific notification? Pop-up?...). I wouldn't personally mind upvoting a proposal like this.

$\endgroup$
11
  • 1
    $\begingroup$ Note : If this receives an overall approval here, I will post it to meta's meta. I prefer getting a smaller community's feedbacks I know before going up. $\endgroup$ Oct 27, 2023 at 22:59
  • $\begingroup$ You got it right that the way people interpret information is critical, specifically noobs to SE. :-) Votes in this case are what we might call a domain language, and would be unfamiliar to someone who has not seen it before. $\endgroup$
    – Nolo
    Oct 29, 2023 at 15:44
  • $\begingroup$ The ability to restrain up votes, rather than using down votes, is probably better. I would fear that humans would forget that tigers are bad and can eat us, but that's a separate problem. :D $\endgroup$
    – Nolo
    Oct 29, 2023 at 15:51
  • $\begingroup$ On the other hand, showering someone with gifts is likely not the best way either. A good speaker will punctuate the lecture with decent jokes, for example. $\endgroup$
    – Nolo
    Oct 29, 2023 at 16:04
  • $\begingroup$ In other words, "make note of this point that came up 10 minutes ago in this obscene or obscure way because that's fun" and it will get you to the next important point that comes up 10 minutes later to help you stich it all together and start thinking about the problem more effectively. So a definite feedback loop seems to be at play. How do we do it? $\endgroup$
    – Nolo
    Oct 29, 2023 at 16:11
  • $\begingroup$ On the negative end, we have attention hijacking, which is no less a form of attention utilization. We can do better than using attention to achieve our own goals and aim it toward achieving the posters goals. The interesting part is that the posters goals and our goals are aligned in this case. $\endgroup$
    – Nolo
    Oct 29, 2023 at 16:17
  • $\begingroup$ You could hypothetically test whether a language model is ready to fulfill this requirement by prompting it to "tell me a joke about the line of reasoning in this question", broken down of course to see if it groks the line of reasoning, maybe sprinkle on some concrete, then tell us a joke. Sort of a litmus test. $\endgroup$
    – Nolo
    Oct 29, 2023 at 16:24
  • $\begingroup$ This is a great idea, but it's a waste to ask about it here as it's a network-wide change (a change to the underlying software). That means it's an enhancement request that must be posted on Meta Stack Exchange. SE hasn't been amenable to new ideas for a long time, but considering between May and October 2023 they laid off about 30% of their workforce, I wouldn't expect much from this. $\endgroup$
    – JBH
    Oct 31, 2023 at 0:01
  • 1
    $\begingroup$ However, I get what you're saying. The Stack has seen a tsunami of new users posting poor questions that get down voted at best and closed at worst. It's really been amazing to watch - but leave it to the latest generation to ignore the rules and complain when they're judged. $\endgroup$
    – JBH
    Oct 31, 2023 at 0:03
  • $\begingroup$ @JBH I'm asking here because there's no point for me if the community I'm involved in the most doesn't agree with this change. Regarding whether it'll get accepted by the SE team, we can't know until we try! Best case it's done, worst case nothing happens and we lost some time writing/reviewing this idea. And I hope it wasn't too much time spent for you ⌛. $\endgroup$ Oct 31, 2023 at 0:15
  • 1
    $\begingroup$ @Tortliena Oh, I agree with it. But I've also worked over at Meta Stack Exchange enough to have a reasonable expectation of success. Nevertheless, if you post over there, let me know and I'll gladly support it. $\endgroup$
    – JBH
    Oct 31, 2023 at 0:27

3 Answers 3

4
$\begingroup$

New users are often unaware of the details of the UI and, as a consequence, have a poor UX.

Adding a notification which ends up in that same UI won't make any good for them.

I still remember when I moved my first steps in SE that I could not understand why at any apparently randombclick I would get a notification, a badge or something else.

Additionally, mods can and do contact users with strikes of poorly received contribution (this is why I say what I say in the first two paragraphs)

$\endgroup$
6
  • $\begingroup$ Still, would it be better to be hidden in a grey button, or to have a new indicator to reach the help? Do you still wish to (or do you) always contact people with strikes of low score? [...] $\endgroup$ Oct 27, 2023 at 23:18
  • $\begingroup$ [...] I know this is not a game-breaking change, but any large addition is ought to be refused, or at least dismissed. That's because we're meddling with a core system. This and the fact a lot of SE employees got fired means I can only ask for small edits. $\endgroup$ Oct 27, 2023 at 23:21
  • 1
    $\begingroup$ This is correct, putting more information in front of the user will not help. Humans equate information to noise when it is not directly relevant to their current task. I've been made aware of studies in this area ( related to process management tools ) through someone have worked and chatted with. $\endgroup$
    – Nolo
    Oct 28, 2023 at 15:35
  • $\begingroup$ @L.Dutch Simple question : When you contact directly someone, how is it done? You send them a mail? $\endgroup$ Oct 29, 2023 at 9:58
  • $\begingroup$ @Nolo Yes, there's noise, but there is contrast, too (the reason the notifications and score changes are coloree while the rest is grey). We also have to recall that most users, after having asked a question, connect back to check what happened to it. That's actually part of their current task, and since both score changes and (usually) notifications are lit up... $\endgroup$ Oct 29, 2023 at 10:02
  • $\begingroup$ @Tortliena I understand why the notification approach is helpful. I voted up the question, but I do not have enough status to actually vote on decisions, 3K is required iirc. My question would be what is the mathematics? Do we have statistics from studies which support that people would use it before we put effort into things? We can reason about why we suspect that they would, but that actually does not tell us anything. $\endgroup$
    – Nolo
    Oct 29, 2023 at 15:42
2
$\begingroup$

I'm thinking more along the lines of bots which can auto comment and also flag moderators.

