Yes, the software definitely needs a close as dup feature, which automatically adds a link to the duplicated question. How that may be used on different sites isn’t the point here, as long as the software supports most of the intended uses.
The dupes would be linked, just as they are currently. Furthermore, the S/O system does not prevent dupes, it merely shuts them down (sometimes). The power users are not compelled in either case to answer the question more than once. In both cases they can simply mark the dupe and link to it.
So @benaston was pondering a feature like that except without even closing the duplicate – maybe call them “similar” oslt instead of “duplicate”.
It is probably site specific. I guess it’s about the design of the “game” rather than the software, per se.
A lot of this is a matter of site culture, and communities here will have to set their own guidelines. SO tried to be a site for expert questions and users there are impatient with beginner questions, but one of my main sites, Mi Yodeya, welcomes questions from beginner to expert. We have to let communities decide what they want to do, and push them to communicate clearly about what that is.
Our “categories” feature would allow a site that’s more expert-focused to also support beginners – add a “getting started” category and encourage people writing their first program (or equivalent for your domain) to ask there. By virtue of it being a different category, users would be “primed” to encourage a different mindset. I suspect that part of what frustrates SO users is that it’s all mixed together, so those who don’t want to see basic questions have to “wade through them” to get to the stuff they want.
3 Likes
Agree. One of the failure modes (if it can be described as that - it’s still a HUGELY successful and useful site!) with S/O was that it was predicated on a central team of super users. This works as long as long as the culture of that team rhymes with that of the important contributors to the site… until it doesn’t. If we could minimise or even eliminate the importance of this central cohort, then this problem might be pre-empted.
You are arguing against the 1% of the 90-9-1 as described in https://www.nngroup.com/articles/participation-inequality/
In most online communities, 90% of the users are lurkers who never contribute, 9% of the users contribute a little, and 1% of the users account for almost all the action. …
All large-scale, multi-users communities and online social networks that rely on users to contribute content or build services share one property: most users don’t participate very much. Often, they simply lurk in the background.
In contrast, a tiny minority of users usually accounts for a disproportionately large amount of the content
And further down:
How to Overcome Participation Inequality
You can’t .
The first step to dealing with participation inequality is to recognize that it will always be with us. It’s existed in every online community and multi-user service that has ever been studied.
There’s also A Group Is Its Own Worst Enemy:
2.) The second thing you have to accept: Members are different than users. A pattern will arise in which there is some group of users that cares more than average about the integrity and success of the group as a whole. And that becomes your core group, Art Kleiner’s phrase for “the group within the group that matters most.”
The core group on Communitree was undifferentiated from the group of random users that came in. They were separate in their own minds, because they knew what they wanted to do, but they couldn’t defend themselves against the other users. But in all successful online communities that I’ve looked at, a core group arises that cares about and gardens effectively. Gardens the environment, to keep it growing, to keep it healthy.
That core group of super users will exist for as long as the community is healthy. SO is currently (in my opinion) making the mistake of disenfranchising the current core group across multiple sites without trying to help another group of that 1% that aligns with their vision.
Trying to make it so that the people who care about the success of the site do not have the tools to properly garden by either restricting the tools, or by giving the tools to everyone.
Now, the software does not always allow the core group to express itself, which is why I say you have to accept this. Because if the software doesn’t allow the core group to express itself, it will invent new ways of doing so.
When the core group lacks the tools to garden / curate / moderate the site, I contend that is when things get rude (as social pressure is the moderation tool of last resort that is the hardest to restrict).
The other alternative is giving everyone the tools and you’ve got something closer to unmoderated usenet or Communitree (also mentioned in A Group):
Communitree was founded on the principles of open access and free dialogue. “Communitree” – the name just says “California in the Seventies.” And the notion was, effectively, throw off structure and new and beautiful patterns will arise.
And, indeed, as anyone who has put discussion software into groups that were previously disconnected has seen, that does happen. Incredible things happen. The early days of Echo, the early days of usenet, the early days of Lucasfilms Habitat, over and over again, you see all this incredible upwelling of people who suddenly are connected in ways they weren’t before.
And then, as time sets in, difficulties emerge. In this case, one of the difficulties was occasioned by the fact that one of the institutions that got hold of some modems was a high school. And who, in 1978, was hanging out in the room with the computer and the modems in it, but the boys of that high school. And the boys weren’t terribly interested in sophisticated adult conversation. They were interested in fart jokes. They were interested in salacious talk. They were interested in running amok and posting four-letter words and nyah-nyah-nyah, all over the bulletin board.
And the adults who had set up Communitree were horrified, and overrun by these students. The place that was founded on open access had too much open access, too much openness. They couldn’t defend themselves against their own users. The place that was founded on free speech had too much freedom. They had no way of saying “No, that’s not the kind of free speech we meant.”
