Both could’ve been asked by someone who had a real problem, and the answers on either (should) help that specific asker with their actual problem.
The former should help anyone with the same error figure out what’s wrong with their code, except in some rare cases which is specific to some library or tool.
But, for the latter, the accepted answer is little more than “line X is your problem, it should be changed to Y”. The other answer expands a bit on why this is a problem, but it does so in a way which is entirely specific to the asker’s problem. Neither answer is likely to help anyone else. And the question itself isn’t likely to convince someone they’ll get an answer to their similar question there.
As humans, we’d naturally tend more towards the latter for both questions and answers, i.e. focusing on the specific problem instead of trying to create something a bit general. This tendency is something we’ll need to actively fight against if we want Q&A’s more like the former than than the latter.
In conclusion, the selling point should be “get your question answered here … while building a repository of knowledge”. Real questions about real problems, but under the filter of building a repository.
I think a big part of ensuring long-term quality on the site alongside friendliness is really finding a good system that makes merging questions easy and worth it, makes updating old answers easy, and doesn’t leave the users who have their questions merged feeling bad for asking – nor leave them without a full answer to their specific problem.
A clear UI will definitely be necessary to communicate all of that, but a large part is also just the initial culture. Are we upset to have to go and find the right canonical Q&A to give to the thousandth user to ask about null pointers, or are we glad to help point them the right way?
In the early years (I got to EE.SE about 2011), I could be quite blunt about what was wrong with a question. I could say things like “Not gonna answer text-speak”, “The English word “I” is always upper case”, “English sentences end with periods”, “No schematic, no answer”, “Volts is not a unit of power”, etc, then vote to close. Those that dumped slop on us were sent home, and it generally worked.
Comments were forced to be ever more polite. As a result, they lost their effect. OPs no longer either fixed their questions or went away. The line insisting on minimum quality before getting the desired response got ever fuzzier and moved further back. OPs could now write very badly, and it would get fixed for them. Of course that meant they just kept doing it, since it worked and there was no cost to them.
Keep in mind that all along new people came to the site that were able to ask good questions. Some aren’t that good with English, but the result of that is very different and obvious from just plain sloppiness.
There must be a minimum standard of quality, and it must be adhered to. But the really important point is that this quality must be achieved by the OP. It’s OK for others to fix honest errors, but sloppiness is not something anyone should be able to get away with. We need to be able to say firmly “Come back when you’re ready to treat this seriously and to respect the time of those you are seeking a favor from.”.
I wouldn’t go as far as closing a question if there are some minor problems with the spelling and/or grammar, but if it’s atrocious I can empathise.
A lot of those comments might be seen as rude. We don’t need to be rude to not welcome questions we don’t want, and I’d argue the site would be better if we aren’t (both by having a more positive atmosphere for the rest of us and also to convince others it’s a nice place to be, as long as they put in the effort to follow the rules).
I can understand where this sentiment is coming from, and allowing questions in dire need of editing may indeed make such questions more common.
Although I personally think good questions are rare and valuable and I’m not that bothered by needing to edit questions (even to the extent of basically rewriting all of it), so I’d be fine with welcoming a good question that’s asked badly (as long as we prioritise creating a post with long-term value very far above whatever the author may or may not want).
I think part of the problem (and this is a societal problem, not specific to SO, not specific to Q&A sites, not specific to the internet) is that if “not so great” is tolerated then eventually it becomes the new “normal”. I could elaborate, but I don’t have time (and most of you won’t want to read my rant profound wisdom anyway).
In the end though, this is a Moderator/Community issue and not a software design issue (other than providing good tools to the Moderators and the Community). The levels will vary quite a bit for different communities (e.g., grammar in Questions is less of a problem for ELL than for Writing) and we can’t “program the solution”.
This is wildly site-specific. I think the endless flood of duplicates is something that mostly appears on the technical and scientific sites. While other sites like Code Review can’t even have duplicates by definition/site design.
This is best solved by implementing a working, user-maintained FAQ system for communities that need it. Finding duplicates was always a major pain on SO, until one implemented a list of canonical dupes to pick from.
I would argue this problem has a lot to do with the rules and guidelines set up by the platform (or the individual sites).
Of course you can’t force individual users to do what you want (you can force them to not do what you don’t want…), but hopefully most users would roughly agree with the rules in general and this would result in them generally getting enforced (especially if the moderators lead the way, whom can generally be told what to do).
I agree that allowing questions a bit beyond what you ideally want can lead to continually extending what’s allowed, but I would argue editing / rewriting a less than ideal post so it’s appropriate wouldn’t suffer from this problem to the same extent as simply accepting such posts (hopefully it would avoid it entirely, but that might be overly optimistic).
