Why are we getting so many off topic 'debug this for me' questions? what can we do about it?

Question

^{Yes, this is a duplicate of Using Programmers as "learning and training area" for Stack Overflow kind questions intended to re-ask the question more clearly and without the tangents of trying to explain it in the comments that have distracted from the core question.}

For some reason, Programmers.SE is seeing an increase in the rate of new users asking pure code and debugging questions. These questions are, and have always been, off topic on Programmers.SE.

The majority of the time, these questions are down voted quite severely, closed, and often deleted very promptly. This is quite simply, again, that these questions are off topic here and asking such a question suggests that the user didn't even attempt to read the tour or any of the help center before posting.

For 10k users, poke at the question close stats for the last 30 days.

To those who don't have access to this, the second and third highest custom close reasons are the all too familiar:

This question appears to be a fix-my-code-dump request, an outright off-topic, totally ignoring what Programmers is about: https://meta.stackexchange.com/a/129632/165773

or a variation on it. The top one is a custom 'unclear' which also often acts as a proxy for the "I have no idea why you are asking this question here." There is also a smattering of other close reasons:

• We don't troubleshoot code here, sorry. No, we don't know where else you can ask.
• We don't debug code here, sorry. No, we don't know where else you can ask. See also How to Debug Small Programs.
• Voting to close as this is a debugging question, but I'm not sure it belongs on StackOverflow since this is a classic example of a bug that can be found simply through proper use of a debugger.
• I'm voting to close this question as off-topic because we don't do fix-my-code requests here.
• I'm voting to close this question as off-topic because Programmers is not a fix-my-code site.
• I'm voting to close this question as off-topic because it's a "fix my code" request.
• I'm voting to close this question as off-topic because it is asking us to debug the code.

There is also a fair amount of "I'm voting to close this question as off-topic because it is an implementation question, but not of sufficient quality to migrate to [so]" and variations on that reason.

The first thing is realizing that no, we aren't going to migrate crap to Stack Overflow. Migrating an awful debugging question to Stack Overflow isn't going to do anyone any favors.

However, this fairly consistent stream of completely off topic debugging questions is draining the community moderation on the site and discouraging to many people who are either active or casual users of the site. It is very discouraging to see "1 new active question" and then look to see it is another poor quality "debug this for me" question. It is also discouraging to for the casual user to look at new questions and see half of them down voted and/or closed. This can be seen in some of the recent meta questions (and many who are active and read meta know that there are quite a few more of this type of question that are older).

---

Yes, there are a lot of questions here.

So, what can we do to help reduce the rate of off topic 'fix this for me' questions?

Some have suggested that changes to Stack Overflow's question visibility for low quality questions have caused people who are asking these low quality questions to try posting on other sites (ours in particular). How accurate is this?

Are there things that Stack Exchange devs might be able to implement that would help new users recognize the scope of the site more rapidly (before they hit a question ban? or get discouraged with posting on Stack* (its just a bunch of jerks who close everything))?

How can the existing community more promptly handle these questions? Do we need to try to hire Oded 7/24 to close these questions promptly? Or is there a valid mandate for the diamond mods to close vote (and delete) the clear cut debugging things that show up in new questions more promptly than the community does?

And for any Stack Exchange community managers - do you have any suggestions for us? How important is it that we provide a good user experience for these users? Could we get another close vote reason for 'fix my code is off topic here' so that it is something that is more official and less curt than the plethora of custom close vote reasons that we currently employ? And while we have anecdotal suggestions of this, is there any change to the "questions asked and closed within 48 hours" rate recently compared to historical trends?

Answer 1

I think the problem is ill-defined. A single misconception underlies most of the issues raised.

The overarching question is not Why can they not understand who we are? but How?

After having been introduced to a long history that stretches at least 3 years back, an obstacle has become clear to me: "...there is no evidence right now to support the claim that the site's name is leading to its problems." So maybe 2012 is not "right now" and Programmers could potentially have a case - provided that there is sufficient evidence. What the previous argument completely neglects is the concept of what is sometimes referred to as internal marketing or (brand) identity. Direction. How can a community be built without a clear unambiguous goal and image?