But that was a requirement. In order to defend themselves against being overrun, that was something that they needed to have that they didn’t have, and as a result, they simply shut the site down.
The core group will exist and needs to have the tools to maintain the site with the vision that they have. Failing to provide for that means you’re going to have the community that cares overrun by people who occasionally contribute but don’t care about the success of the site.
7 Likes
Strictly speaking this is not correct, and uses somewhat prejudicial language.
As far as I know, there is no routine, intentional use of Meta (or any public part of the site) to organize downvote brigades. That wouldn’t be allowed. The heightened, focused visibility that Meta turns on a post does tend to create brigade-like behaviors, whether intended or otherwise, even if the Meta post that brings this attention is written by the user the Meta effect is hurting. But this is not coordination: it’s a strong selection effect combined with a considerable priming effect, so quality-conscious users (and almost only those) have their attention drawn, and their critical awareness primed, by the extra details in Meta. They then naturally vote exactly how you would expect, without consulting anyone else. The only ways to prevent this would seem to be either a) forbid talking about specific posts and how to improve them on Meta or b) somehow disable voting on posts linked from a Meta discussion. Both require software support, although the former may actually be possible in Codidact if we push hard on per-question discussion areas.
(There is coordination done to maximize close vote effectiveness, because SO’s close vote queue is out of control; this is not technically done on Meta, but on chat. Not that big a difference, although it does mean that the regulars there are a slightly different group of people. For example, I was very active on Meta SO, but never did much with SOCVR, the chatroom. The need for this would hopefully be much reduced in the long term with better close tools, and in the short term very simply by not having 5000+ Q/day.)
Tolerance for homework and poorly researched questions is a matter of site policy and culture, so for now we mostly need to ensure the site software is flexible enough to be useful in managing a range of possibilities.
Duplicates really aren’t, though. And it’s not about making sure site regulars don’t have to see, or answer, the same question again. (Well, that is a good thing to aim for, but it’s not solved well by having a way to close questions as duplicates. Instead, having better automatic search so that askers more often see the answers they should be looking at, along with better instructions so they have an idea of the sorts of mental adjustments they should be making to apply the existing answers to their situation.)
The value of closing as duplicate is really in maintenance and searching, since the value of a Q&A site is in having expert answers already there so that the vast majority of people with a question never need to ask it. Two hundred answers to the same basic question, with various minor adjustments for various cases, are far harder to check for subtle errors and potential problems, so you end up with large chunks of your Q&A database giving bad advice that opens security holes, instills bad habits in your kids, or exposes you to legal hazards. And needing to read through a sizeable chunk of those answers to be sure you didn’t miss some possible solution wastes the time of later searchers.
It’s good to be focused on making things better for askers, as long as you keep in mind the two different groups of them, and how much larger the group of silent browsers is than actual, visible question posters.
This does bring up a subtler problem. Often, the canonicals aren’t wrong, they’re just poorly specified. Rather than teasing out the subtle root cause of a question by leaving it closed until it’s edited to have unambiguous diagnostic information, an answerer will seize on one particular scenario they’re familiar with that is superficially exactly what the question describes, give a working answer for that scenario, and pat themselves on the back for a job well done. What they don’t realize is that there are in some cases quite a few other scenarios with completely different root causes and exactly the same set of described symptoms. (The numerous upvotes on these answers mostly come from others who, without realizing the differences, also had that same specific scenario.) A very common case of this is toolchain error messages on SO. You can find dozens of very visible, frequently-viewed questions about various specific errors, with dozens of upvoted answers each, all of which sometimes work and sometimes don’t, but they usually make little effort to explain when they would work and when they wouldn’t. So you can find the canonical question about some particular error message in Android SDK, and each of the many answerers is, without realizing it, actually answering a different question.
The only way I know of to fix this is to ruthlessly force these questions to be painstakingly diagnosed until there’s genuinely enough there to confidently determine which of the half-dozen reasons or more for that exact error message is actually applicable. That requires a lot of expertise and a lot of patience, and potentially several rounds of tedious clarification.
Disallowing duplicate closure doesn’t make this better; it actually makes it worse. Now, instead of one question with a few dozen mostly-irrelevant answers to wade through, and some chance of a real expert taking the time to analyze the variations, you have hundreds, even thousands of answers to wade through, most of which are so similar that you can easily lose track of the differences.
4 Likes
When benaston wrote …
… I assume that was talking about changes to site policy – not coordinating attacks on specific posts – for example to agree on policies like how to handle homework questions … which might be contrary to the wish of some other not-power-users who don’t use Meta.
So …
… maybe true but I don’t think that was what benaston was talking about – was instead talking about policies (agreed on meta) like whether homework questions should be closed, whether duplicate questions should be closed, etc.
