There will be the need to somehow introduce new people into how our system works. This is in some parts MVP, but in most parts probably soon-after-MVP.
As a sidenote: I think we should (must) welcome all contributions by people, who are willing to be helpful and to learn (“badly written post”). If someone is not good in English, that’s fine. Let’s edit their post to improve spelling and grammar. This will help our site and their English skills, hopefully. That doesn’t mean, though, that we shouldn’t moderate/criticise people, who produce bad content and aren’t willing to learn (“bad post”).
Here are my ideas:
I am talking about help, specificially directed at new people. This is not about “help documents”/etc.
All help should be highly context-sensitive.
It is known, that users never rarely read long texts/explanations/help documents. However, when we provide context-sensitive, short messages, there is a chance of the users really reading through them.
The help should be – wizard-like – separated into smaller steps
The feed back should be given one-by-one and with verification, whether the step has really been completed.
Most help should be directly actionable.
For example: When someone’s post is closed, we shouldn’t tell them, that they need to read the help center, but we should give them hints, what they might change for a better reception.
Going from this, I’d imagine the following types of help tools:
General introduction (AKA “Tour” on SE)
Question assistant
Answer assistant
Edit guide
Closed question → get reopened assistant
All of these should be optional, but on by-default. This allows users experienced with the system to skip them (in general or in parts), whilst directing new users to it.
I like the First Posts review queue, where the first post that anybody ever does gets kicked to a special review queue by the more experienced users for editing etc.
Beyond that a lot of it is culture and a need to be welcoming.
If its a choice between
A. Making the small edits to salvage a question like capitalization, proper tags,making it on on topic or explaining how to upload images…
B. Yelling at the user and closing the question.
I would have to hope that people would choose option A.We also need to not yell at the experienced users who choose to give a hand up to new users like happened to me and I am still not happy about 2 plus years later.
There are plenty of times where the difference between a good question and a bad question is 30 seconds of editing, we should encourage helping the new users by editing their posts.
But yeah. That’s what I mean with the difference between bad posts and badly written posts.
The former should be moderated (closing, deleting; while still being nice towards the user). The latter should be improved (editing), because the user is probably not that good in English and we won’t help anyone by removing that post.
That’s a false dichotomy. In many cases I prefer to leave a (non-yelling) comment on how to improve the question (and I also appreciate it if someone does the same on my questions). Without casting a down vote or even a close vote. Obviously that doesn’t apply to obvious typos, or generally to cases where the issue clearly is one of mastering the language (although there I may comment simply for the reason that I’m not sure whether I understood correctly — I’m no fan of speculative fixes; someone else might react to the “fixed” post before the OP has a chance to correct it).
An example of such a comment could be:
Hint: The typesetting of sin and cos is much nicer if you prefix them with a backslash (like \sin).
I can’t imagine anyone reasonably getting annoyed at that type of comment. Of course, if you just write
Please fix your MathJax.
that isn’t much helpful (although it’s still not yelling).
I really like the idea of small wizards/dialogs showing up only when needed.
Currently, SE heavily relies on veteran users to guide the new ones. If someone asks an off-topic question, write a comment to help; if it’s unclear, help the OP to clarify it; if the OP says “thanks, it worked”, remind them that the answer should be accepted, and so on. It might kinda work for smaller sites, but for bigger ones like SO it’s unrealistic to believe that leaving this task only to the community will be enough (even in SOpt, which is much smaller compared to SOen, with an average of 100 questions per day, it’s too much workload for the community to handle). There are more questions needing guidance than people to guide the respective OP’s.
But SE refuses to improve the site, to make the interface itself provide guidance to those users. They prefer to leave the burden to the community - which is not enough - instead of making a smart UI that could guide newcomers. Things like these ideas are never implemented. IMO that’s a mistake that we shouldn’t repeat.
I’d say that every “core action” in the site (voting, asking/answering, editing, what to do when someone answers my question or edits my post, anything that requires some learning about how the site works and it’s not obvious to new users) should have their own assistant.
