How can we help users who repeatedly ask similar questions or update constantly?

Question

Before I ask my questions, I need to describe two types of low-quality users.

---

Type P:
A type P user will post a block of code, possibly with a stack trace, and text that boils down to "this doesn't work right, tell me what's wrong." After an early answerer says something like "well, you declare foo but you don't initialize it," the user will stop paying attention and run back to his IDE, perhaps leaving a comment saying "oh I see it now thanks."

Then, five minutes later, the same user will post a new question with the same code (except that foo is properly initialized). Someone else will answer with "oh, you can't put final inside the parentheses, it has to be at the start of the line." This cycle repeats until there are half a dozen questions, each identical to the one before it except for one change.

Type 84: ← AKA chameleon question askers (thanks Grace Note)
A type 84 user will initially post the same sort of question as a type P user. After getting an answer, though, he'll update his initial post over and over until the early answers no longer make sense.

---

Type Ps often turn into type 84s, and vice versa, either on their own or in response to well-meaning commenters who say that their behavior isn't constructive. I'm writing this post in large part because I've been one of those well-meaning commenters and I wish I had a better response.

Of course, the real problem in these cases is that the users are posting code with multiple errors that they don't fully understand. The way we handle things now — which is to say, not very well — gets them frustrated with mean Stack Exchange users, gets us frustrated with clueless newbies and wastes everyone's time in the process.

I realize that I'm not describing a new phenomenon by any stretch of the imagination, but I think we may need a different approach here than we give to low-quality users in general. The low-quality asker algorithm won't pick up on 84s as much, I think, and Ps can tie up a lot of resources before they get noticed. The good thing about Ps and 84s is that they're engaged, so we at least have an opening.

What is the proper response to the situations these users create? How can we best help them with their immediate problems and grow them into productive community members? I realize that a certain subset of these users will never become productive community members. How can we minimize the pain on both sides when dealing with them? It does us no good for people to go around saying SE sucks, after all.

_{Some of you are probably wondering why I chose P and 84. Well, because it's boring to always use A and B, or 1 and 2.}

Answer 1

Type P

The worst of these offenders will be caught by the low-quality-question-you-can't-post-here-so-go-away filter. The new 70/month rule will help, too.

Jeff has encouraged us to use the power of the downvote to help the filter figure out who the bad askers are. So downvote away!

Type 84/Chameleons

Flag these for moderator attention. These questions need to be reverted and locked. This is a total perversion of how a Q&A site works, and must be stopped outright. Repeat offenders should be dealt with by the mods as they see fit (Ideally this involves Ninjas, but suspensions work, too).

Answer 2

Type P users usually post stuff that is very localized (forgot a semicolon or something). If you see no improvements from one question to another, just start downvoting them and voting to close the questions. The questions offer no value and aren't a real loss to the community. I've seen users waste more than half an hour just answering variations on the same type of question, basically doing debugging for users that are too lazy to just use their brain...

Answer 3

Ideally, the Type 84 Chameleons should be encouraged to revise their question, leaving the current version of the question/code on top, and leaving previous editions of the question below (reverse chronological order). The horizontal rule and 3rd level heading (prefixed ###) can be useful for this. They needn't leave everything from the earlier editions, but enough to make it clear that there was a previous edition, and the key points that answers (or comments) addressed.

Sadly, these people don't have the 'maturity' to do that. They're in a rush to complete their homework and don't realize that what they're doing is somewhat antisocial. So, realistically, what 'should' happen (under my hypothesis) does not. If it is not too frenetic a pace, someone can go back and retrofit previous material into the later version(s) of the question, but it is hard work for minimal reward. I've done it maybe a couple of times...but not much more.

I also have to agree with another point made by others; Type 84 Chameleon questions often don't have much lasting value. So many of them do not warrant the effort it would take to make the edit trail comprehensible.

Answer 4

Type P is legitimate. We can see it as one question for each part of the whole task that the OP doesn't know. Not everybody has the same level of abstraction. Good answers (completes) would help to reduce the cycle of questions. Whether those question are by themself a good entry point or not (it might too much "localized") will be answered by the community's votes. Type 84 should be discourage (One should not keep answering and editions to the question should be rollback, encouraging the OP to ask another question through comments).

Answer 5

I'm tempted to say flag for deletion, and if there was anything unique in the question, post a new one with just the relevant info.

I realise that this sounds harsh, but most 'Fix my code' questions are of little to no use to anyone other than the OP, and probably shouldn't survive longer than the OP needs them.

Of course, this does increase mod load, so maybe mods should have a power to restrict the number of questions a user can ask (with suitable expiry on the limit), with a note to say 'update a previous question'

Answer 6

I just wondered why P & 84. But nice to see some thing unexpected there. :)

For Type P:

I was unable to agree with the down rating the questions as others suggested here. What if user L asks a question and user M modifies it asks the following one. Unless we are very keen on noticing the users and tracking their behavior it is tough to identify such users and down grade only those questions. In general if I see a question I would rather try to solve the problem than concentrating on who posted it and why. I feel this is the best use of my time.

Also, some users are trying to fix their program with out thinking or understanding much, most likely who are not interested to learn or were forced by to fix some issue.

In this case my approach could be posting the link back to the original question, if I can find (by searching) it in minimal time, so that the people answering the question quickly realizes the pattern and probably use the method of downgrading the question or sending some helpful info to the user, who is posting these questions, privately to solve his/her problem and pointing them to some useful study material.

For type 84:

I agree with the suggestions made by others. How ever we can reduce the Mods work by adding the explanation why the user can not keep editing question with some examples into a FAQ and refer such users to look at the FAQ in a hope they read and understand the issue.

Answer 7

I imagine the desired behavior for Type P users is one of the following:

1. Editing their original question adding a line at the end similar to the following:

   So-and-so's answer was helpful and fixed the immediate problem, but there seem to still be issues with the code. For example I'm now getting this error: blah blah. Does anyone know why I'm still having issues?

2. Leaving a comment on the helpful answers with a similar sentiment to the above

Can we encourage them to do that by commenting on any new questions they open up? If there's something on Meta or the Stack Overflow blog that encourages them to do this, can we link them to it?