What's going on with icon questions?

Question

tl;dr: Lots of icon questions seem to have been closed in a way that doesn't match the FAQ, doesn't match the existing meta discussion, isn't consistent, and doesn't seem to make sense in terms of what is and isn't UX - with no explanation. What's going on?

---

Edit: There seems to be some confusion about the focus of this question. When I refer to "gimme icons" questions as "I can't think of an icon for [app], think of one for me, I'm lazy" questions, I am not arguing that these are great questions that should be allowed. The focus of this question is, like the title, "What's going on with icon questions?". My argument, if there is one (this is a question after all), is that for reasons described below, new users (like me) are not currently given a clear picture of what sorts of icon questions are allowed - the FAQ, meta thread, and reason for closing all seem to be saying different things.

---

I thought this site - a site of UX experts - would be a good place for expert opinion and fact sharing about the communications and UX aspects of icon design. Things like...

• Knowledge on how certain common UI elements perform in usability testing
• Rationale on why certain iconography trends work in terms of easing user experience, and why others don't and confuse users
• Caveats about under what circumstances common visual metaphors that often work actually aren't appropriate (and vica versa) - e.g. important generational and cultural differences that UX people should take into account

...as opposed to the design side of how icons should look. I was really surprised to see that...

1. Every 'icon' question looks to have been closed - even with loads of upvotes and good answers
2. There's a meta discussion where the consensus seems to be that app-specific icon questions should be closed as 'too localised' but (in Ben Brocka's second answer) general questions like those above about the user communications side of iconography should be allowed
3. None were closed as 'too localized'. They were all closed as 'off topic', and none that I saw had any kind of rationale or explanation for closing left as a comment - just abruptly closed

What's going on?

I came to this site because I wanted to see what UX experts had to say about the common convention of using arrows like ► ▼ to indicate that there is hidden content that can be expanded (►) and that the below content can be hidden (▼). It's a visual method that works for me, but I wanted to be sure that there is a good rationale and (ideally) real evidence that it works for people with lower than average computer literacy before relying on it (and if there is a known problem with it that I'm not aware of, naturally it'd be useful to know if there are alternatives that are known to avoid this problem).

Sounds like a reasonable question to ask a UX expert, right?

So why on earth are questions about what is an appropriate visual metaphor for a general user experience problem (such as clearly communicating common functions like delete, show/hide, moving in time) closed? That last one even has a pile of comments from people pointing out that it's a genuine UX question.

The FAQ explicitly says "the subject of icons is on topic" (while explaining, rightly, that "I can't think of an icon for [app], think of one for me, I'm lazy" is off topic). It may be true that the askers are usually in the process of trying to make an informed iconography choice - what else would prompt someone to ask? - but the essence of the questions are "What's known about what works?" not "Gimme icons, I'm lazy". Any ambiguity can be fixed by editing.

And why does it seem to be so inconsistent? This question here is along precisely the same lines - asking about the suitability of a certain visual metaphor - and isn't closed. What's the difference? Is the difference because the 'floppy disk' question invoked the name of some established UX Guru?

Answer 1

What Jon and Ben said already is very true, but I would just add that for most of these things, there's a "right way" and a "wrong way" to ask questions. For instance, you said:

I came to this site because I wanted to see what UX experts had to say about the common convention of using arrows like ► ▼ to indicate that there is hidden content that can be expanded (►) and that the below content can be hidden (▼). It's a visual method that works for me, but I wanted to be sure that there is a good rationale and (ideally) real evidence that it works for people with lower than average computer literacy before relying on it (and if there is a known problem with it that I'm not aware of, naturally it'd be useful to know if there are alternatives that are known to avoid this problem).

The "wrong way" to ask this question is

"What's a good icon for showing/hiding hidden content?"

We get this type of question all the time, and it can't be salvaged into a good question with a lot of work.

The "right way" is to ask the following:

User acceptance of disclosure triangles: It's a common convention to use arrows like ► ▼ to indicate that there is hidden content that can be expanded (►) and that the below content can be hidden (▼). However, is there any evidence that average to below-average users understand this convention and make use of it? Are there any known problems with this interface element, or alternatives that perform better? ...etc...

Answer 2

The floppy disk question is on topic because a 'correct' answer potentially exists - is this metaphor out of date? (the correct answer being either 'Yes, because of X' or 'No, because of Y'). The other icon questions are purely asking for suggestions as to what type of icon is a good representation for X. There is no correct answer to questions like that, hence them not being constructive questions.