Lots of rules in SE sites are spread through Help Center articles and Meta posts (the former is more “obvious” and “easy to find”, there’s even a link to it in the top of all pages; the latter isn’t, specially for new users - I wonder what’s the percentage of users that knows about meta), so people usually take some time to learn how the site really works, what they should and shouldn’t do, and so on. I believe that the context-sensitive assistants will be very useful in order to speed up the learning curve for new users, and to minimize the burden on the others, who will have more time to do Q/A things instead of “please fix your question” stuff.
The location of “the Wash” is not obvious to a lot of people so there are options
Nicely ask the OP where it is.
Do a Google Search
Wait for someone who knows where it is to answer.
Close the question and accuse the OP of being arrogant for not specifying the location in the original question.
Claim that the experienced users who edited the question with the location are making the site worse.
All of those things happened, what I am saying is that options 1-3 are fine, but we need to avoid 4 and 5.
Being nice does not mean that content doesn’t get moderated, it means that we are not deliberately rude when moderating or being rude to users in hopes of driving them away.
I have helped close and delete hundreds of posts and over 10% of all the downvotes on Outdoors.SE are mine, I don’t have a problem with moderating content. What I have a problem with is users being rude, which in turn drives away away the users that want to help, which means that there are fewer users to do the moderating.
Can you clarify whether “3a. Politely close the question until someone knowledgeable edits in the correct location” would be acceptable to you in this context? It seems like voting to close, and claiming the OP was arrogant and thus deserved to have their question closed, are two distinct things that don’t need to be connected.
Closing can absolutely be done politely, however it’s a heavy handed solution compared to editing. Closing takes time and multiple people while editing only takes one.
Here are two questions that were unclear due to lack of pictures.
In both cases, they first got incorrect answers due to the lack of clarity. The solution though is to add pictures to the question, and I don’t know that closing the question would help with that.
Answering a question before it totally clear is a high risk/high reward type of situation, the first answer often gets more points, but every so often you misunderstand the question and so have to rewrite the answer.
I would rather people not answer questions that are unclear to them, than try to protect them from answering an unclear question by closing it.
However, editing a mess to make it acceptable also has a significant long term cost. It teaches that dumping a mess on us works, in that the desired result is achieved. The OP will likely be back doing the same again, since it worked with no apparent downside to them. Perhaps even worse, bystanders get the same message.
The only thing we have that they want is an answer to the question. Withholding that is our only leverage to enforce a minimum quality level. That’s what closing a question is all about. It in essence says “fix this mess or else”. Without the “or else” part, experience has shown that we largely just get ignored.
So editing a bad question to make it good may seem like a good thing to do in isolation, but it often is not when looking at the site as a whole over a longer time frame.
Yes, of course, but in reality it doesn’t work that way. All it takes is one person who either thinks they understands the question or guesses right, and the quality controls have been subverted.
Answering a bad question is a case where it is of benefit to the answerer, but a detriment to the site as a whole. That’s the point of closing. Given that someone will always jump in and answer a bad question anyway, it is our mechanism to lock them out.
As demonstrated on SE, the problem is that answerers can easily answer before questions get closed, which undermines our quality control all the same.
The ideal is to prevent answers from the start if the question should be closed, and the only way to guarantee this is to require that every question passes review before being made public.
This may not scale as well as only reviewing questions if someone flags it, but overall quality of visible questions should be much better.
Of course there can be some sort of prioritisation in the review queue to avoid wasting time reviewing garbage.
This assumes we don’t want to more new bad questions. If the goal is simply to get rid of them after the fact, closing can help. But that doesn’t really discourage more such questions, since people get answers one way or the other. And if you keep getting more and more such questions, eventually they’d overwhelm reviewers and either review standards will drop or many things will slip through the cracks, or both.
It benefits them because they get rep from the upvotes, changing the system to not reward this sort of behaviour (and instead reward based on a better voting system perhaps?) should go a long way to fix this
In cases where the problem is bad markup or lack of markup doing part of the post and leaving a comment with advice on how to finish has worked for me in the past.
Friendly and helpful but also making it clear that we expect them to work up to the standard.
That strategy fails for cases where the issue in incoherence or utter lack of clarity simply because the isn’t a clear way to fix part of the post.