The Name

The popular understanding of "programmers" is that programmers code, but the current stack is to my understanding more about design and development - what programmers actually do in addition to coding, or should I say software developers?

The Logo

Do you really need three screens for finalizing that UML diagram or a requirements sketch? Besides, coffee is culturally engrained with "crunch time" and if you work in an office, perhaps even soft skills. It is not surprising that out-of-scope questions are being asked.

Conclusion

Neither the name nor the current logo reflects anything more than the stereotype programmer as a coding clerk. For the generations that have grown up in the digital age of mobile apps, usability is in the brand.

IMHO, a new logo and name that reflects the scope of the stack could catch potential users at the earliest stage of involvement. Perhaps it could even attract new followers, who have have been asking good questions over at SO, which are more relevant to the current stack.

---

Assign a task-force to gather evidence for how the name affects user-interaction. Statistics could be derived from time spent on site for samples of poor question (e.g. a 5 downvotes criterion) compared to random samples of good questions and the number of click counts for the FAQ. A sample size of 6 on an arbitrary day in 2012 [eternal September pun goes here] is inadequate and lacks verification. Furthermore, demonstrate that the site has been stalling compared to other similar sites.

Answer 2

I am only barely a programmer, by the definitions summarized on this site. I possess a lot of experience in linear programming and process management but object-oriented programming is not yet within my realm of experience, and that seems to be what all the kids are doing these days.

With that said, I came here because I KNOW I don't know the answers to some of my questions. I am not in school (quite old and can't afford it) so I am self-educating and asking questions on a board somewhere when I am having difficulty figuring out where a piece of code is broken after working on it for several hours has been quite helpful.

I am a reasonably intelligent guy who is new to the site (< 30 days), so I like to think I am fairly representative of the folks who are clueless as to how to use this site. I know you don't want to be doing other people's homework for them, and perfectly respect that, but without checking for student IDs there is no reasonable way to tell the difference between a struggling programmer and a student programmer.

There is no reasonable way to prevent folks from asking debugging questions. What you CAN do is channel them into a place specifically to ask debugging questions or force them to identify their debug questions. If there is a place to ask debugging questions, it is unclear as to where that is. I think it might be here: https://codereview.stackexchange.com/ but am unsure. It sure took me a lot of searching to find it after I started typing all this out, if it is the right place at all.

So, the result should be to make a Debug community (if the above is still the wrong place) and send folks requesting Debug help there. Or, alternatively, tell those folks to preface their question titles with a DEBUG: to let users know before they ever open the question that it is a debug-related question.

Answer 3

Why are we getting so many off topic 'debug this for me' questions?

The date the question was posted (towards the end of September) and a serious uptick in 'homework' questions does correlate well with the start of term in many education establishments around the world.

Many of them are kids who have all grown up with a 'sharing economy' with an expectation that because it's the internet it's all free and someone will do their work for them (for free).

what can we do about it?

That's the $1 million question - ultimately we either need to put up with it or make the process of asking other people to do your work for you more hassle than doing the work. We would also need to do those things without adding excessive amounts of work onto the mods.

I've a few ideas, but they're musings more than fully-fledged solutions and they really depend on how the mods want to run the community...

• put new questions on-hold automatically if they meet certain criteria - rep (both here and at SO), having a question rejected at SO might be a couple of many signals that could be used. If the moderators want to police questions in that way (it's extra work and they may not want to.)

• deleting answers that are too obvious so that the people asking the 'do it for me' questions are less likely to get a copy/paste-able answer. That might upset some users who provide answers.

• Use a [homework] tag that triggers a more visible reminder to those answering (who may not fully understand what programmers.se is about) that "teach a man to fish" is more useful than solving a problem directly?

Just some ideas, I have the flame retardant suit on so feel free to let loose in the comments... :)

Answer 4

Why are we getting so many off topic 'debug this for me' questions?

I believe that most of the questioners asking debugging questions here are indeed expecting to get an answer here, i.e. they think that these questions are on-topic while they are not. It's a big misunderstanding.

What can we do about it?

Make it as clear as possible that these questions won't get an answer here. Indeed from the start page it's rather difficult to understand the topic of this stackexchange.