Another subtle problem is that people (authors) own answers.
For example if I see some answer as wrong (but highly-upvoted) there’s some cultural pressure against my editing the question – after all, who am I to edit someone else’s highly-upvoted question?
SE’s ethos – of separate answers from separate users – reduces inter-user argument but is IMO a root of this problem, or at least the root of what prevents its easy fix (i.e. editing the allegedly “wrong” answer).
It seems likely that people who are active (“contributors”) on the site are also active on the Meta-site; but I’ve read that’s not invariably true (e.g. on SO and meta-SO).
Also the population of users on this (Discourse) site is presumably self-selected to be highly meta-oriented. It may be, that meta-users and what meta-users do is very “cool and normal” – but perhaps we’re also biased on that topic. 
A different perspective if I may is that a lot of argument and suffering comes from:
- “I-making” – e.g. this is my site, my reputation, my view, I am right, I am experienced, etc.
- “Conceit or pride” – which involves comparison, like I am superior, I know better, etc.
That (conceit) is especially associated with “sectarianism”.
Doctrine has it that (rare) enlightened people don’t experience conceit at all. And semi-enlightened people only experience “true” conceit, e.g. thinking they’re superior when they are superior.
It might be conceit that’s a root of rudeness (“you’re inferior or wronger than I am, therefore I can and should be rude to you”). It might also be, whether it’s or intended or not, perceived as rudeness or hostility (“my question was closed by a bunch of know-it-alls who think they’re relatively super-duper, and who are members of a clique or sect”).
That might also be partly why people have been worrying about “reputation” – because it might have beneficial effects but also not-so-good (incidentally Buddhism too teaches that although conceit is a hindrance and evetually to be abandoned, there is a use for it on the path – i.e. comparing yourself with someone whose practice and attainments are superior, and determining to practice like them). I probably shouldn’t talk about reputation here though.
Anyway, talking about super-duper users and resenting their cliquishness sounds like classic sectarianism and conceit.
Incidentally I’m told that ancient Athenian democracy went out of its way to be democratic (albeit the demos i.e. citizens only of course, excluding women, slaves and foreigners) – all the important positions, e.g. “magistrate” and “president” were temporary, and chosen by chance i.e. by lottery (maybe like the modern jury system). So, I was told, there was like a 25% chance that sometime in your life you might have been president for one day. This was to prevent the existence of a permanent political class, i.e. career politicians IOW oligarchs. There were some famous orators and “demagogues” though.
I kind of wish there were better integration between the site and Meta and chat (noting that I am a meta-user). SE, wisely or otherwise, keeps them somewhat separate. Instead I’d prefer some way, I don’t know what exactly, for meta and chat to be as discoverable and as welcoming as the main site is.
And maybe a way for certain “important” meta-topics – e.g. deciding site or community policies – to be advertised as widely as possible, something a little more effective than the SE’s featured tag.
I’d like you to consider what happens if a site is long-lived, how does it handle change. Does it keep the same policies it had when it started? Adopt new ones? Who does the adopting? Can you satisfy both the old-guard conservatives and new-guard modernists? A new-guard may be seen as arriviste, by the way, another form of conceit.
Can the site software handle (accept, satisfy) sectarianism? For example would a homework tag – which is arguably a meta-tag – serve to sufficiently permit but hide that type of content (instead of banning or closing that type of question)?
Perhaps it should try to. But (from the “those who don’t learn from history” department) do people understand why SO don’t or didn’t, e.g. the “death of meta-tags” episode?
I found that being able to “ignore” a tag was useful to me once, on one SE site.
1 Like
No. This completely the wrong mindset. It is the core group that provides most of the value of a site, and has the most interest in keeping it clean. Their contributions is what keeps the questions coming in. Disenfranchising these contributors is the mistake SE has been making over the last few years.
No, that is not the solution. As a site regular, it always pissed me off when someone (usually with far less effort put into the site) said that I should just ignore bad questions instead of downvoting and closing them.
That solves nothing because it doesn’t address the problem. It’s not about me seeing a bad question. It’s about the bad question existing on the site. Bad questions hurt the site because they drag down the general quality level, add noise, and give the appearance that asking that level of question is acceptable. That only begets more bad questions.
2 Likes
On the closure as duplicates issue: I think the main problem might be that the question is closed as duplicate to another question. A better policy would probably be to write a new canonical question as soon as two non-closed questions to the same topic exist (ideally in its own category, so it’s easier to find), which then is a collaborative effort, and then close both questions as duplicate from that. Whenever a new possible duplicate arises, the canonical question is checked whether it already adequately covers that new question, and if not, whether it can be edited to do so. In that case, do the edit, and close the new question as duplicate. Otherwise the new question clearly is not a duplicate, but only closely related.