Please read first and react after reading. :D

I am aware of the controversy. I feel it in my life, so I am not in need of a lecture on this.

The trick question in my question, How to limit the range of answers to questions which are not technically considered opinion based?, was about the easy part and the hard part.

Detecting these kinds of questions is the ridiculously hard part... but maybe the problem is mostly solved or can more easily be solved with a bot that interfaces with a language model.

The easier part is managing the tutorials because we already know a lot about how to do that, but that is also a very, very hard problem. We don't recognize them as they are because we are human and we take these things for granted.

A language model is also probably very capable of helping to better organize our tutorials.

The problem is in how we do our gate keeping and how much effort that requires.

I understand that the philosophy is to reduce both friction and effort, but humans are notoriously bad at doing that. :D

Maybe we shouldn't allow bots to close questions, but there's no reason why bots cannot help to facilitate how and when to interject relevant information about quality control for the purpose of making the site more focused, more searchable.

If you are repulsed by the idea of language models, I get it. But I would have to ask what the difference is between an artist taking inspiration from other art, or a child learning from their environment, and an unconscious tool optimized on human knowledge for the purpose of making output utterances maximally relevant to the input utterances we provide? What makes it more wrong? We do not expect children to ask if it's ok if they listen to our conversations. But a language model is not a child. It is however aimed at actually helping everyone, even if we don't understand how it works or why we are aversive to the idea of it.

One example I can think of off the top of my head is that I can clearly see a feedback loop between Discord and Youtube. And the time window of that loop is at least less than an hour. When I've brought up relatively obscure topics in Discord, things I may have never talked directly about there, i.e. subjects new to me or specific to some context I've never been in before or recently by a long time, miraculously suggestions to videos on these topics show up after a little while when I refresh the Youtube tab.

I don't know the relationship between Discord and Youtube, but I'm very confident there is a feedback loop there. And I find it occasionally useful.

A bot that does this directly is probably much simpler than the above example and since it wouldn't particularly be trying to sell anyone anything, would likely be highly relevant, direct and very useful.

$\endgroup$
8
  • $\begingroup$ I honestly have troubles understanding what you're saying, I guess I'm not exactly used to this kind of wording 😅. So far I get it you're suggesting to add some deep A.I. 🤖💬 (ie. Like ChatGTP or llama.cpp) to analyze the question and to provide an in-depth comment one way or another, is that right? What do you mean by the "hard and easy parts" you intended to ask about in your question? $\endgroup$ Oct 29, 2023 at 10:15
  • $\begingroup$ @Tortliena Probably the difference between US English and European English. Even in the US there are deep dialects which make it difficult for Americans to understand each other, so this is fair. :D My awareness will help me communicate better in the future. What I imply by the "hard and easy" part is about how the reader interprets the question individually, but language barriers are a good example as well. We all think differently. I am asking about something fundamental which addresses this deeper level of variation in thinking. $\endgroup$
    – Nolo
    Oct 29, 2023 at 15:17
  • $\begingroup$ What I suggest is to try to address possible logical issues in questions as well as to try to answer the question. In other words, to help, fundamentally. Of course this is what we try to do, but I think we can do a little better, and I think we can continue to improve, perhaps more efficiently. ;-) $\endgroup$
    – Nolo
    Oct 29, 2023 at 15:20
  • $\begingroup$ In the limit, you could say, if the poster has mastered the art of thinking, then the questions would become boring and the answers would be apparent. But that is an alien concept to all of us. I admit no claim to recognizing what that means. But it serves to reason about asking questions and answering them. $\endgroup$
    – Nolo
    Oct 29, 2023 at 15:27
  • 1
    $\begingroup$ Ok, I think I understand better now. The idea of an automated comment (like when the bot tells your question is unclear, for instance when it lacks question marks or is too short) could be a good idea indeed. The phrasing should be carefully chosen though, since it's public and not personal. The goal is not to throw rocks at people on the public place ^^. $\endgroup$ Oct 30, 2023 at 13:29
  • $\begingroup$ @Tortliena indeed. You mention clarity, but I'd think it's not impossible to look deeper into the framing of the question and to propose suggestions for separating out concepts from details, to help focus on concepts which depend on details - i.e. collectively on the details provided in the question. The way I see it, it's very easy to become distracted in details, which is where the variability in the answers comes from. Not that the variability isn't important, but it simply makes for walls of answer to a question that may benefit from more concision and focus. $\endgroup$
    – Nolo
    Oct 30, 2023 at 18:54
  • $\begingroup$ @Tortliena I referred to this example comment which helped to frame the question more concisely for those who chose to answer. worldbuilding.stackexchange.com/questions/250774/… $\endgroup$
    – Nolo
    Oct 30, 2023 at 18:59
  • $\begingroup$ "If you are repulsed by the idea of language models, I get it." No, i just think that they are in general, incompetent, and perpetually simulating being on hallucinogenic drugs. $\endgroup$
    – Topcode
    Nov 1, 2023 at 2:22
-1
$\begingroup$

Yes, I agree with the proposal

You can upvote this answer if you wish to explicitly tell you like the proposal and wish to see it developped. You can also downvote it if you dislike the proposal and don't wish to see it happen.

Upvoting or downvoting this answer doesn't necessarily has to mean you agree (or disagree) entirely with the proposal. For instance, you can agree for the most of it yet would like to change certain parts, or you could disagree with it because you strongly dislike a key point of it, but like some others which are more minor.

If you have points to debate or want to explore another approach, you can add it as another answer (or as a comment to the question).

$\endgroup$

You must log in to answer this question.

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