Given that most of my work currently revolves around Unity, I have recently become a bit more active in the tag. Most of the questions use some form of C#/UnityScript/Boo and are simple enough to answer.

Problems arise when it's not necessarily a script that exhibits a problem. A lot can be done inside of Unity by hooking scripts (or components/MonoBehaviours) to GameObjects which are nodes in a tree. There you leave the realm of programming to some extent. Or you rather enter the realm of visual programming perhaps. Problems might well arise due to your particular tree structure, the attachment of components, or settings entered in a UI/Editor instead of a bit of code.

I have been tempted to ask for a small sample project in several cases, to be able to open them and have a look. The user could share it online, and in many cases the problem is trivially demonstrated and easy to resolve. But there is no real guarantee that the file will remain available, and you can't really see what the problem is from the visual content of the question alone. (This of course besides the generally iffy practice of opening unknown content found online) Of course proposals have been made regarding SO/SE-hosted attachments, but those have been unsurprisingly and understandably declined.

Alternatively one could imagine asking for screenshots of the situation, but that gets messy and complex quite fast.

So the question I have is, how can we support cases like this without ending up with tons of questions that no longer demonstrate the original problem? Are there other tags/communities of SO I'm unfamiliar with, which have solved a similar situation in a reasonably supportable way? Is there a good practice to recommend and perhaps even to include in the tag wiki?

Granted, some might argue it's no longer a programming problem at that point (I wouldn't agree, but opinions can differ). Even so, I still think it's something to consider. Deferring such questions to a site like SU as a general software usage problem doesn't solve this situation. (Well, other than it being their problem from then on. ;) )

  • There is an SE site specific to game dev - gamedev.stackexchange.com. Perhaps try there or the meta there? – chue x Dec 14 '13 at 16:57
  • That is not all that relevant to the issue. It would only shift the question to another site, and I'm not sure it's off-topic for SO to begin with. – Bart Dec 14 '13 at 17:17
  • 3
    You can take a look at a lot of the Xcode questions tagged Interface Builder to get an idea of it. Most people just describe what to do in text ("Click on the Attributes Inspector in the top right corner and check the box labeled...") but I don't know how well that works with tools that are more visual than a bunch of fields and checkboxes. – Linuxios Dec 14 '13 at 17:23
  • @Bart - Point taken, but maybe they have more experience with these types of questions there. There definitely is overlap in terms of what questions are valid on GD vs SO. I would say Unity related questions would be more appropriate for GD. – chue x Dec 14 '13 at 17:24
  • @Linuxios interesting. I see several of those indeed also go the screenshot route. Though that's particularly useful for answers, more so than for questions it seems. – Bart Dec 14 '13 at 19:11
  • 1
    Unfortenately Unity3d project are much larger than the maxium allowed post size of 30000 bytes. You could imagine that you base64 encode rar files and include those in the post. Not only is that a huge misuse of features offered for larger files you have to split acrosse multiple posts (rar in files of 20K, giving a bas64 encode of nearly 27K) were it all becomes very messy...a size-limited upload feature hosted by SE would be preferable. – rene Jan 20 '14 at 14:59
  • On MSO there is a post about storyboards in iOS, seems related: meta.stackoverflow.com/questions/313330/… – rene Dec 24 '15 at 8:45

If a question cannot be made sufficiently clear using the options provided by the site, it's more than likely simply not a fit. Questions like this have several issues.

  • They are "debug this for me" questions. If you are unable to show us the code or take the specific screenshot with the information we need, you've not narrowed it down enough. We shouldn't have to go through your entire project just to figure out what's wrong, even if that project is not all that large. So do your own debugging and narrow your problem down to something you can demonstrate. Dumping your project and stating "this does not work, what's wrong?" will not do.

  • They are not useful to the wider community. Hooking up components in an incorrect manner, messing up your hierarchy or entering wrong values are in essence the equivalent of typo questions. They would be "too localized" if that close reason was still around.

And you know full well that the only "solution" to the issue you raise would be some form of permanent file hosting support. You can skirt around the issue, but that's what it comes down to. And that solution has several problems. Disregarding the hosting itself, which could come with a size limit, how do you ensure that all uploaded content is safe? How are you going to respond to the question "Hey, I wanted to help this user and now my computer is summoning the devil". I don't particularly like to download stuff I don't entirely trust. And to put that responsibility on the site is not fair.

So all in all I'd say that if a question can't be asked with the features currently available, it's not a fit for Stack Overflow, and unfortunately neither for any other site in the network. Asking "how can we support them" is thereby irrelevant, because we shouldn't really support them any more than we already do.

  • 2
    There, let's see if this sparks some discussion. Two bounties don't seem to do it, so there you go. :) – Bart Jan 23 '14 at 22:50

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