Software-wise this would basically mean to support two things (which would be configuration parameters for the individual communities):
-
Allow closing a question as duplicate of a question in a different category
-
Make it possible to disallow closing a question as duplicate of another in the same category.
Actually thinking about it, both are actually two sides of the same coin: Each category would simply have a list of categories that may contain duplicate targets. If a category doesn’t contain itself as duplicate target, then closing a question as duplicate of another question of the same category is not possible.
Note that this design allows communities to use the SE-style close-as-duplicate and the close-as-duplicate-of-designated-canonical I suggested above, so it would give maximum flexibility.
Probably the default duplicate-target list of a category, if not explicitly edited, should consist exactly of that category (that is, the SE model is the default if not configured otherwise). This is because there may not be a category for canonical posts to begin with, and even if there is, the software won’t be able to identify it. And we certainly wouldn’t want to enable closing main posts as duplicates of meta or vice versa.
1 Like
That’s an interesting idea.
There’s a bit of a problem with applying it, though – depending on the type of topic.
The problem I’m thinking of is, for example, on Buddhism.SE there are more than a hundred questions each asking variations on a theme, “what is anatta?”, and “what is rebirth?”.
Iterating on your suggestion then, you (a community) might end up writing a long canonical essay or a book on the subject – which (writing books and essays on the subject) has already been done elsewhere.
Maybe on this site such an “essay” could be reformatted or built as a set of “FAQs”.
I have sometimes been interested in that meta-topic – i.e. the ability to collaboratively author some long (e.g. with subsections) body-of-knowledge on a subject, complete with comments and with references and footnotes and so on – that is however more than (or different than – or subsequent to, post facto) simple “Q&A”, as a vision or mission statement.
Even so what you’re suggesting is an interesting alternative to the status quo, which might fit for certain types of topic (e.g. the scenario where there are several causes of a given error message each with different fixes required) – does that require any specific software support/functionality, or is it only convention i.e. how the software is used?
There has been some discussions around handling duplicate questions here. I have made a more concrete proposal, please discuss there:
Does the intentionality of a meta brigading really matter? Intentional or not, this emergent phenomenon still occurs. Usually it looks like a flurry of downvotes followed by a rapid-fire closing. Where’s the upside in this sequence? It’s a horrible user experience. The site could have just lost a future expert.
We shouldn’t conflate two things:
- Mark as a dupe, with a link to the duplicate(s)
- Closing
I am arguing that the closing of such questions buys very little, but costs a great deal.
OK thank you for highlighting this. I did not know about that.
That value is afforded by the marking a question as a dupe, not by the closing of the question. The search engine can easily prioritise the showing of higher-upvoted questions - which conveniently will tend to be canonicals.
That is one of my key points in Handling duplicate questions - #2 by manassehkatz In essence, duplicates are a problem but making them almost the same as Close because “evil” turns off new users.
4 Likes
I am not advocating for not providing tools to garden/moderate.
I am advocating for providing different tools within different user flows.
For example: on S/O, marking a dupe closes the question.
I am highlighting that marking a dupe and closing a question are two things that need not occur together, and that we might derive some benefit from separating them (for the reasons supplied in my other posts).
The Communitree example and the 90-9-1 rule were very interesting.
2 Likes
Indeed, FAQs were the form used back when questions were asked on Usenet, without any special software support. And several of those FAQ lists still exist on the web, so I’d say it’s a time-proven form.
Indeed, a FAQ is by its nature a canonical question. “FAQ” might be a good name for the corresponding category.
The software support required is outlined in the post you replied to (basically, it’s support for cross-category duplicates, for the case that canonical questions get a separate category). Beyond that, it’s pure convention.
Interesting concept. Trying to automate this is unlikely to work well. How would the system decide the topic is the same, for example? Tags are a hint, but you can ask two very different questions that still fall under that same general topics as indicated by tags. Keywords will probably be worse.
Then how is this canonical question going to get generated? It takes real humans, preferably the resident experts, to do that.
Just two questions that appear to be on the same topic is a bit thin, and probably not worth a canonical question. If you have too many canonical questions, you’re back to the original problem of not being able to find the question that already addresses a particular problem.
The best the system can probably do is to use AI to identify possible multiple related questions, then let those with high enough score (this is an area where you really want the experts writing the answer, at least the first answer) peruse the list. There then need to be ways to say, “no, these two questions aren’t the same”, so that people don’t have to keep looking at the same possible related questions over and over again. This could work well, but will probably need a lot of tweaking after getting some experience. Definitely NOT MVP.
3 Likes
Point of order: Please use the thread I linked above about handling duplicates to discuss, how duplicates should be handled, because otherwise we’ll duplicate (pun intended!) the discussion.
1 Like