That’s a bit drastic, and requires a lot of drudge work by volunteers who would probably rather do other things.
I proposed something a while ago that maybe should be mentioned again. Questions require a positive score to be answerable. For new users, questions start out a 0, meaning they are not immediately answerable, but a single upvote by anyone makes them so. For subsequent questions, the initial question score is the recent average (3 months or last 10 questions or something) of your questions score. This means that as long as you don’t write bad questions, only the first will be on hold until it gets the first upvote, and all your other questions are answerable immediately. It also automatically puts questions on hold from users that have proven to be problems.
Basically, if you have a good track record, your questions are considered good until judged bad. If you have a bad track record, your questions are considered bad until judged good.
Olin, I’m not sure how to square this with what you’ve said elsewhere about not really caring about the specific asker of a question. If something comes in that’s kind of a mess but I think there’s a good core question underneath, and I clean it up, and you write a good answer — doesn’t everyone win?
In my experience, most “bad” questions of the type where editing helps are cases where English is a second language. Or, they’re cases where the asker doesn’t quite know encourage about the topic they have a problem with to phrase the question well. I don’t think closing those really helps get better results.
On the other hand, when there’s a case of “What’s this effect” and we can only guess — I’d really like the original poster to try to describe, or else we end up with a series of guesses.
Maybe there is a mixed option, where a single flag from a trusted user moves a question from the front page to a review queue? That way most of the time questions could just with no delay, and the messier cases also handled.
At SE, I’ve often wished there were a sort of meta-ask site, let’s call it ask.se.com. The whole point of Ask would be for new askers that probably don’t know exactly which stack the question should belong to. Rather than actually answering the questions at Ask, the actions there are to quickly find the appropriate stack the question should land on.
One idea would be that “tags” at Ask strictly just map to the different stacks (i.e., “photography”, “video”, “graphic-design”, “code-review”, “unix-linux”, etc.). The asker probably has some idea of where they think the question should go, but the regulars at Ask are the high-rep users and diamond mods of individual stacks. So if you, for instance, were to opine that a particular Ask question isn’t really suited to Photo-SE, your high rep at PSE would generally outweigh another Ask who isn’t active on PSE (but would otherwise have a fairly decent sense of most questions’ topicality).
Perhaps all first-time askers have to ask new questions at Ask-SE (or rather, I’m advocating Ask-${MVPname}). That way, the collective wisdom at Ask helps weed out the obvious junk from all topic-specific sites (obvious spam, questions that are clearly poorly worded or asked and aren’t likely to be salvageable). And also, the rep system at Ask can be tuned to how successful the regulars are at directing questions to the most appropriate sites, rather than being judged on individual sites’ question/answer rep.
It’s a different sort of incentive system, essentially a mod or meta-mod rep system, separate from topic-specific rep. Ask rep specifically rewards / tracks a person’s sweep-up or shepherding duty, which I feel SE only really recognized with things like review queue badges.
Not necessarily. If this shows the OP and everyone watching that they can dump crap on us and we’ll just fix it for them, then we’ll get a lot more crap that others expect us to fix.
I think most are just laziness and an “Eh, who cares?” attitude. Some of that comes from previously getting away with posting bad questions. No matter how little you know about English, capitalizing the word “I” is such a universal and simple rule that there is no excuse for getting it wrong consistently. However, we see a lot of that, and other blatant crap for which there is no excuse.
As for those not that good with English, maybe there should be a separate ESL review queue that the asker puts the question on voluntarily. Anyone that feels they have the necessary English skills can edit the question, and release it from the queue. It can’t be voted on until this is done, so the OP won’t get penalized for the initial bad question.
This would also makes it clear that all questions posted directly are expected to be well enough written. At that point we can be ruthless about downvoting and/or closing.
… in fact, I can’t think of any other languages off the top of my head that do capitalize “I”.
Yes, some people are lazy and don’t put effort in. But in my experience moderating on Stack Exchange, the vast majority of badly-written questions (or answers) are badly written not through malice but through inability. Why should we penalize people for making the effort to write in a second language?