I have two ideas to discourage people more actively.

Idea 1: An explicit warning notice when asking a question that questions asking debugging help are offtopic and will be closed without an answer. This warning notice should only appear to every new user asking the first X (=1?) questions.

Idea 2: Since a name change is outside of what we can do and also seems unlikely, change the remaining layout (subtitle, logo, other notices) to make clear what this community is focused on and what it is not focused on.

Just a primitive visualization with a terrible logo:

Answer 5

Here's the approach I am currently using when I stumble upon this kind of questions. (Those interested in what else I tried before coming at it, can find details below.)

1. Add comment referring to MSE guidance Which computer science / programming Stack Exchange do I post in?
2. Vote down.
3. Vote close as unclear.
4. Vote delete (when question becomes eligible for that).

Mentioned MSE guidance is sufficient to help asker figure another, more appropriate site to focus their efforts on. Worth noting that in the past, I've been referring it rather judiciously for the fear that it may lead to reposting very low quality debugging questions at SO. Nowadays, I am not concerned about this, because SO (finally) found an efficient way to handle garbage questions by burying them in Triage queue.

Vote down is because it's off-topic. Additionally, it makes it easier for the system to trigger blocking, rate limiting and anti-recidivism protection in case if asker intends to persist in asking inappropriate questions like that, for example in order to use Programmers as a proxy for Stack Overflow.

Close reason picked (unclear) is because I abstain of wasting effort on checking whether the question is good enough to migrate to SO or not, or maybe it needs some improvements to become good enough to migrate. Another facet of this reason is, it's unclear to me why asker picked Programmers instead of more appropriate and much better known site (Stack Overflow).

Vote delete serves same purpose as vote down, to help system act in case if there's intended and persistent misuse. FWIW when voting this way, I expect that for occasional mistakes it doesn't harm, ie that question blocks work as explained by SE team here: "first question doesn't count".

---

As promised, here is a brief overview of what else I tried before coming to above approach.

1. "Classical" review. You know, it's when you start with checking whether question is good enough for Stack Overflow and if it is, vote migrate. If not, you additionally study the question to either figure how to edit it to improve, or pick close reason that would help the asker find how to improve, maybe also comment with more explanations, stuff like that.

This way used to work pretty well until amount of debugging questions increased so that studying them started taking too much time. When I noticed this, I dropped it. Quoting self, "I come here to learn and contribute to Programmers topics" - I invest extra effort into studying off-topic questions only if it doesn't take much time.

2. Trying to use system capabilities. You know, system can tell many cases when debugging question is posted here to abuse it. It can tell whether asker was recently blocked at SO and likely tries to circumvent that, or if the question has been cross-posted here, possibly in hope to gather more feedback if it needs to improve. It can tell if the question is of such a troublesome quality that Stack Overflow would hide it (or maybe already did that) in Triage queue.

I tried to figure how above can be used to simplify handling these questions and posted various feature requests. So far, all of them have been ignored... and likely will stay that way, given that even very easy to implement and reasonable request to block same titles as are blocked at SO (with words 'problem' and 'help') stays ignored.

3. Flagging. For readers unfamiliar with how it works, you explain in the flag message what specifically is wrong with the flagged post and why moderator involvement is required (as opposed to regular means like community voting, comments, edits, canned flags etc).

Similarly to "classical" reviewing, this used to work like a charm in the past, and just like it, this way broke when amount of debugging questions increased so that flagging them started taking too much time. One can't just type in flag message "hey, this is debugging" and expect moderator to take it from there. To make a compelling flag message, you have to study the post quite thoroughly, that's why I dropped it when site started getting too many debugging questions.

See also:

• Perform automatic checks for cross-posting and question block at SO when post enters close queue, not only at attempt to migrate

• Could we please detect troublesome questions using Stack Overflow data and offload them to Triage review for an hour or two?

• If "Problem" and "Help" are to be blocked on Stack Overflow question titles, block them here too

• Should one advise on off-topic questions?

  Don't ever answer a question, not via chat, comments, or any other way. It is even dangerous to say: go here, Google for that, because next time, they will do the exact same and eventually get what they want.