On a per-tag basis, I hope.
Trying to have a single FAQ for all of SO would be a nightmare.
I don’t think so. First, let me get a rant off my chest:
<rant>
Please let’s stop calling introductory documentation a FAQ. First, most of the information is not asked about. It’s something we want you to know. Generally you don’t even know to ask or don’t care. Second, it’s not a set of questions. It can be a set of answers, but usually not even that. Its rules, background info, site operating procedures, etc. Call it that.
</rant>
The problem with such documentation is that it won’t get read, particularly by the people that need to know it most. The kind of person that wants their problem solved now, doesn’t care about site norms and policies, and doesn’t feel any obligation to take some care when asking a bunch of other people for a favor is just going to blurt out their question. No amount of help documentation is going to fix that.
What we need is to force new users to read some basic introductory material. It should be kept to 1 page, and shouldn’t be burdensome to read. However, we have to make sure it is actually read. I don’t know if a button at the bottom “I have read and understand the site guidelines” is good enough. Probably not. Another possibility is to actually quiz them once they claim to have read the rules. You need to get 4/5 right before proceeding.
I think (but could be wrong) that some of the discussion here about FAQ isn’t referring to “every website has a FAQ =Help page” but rather the Q&A-specific concept of “Questions asked frequently enough that they have Canonical Answers” and where a good set of those would eliminate a ton of duplicate and /or low quality questions. For that we need (and have had some discussion about) a blog or similar that would have the same editing etc as regular questions but would be a separate list because doing this as a regular question gets “lost”.
I thought Lundin meant an actual FAQ - a list of links to frequently duplicated questions on the site, to make it easier to find one when encountering a duplicate of it. But also well-integrated and highly visible - having a list hidden away in a text box somewhere (like what would happen if you try to put this in tag wiki’s on SO) doesn’t help that much.
On any given major tag, a good chunk of the questions are duplicates of one of like a dozen or so posts, so having easy access to those would be helpful.
As far as I’m concerned, this is not primarily intended for new users (although it may help them too), but rather for people looking for duplicates to use for closing something.
Let’s be clear: requiring users to be polite is not a problem. You can be pointed and firm while still being polite, but if you feel unable to participate because you’re not allowed to be impolite about bad questions… that’s not okay.
When SO was young a lot of the most useful questions to me were of the form “What is the best library to use for X?” They were always closed as not constructive, which was annoying, but not before accumulating some observations from people with experience using those libraries. Now searches for questions like that never return pages from SO, and it’s much harder to find that kind of information from people who don’t have something to sell you.
I think best is sometimes misconstrued as opinion based == close as opinion based. Yes, those questions are, except in limited contexts, generally at least partially opinion based. But opinions of experts can be very useful. Plus often that is what OP actually needs - a way to find out “what other people recommend”. Not in the “tabs vs. spaces” or “vi vs. emacs” way - those will never be constructive. But knowing that “xyz tool” has “a, b, c features that can make life easier” can be a great answer to the right question.
Or that library pqr looks great on paper, but in the time it takes working around its bugs and odd design choices to get it to work for you you could have written a better library yourself.
And you may also have sites that combine both. Sites that have questions like ‘what is the difference between theory X and theory Y?’ and questions like ‘I have this A with some B but now I want to add C and D as well, how can I be sure that E remains while I do F and keep G intact?’.
Every site is gonna have different levels of complexity (and simplicity, quality, spelling errors, usefulness, etc.). It is the point of a big community to be able to deal with all sorts of questions. In the end we get some sort of wolframalpha-like terminal in which you can enter a question, and we want to diminish (as much as possible) responses like “Codidact does not understand your question, try to…”.
This is not the countryside but the big city. If we want to be a big community then we need to give each other sufficient space to do our own thing. If we want codidact to become big then we need to organise it along that path. Allow variety and do not implement too many ideas/technology that will fix/restrict it (or better put technology in place that helps diversity to flourish instead of complicate things).
FAQ != site usage guidelines. I’m speaking of a topic-specific FAQ to close duplicates on that specific topic. Example of what I mean from SO, C programming tag:
This one is entirely user-created and maintained. The site itself provides no mechanism for it so we had to show everything into the tag wiki, which is hard to find. New users certainly won’t find it - this is mostly for user moderators who need a list of dupes to close down duplicates with.
Yes, in my experience, allowing the existing user base appear rude in any circumstance can make the impression that it’s okay to be rude in other cases, which then spreads and is hard to eradicate. I would prefer to stay calm, objective and constructive 100% of the time, and teach others the same by my example.