Stack Overflow community standpoint on find-my-error questions

Question

There are multiple questions asked on SO which are actually about finding an error in somebody else's code. These questions of course differ in their quality - include code, error messages, detailed explanations, sscce or not. Let's not take that into account right now.

Examples: most of segmentation-fault

• Segmentation fault in my C application
• segfault on second calling of opendir

The reasoning behind closing would be that, since the person is not able (or doesn't want to) provide a SSCE explaining the problem, the odds are it'll be a very simple programming mistake, most probably not useful for other people. Of course there is always a tiny chance that the person has actually encountered something very interesting, discovered a bug in a compiler or JVM.

Another problem is that the topic and contents of those questions have little to do with the actual problem, thus making it almost impossible to google if you have the same problem. Because of that we tend to do the same work again and again.

tl;dr: Should find-my-error questions be closed as a rule?

Answer 1

Such questions might need to be closed, because we've covered that topic quite a bit. These questions generate quite a few duplicates because beginners often don't know what they should be searching for.

This is why I don't mind quite a bit of duplication, because every different way a question like this is asked, it's another way someone might find an awesome canonical answer. Looks like we have a little cleaning up to do there.

Otherwise, provided a question contains useful breadcrumbs that might lead someone to find an answer that relates to the problem at hand, such as:

• Compiler warnings / errors (the first thing many people do when seeing a compiler error is paste it into Google to search)
• Stack traces / Valgrind output (be kind, tell them about Valgrind, they'll thank you!)
• Source code
• Some context to the problem that amounts to them realizing where it's likely happening, but just not why.

... there's no reason to tell someone that we're perfectly capable of helping them with a valid programming problem but won't as a matter of policy, because no beginner anywhere would ever do something as dumb as they're doing. That's how it comes across, and that's not very nice.

In quite a few cases, you're probably looking at a duplicate and this becomes a matter of curating our content more than anything else. The questions are answerable and the folks avidly participating in those tags will eventually connect the dots and flag for duplicates / merges - so I really don't see the harm in answering.

If it's a total stink bomb of a question, then other reasons for putting it on hold apply.

Answer 2

Perhaps it would be a good idea for those more experienced to create a generic answer whenever they find and fix an error in someone else's code.

The main advantage of such generic question-answer pair would be that it would be easy to find it if you knew the keywords being used. That way we could quickly close the question while also providing an useful resource for the OP to use to fix his problem.

Another advantage (IMO) of doing that would be that the "meaningless reputation" gained from trivial questions would be reduced. And the person who created the generic answer would have actually served the community, not only helped a single person.