User Experience isn't about what the nicest sort of icon is, or asking for people to design an icon for you. These aren't UX problems that can be given a correct answer.

Going directly to the FAQ, one of the basic StackExchange principals is to ensure that questions are not asked where:

every answer is equally valid: “What’s your favorite ______?”

This is the type of question that icon requests fall into.

Answer 3

I'm going to try to answer my question with a summary of the on-topic various things that have been said in the above answers and comments.

• When icon request questions are closed with this rationale:

Off topic. Questions... are expected to relate to user experience within the scope of the FAQ

...this does not actually mean that questions about the suitability of icons are not related to user experience. It should be read as:

Not constructive. As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or specific expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, see the FAQ for guidance.

• This is consistent with the real reason why those icon questions that have been closed were closed.
• Both the above are consistent with (albeit not currently clearly explained by) the FAQ. Icon questions are on topic. Icon questions that are not constructive (including the particularly common case of icon requests) are not allowed as per usual SE rules. UX.SE closes these "Not constructive" questions with 'Off topic' as an explanation because apparently, this leads to less arguments.
• The meta discussion I linked to (and which, confusingly, the FAQ links to) is obsolete. Since that discussion, the decision was taken that it doesn't matter how localised an icon request question is: icon requests, even when they're about common universal needs, are usually "not constructive", and so are therefore closed
• When a user posts a question about how to communicate a meaning to users, it is up to them to make it explicitly clear that they are asking for facts and principles to inform a choice, not a list of icon suggestions. If they are not explicit enough and a question is closed with the explanation "Off-topic", it is up to them to figure out that the problem is that the question is not constructive and doesn't afford a concrete answer, and once they have figured this out, they should edit their question accordingly.
• UX.SE moderators will usually assume questions that resemble icon requests to be not constructive, and will tend to close them rather than editting them, because there are so many of them. If a question that looks like an icon request question does contain a real, constructive question that can be answered with facts, users can edit the question such that it is no longer "not constructive" and propose (flag?) it for re-opening.

Examples of types of icon questions that are constructive include:

• Questions about specific UX implications of a specific type of iconographic representation (so long as it's sufficiently general to not too localised). For example answers to Save icon, is the floppy disk icon dead? that suggest better visual metaphors will be constructive in a way that answers to "What's a good save icon that's not a floppy disk?" won't, because they will need to demonstrate specific advantages over the representation specified in the question
• Questions about a specific tradeoff or choice. For example, answers to Should a toggle button show its current state or the state to which it will change? that suggest better icongraphic alternatives will be constructive in a way that answers to "Best way to make a toggle icon" won't, because they will need to demonstrate how they compare to the pros and cons of the choices in the question
• Questions about how to represent something that make specific reference to UX problems and concerns and that explictly ask for guidance and principles, not icons. For example, answers to How to graphically represent a language that suggest a type of icon that solves the problem will be constructive in a way that answers to "Icon to graphically represent a language" won't, because they will need to show how that icon type addresses the specific problems and concerns that were written into the question.

Questions that resemble an icon request, but contain a real, solvable UX problem, should be editted so that there is:

• A clear outline of the problems to be solved that answers can be assessed against
• Indication that there is a real UX problem and not just a failure of the imagination or research of the asker
• Where possible, examples of existing iconography and explanations of in what way these don't solve the problem

...and, because icon questions are so frequent, moderators shouldn't be expected or relied on to do this.

Answer 4

Icon requests are off topic; that's the rule. It's just not a constructive area, as evidenced by the dozens of different icons used for dozens of common functions; there just aren't correct answers in 99% of cases.

Another problem is with the "expert opinions"; as a Q&A site, we solve problems rather than just share opinions. Yes a good deal of opinion is sometimes involved, but it should only exist as tangential to the answer. Answers should be at least "This works, and it works well in my opinion" rather than "My opinion is X." The latter is just polling for opinions, which just clutter up the place. We're here to solve problems, not muse endlessly. The classic problem with forums is that they muse endlessly on the same stuff; we don't want that here.

So why on earth are questions about what is an appropriate visual metaphor for a general user experience problem (such as clearly communicating common functions like delete, show/hide, moving in time) closed?

Those three are all pretty clearly "what icon should I use". Unless there's a painfully obvious answer like "use X, the standard icon for Y forever", it's just a matter of opinion and thus Not Constructive. Having the rule just limits these arguments.

I'm not seeing your concern, honestly; I don't see this "too many icon questions are being closed" issue. Please list some examples of questions you think have been closed but aren't in violation of the "recommendation" problem, otherwise it's hard to tell what your actual argument is.