2

If one runs into a serious and/or annoying bug in a piece of commonly used (open source?) software, is it okay to ask a SO question asking others to verify the bug and/or suggest workarounds for it? If not for software in general, how narrow does it have to be? Must we limit it to libraries and such used in programming?

Perhaps the straightforward thing to do would be to find the bugtracker for the software if it's open source, and post something there. It would certainly make sense to do eventually. But the traffic there is low, and you might not get a response for a while, and even less chance for a fix or workaround in a reasonable amount of time.

However, by posting on SO, I believe there are many benefits:

  • Almost immediate, high-traffic response
  • Many-eyes analysis of the problem
  • Possible multiple solutions to get around the bug
  • Generates permanent documentation of the bug that you can send somewhere later

Some might say this type of post is inappropriate for SO. What is the consensus on where to draw the lines?

3 Answers 3

3

I think it is appropriate if the bug pertains to things within a software developer's domain (and perhaps a developer's tools, though I would be more careful with that).

Inappropriate Bug Question
"MS Word hangs when I format a paragraph with a sans-serif font."

"Bug" Question that Shows Misunderstanding/Lack of Effort
"x + y != 2. But I'm sure both x and y are 1...is there a bug in the compiler?"

Likely Appropriate
"I'm developing an add-in for MS Word, but when I cast IFoo to IBar a runtime exception is thrown, but only when the class implementing IFoo also contains a static constructor. Here is a snippet which reproduces the problem on .Net 4.5"

I think it is especially appropriate if the developer is not 100% sure that it is indeed a bug but has a clean, reproducable test case about which they can formulate a clear question.

This also assumes that the prerequisite research has been done (i.e. the issue isn't already clearly documented elsewhere). In fairness, the more stable a library the more obscure the bugs tend to be, so it may be hard to search accurately.

Many bugs in both open and closed-source code such as this one have been identified, explored, and even escalated to the right folks (Eric Lippert, in the aforementioned case). I've seen folks like Phil Haack and John Resig weigh in from time to time.

Escalation like that certainly doesn't happen with every question, so I feel it is also useful to report the bug in the appropriate places. Even if the author happens to respond, the issue should be entered in their bug tracking software for consistency and ease of access. No one wants to be collecting stray reports from unstructured sources.

3

If your intent in posting here is to get help fixing the bug in code, then I say go ahead. But if you are only hoping to bring attention to or verify the bug, that might not be so helpful.

Only exception to the above: If you are trying to use an existing library in your code, and you suspect your problem with using it is due to a bug, then it might be OK here because, again, your question would be about code.

But if, for example, there were just some user-level bug with, say, Firefox and you want to know if others experience the same bug, that's not on-topic on Stack Overflow - again, unless your intent is to fix that bug.

2

"commonly used (open source?) software" is a bit vague, for the rest of the answer I'm assuming you are talking about a piece of software that somehow relates to programming. If not, I'm not so sure why you'd want to ask about it on Stack Overflow.


First:

  1. Find the project's issue tracker
  2. If the bug isn't reported, report it

Then:

  1. If there weren't any workarounds suggested on the issue tracker (or the bug wasn't reported at all), feel free to ask on Stack Overflow for workarounds (don't forget to link back to the bug report).
  2. If there were workarounds on the issue tracker, but they didn't work for you, free to ask on Stack Overflow, but please explain in detail why the proposed workarounds didn't work for you.

If you find a great answer on Stack Overflow, you'll get bonus unicorn points if you link to it in the bug report.

Also:

If you found an awesome workaround, you might want to risk posting a Q&A combo. The community generally frowns upon self answered questions that aren't stellar, so if you decide to go this way:

  • Your question should stand on it's own,
  • Your answer should be awesome.
5
  • I don't understand your last statements. That link seems to imply that more self Q&A should be on StackExchange sites, and is actually encouraged. What's with the secret "frowning upon" self answered questions?
    – Andrew Mao
    Commented Feb 28, 2013 at 7:14
  • @AndrewMao Self answered Q&As is a good thing. But the question must be a good question in itself, it won't be treated differently by the community because it's paired with a good answer. If the question is crap, it will be closed and eventually deleted. That's something that several people who posted self answered Q&As missed, (probably) thinking that since they are giving a good answer, the quality of the question doesn't matter much.
    – yannis
    Commented Feb 28, 2013 at 7:23
  • @AndrewMao Furthermore, because self answered Q&As may generate more rep for the author than questions or answers (as there are two posts by the same person to vote on), the community is generally a bit more strict with them.
    – yannis
    Commented Feb 28, 2013 at 7:24
  • I only post self Q&A when I actually have a question and then end up solving it myself before someone else does. It never occurs to me that I might have a great Q&A to share with the community.
    – Andrew Mao
    Commented Feb 28, 2013 at 7:25
  • @AndrewMao Then you are doing it right, and you should continue doing the same.
    – yannis
    Commented Feb 28, 2013 at 7:26

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .