List questions - another take

Question

[EDIT: this applies specifically to list/poll questions. I opened a separate question about discussion questions in general]

My question "Worst single user experience you’ve had?" just got closed. I don't disagree with that moderation in particular (the wording of the question was probably too negative/argumentative), but it does bring up the question of list, poll, and best/worst questions. Closing the question because a similar question was closed on SO doesn't seem like a very compelling reason (it's citing precedent -- but is that precedent relevant to this this jurisdiction?)

Consistency -- some list questions seem to be OK

Consider the top-voted question on ui.stackexchange "Must-read User Interface Book?". This question has no close votes, and seems to have been endorsed by the community. So the line is not an arbitrary "no list questions" or "no questions that cannot be answered". Obviously, there is some kind of subjective test. Why is a question on good UI books OK, but a question on bad/good UI implementations not? Both are not directly related to a specific issue, but intended to allow a budding designer to gain a broader perspective on the discipline.

The question is relevant and informative

Learning from design mistakes - or design successes - of past UIs certainly is useful. To address Carson Myers's response:

By asking about the worst UI screwups we've ever seen, you're really just looking for some light reading and entertainment (IMO, since the answers you probably expected aren't common and are probably obviously wrong to the community)

Most "worst UI" screw-ups are probably well-intentioned decisions that went wrong, and discussing why will help us become better designers. "Common UI pitfalls" is a different question. I'd argue both are relevant and worth discussing.

For example, Coding Horror's worst UI post addresses developer decisions that were well-intentioned -- wGetGUI's creator made that UI purposefully. Lotus Notes used to be ahead of its curve in user-focused design, one of the reasons it was so popular back in the 80s. What went wrong? IMO, that's exactly the sort of thing this site is for.

Common UI pitfalls is a separate, and equally relevant, question. These are usually made due to negligence or lack of knowledge. Few UI designers would purposefully use a check box as a radio button, but to many users or developers, the distinction is lost.

The opposite question (best user experience you've had) would bring up some good material. There could also be "Small things that make the UI feel better".

"fun" != "useless"

If the argument is "fun" questions shouldn't be allowed (but relevant open-ended could be acceptable), what if the question were rephrased to something like "Design mistakes to learn from?" or "Designs that affected you emotionally?"? I would argue that this would make the question less appealing for users to answer or read, while not significantly improving the scope or quality of the answers. But that argument could go either way.

Promotion/improved traffic

All but one link I've seen in the wild to StackOverflow was to an open-ended question (the "best source code comment" one was actually mentioned in a training seminar I went to). Especially as this site is about to go public, bringing traffic to/promoting this site is very important. Having a question like this could have users to sign up just to answer the question, which would encourage them to return, interact more, and possibly post on a more specific discussion.

Carson Myers brought up in his answer:

Even if it did draw traffic, it might not be the traffic we want, or rather, the traffic might not have the right impression of what this site is.

This is a really good point. I'd argue that traffic that would be interested in reading and discussing good/bad existing UI implementations is exactly sort of traffic we want, and as long as there aren't too many open-ended questions, anyone who explores the site will see that focused questions can be asked & discussed as well. However, if you feel a smaller, more elite community is healthier for the long-run success of the site, then by all means - open-ended and poll questions should almost certainly be disallowed. But be sure to weight the pros and cons before making a conscious choice to shut off a significant user base.

Questions like these a are great way to generate content

Especially as the site is new, and the volume of questions has been dropping lately, expanding the scope slightly will increase the amount of relevant content. This might make the site seem more attractive to potential visitors and increase search rank. When this site goes public (and in the first few weeks) having good content is IMO more important than having focused content (though again this is subjective).

Users like these questions

Look at the top-voted questions on StackOverflow, SuperUser and ServerFault. Many are open-ended/subjective. While "community-driven" != "democratic" (and "never trust the user" is an oft-cited principle in UX!), the users seem to have spoken strongly in favor of these questions.

The interface & site design lends itself to these questions

Features like voting & commenting plus the ready-made community makes these questions a very natural fit for this type of site. This may not be what Joel thought of when originally implementing StackOverflow, but that doesn't change this fact.

The precedent goes both ways

SO may be becoming harsher on poll questions, but it appears SuperUser and Electronic Gadgets are more lenient about them. UXExchange seems to have no problem with these questions, too. Precedent and consistency are certainly important to take into account, but it seems to go both ways on this one, so this argument isn't wholly convincing.

The scope of the question is important

No an argument for or against them, just food for thought. "Worst UI ever" or "Must-read UI books" might be too broad. But there are many list/poll questions that are more focused. "Are there any good resources about designing touch screen interfaces?", for example, has a narrower range of answers than "best books", "best blogs", or "worst interface", and could be considered more useful because of this.

---

TBPH, I feel the "crackdown" on poll-type questions on StackOverflow was a mistake, but obviously a lot of people a lot smarter then me felt otherwise. From what I understand about StackExchange, however, each site is given some latitude to decide what types of questions are OK, and AFAICT there's no StackExchange commandment prohibiting poll, list or open-ended questions. So this site could go a different way on them. Since we're user-moderated, a FAQ entry about them would go a long way towards making clear what's acceptable.

Sorry for the rant -- Spending 11 days cooped up in your room with the flu does that to ya. You've been... WALL OF TEXT'D!

Answer 1

When I read the title of your question, I thought: "this doesn't belong here."

After I read this meta question regarding it, I thought: "maybe this belongs here."

I have to conclude that it was mainly the phrasing in the question. I think questions that are of the type "(best|worst) x you've ever y" are obviously too subjective. If it were rephrased (as you suggested) to something like "Design mistakes to learn from" or even better something like "UI pitfalls everyone should know about" it may have survived. However, it still would have been a best/worst question in my eyes. By asking about the worst UI screwups we've ever seen, you're really just looking for some light reading and entertainment (IMO, since the answers you probably expected aren't common and are probably obviously wrong to the community) and, if you wanted an open-ended, list question of this sort, "common UI pitfalls" might be better.

Searching for list questions in a useful manner doesn't really work. If I google searched "should my navigation be a drop-down box?" on google, and your question had something related in it, I might not find it or be compelled to read it. While list questions can be entertaining and draw traffic, something about this particular question (and most best/worst questions I would assume) just nags at me about not adding much value to the site. Even if it did draw traffic, it might not be the traffic we want, or rather, the traffic might not have the right impression of what this site is.

All else being said, since everything is community moderated, and not necessarily always by the same parts of the community, some off-topic questions may survive while others will surely be shot down. This will probably result in the argument that "question X is allowed, why not question Y!" just as it did on stackoverflow. It's not really something that can be avoided.

In any case, you have been WALL OF TEXT'D as well, my friend :)

Answer 2

I just asked a list question before realising the current attitude towards them; mine is here: What are some great web-based prototyping tools?

I think they're relevant in the sense that my understanding of the StackExchange network, coming from StackOverflow, has always been that it's about replacing inadequate, outdated, or incomplete sources of information with better, more open, more recent and more community-driven ones. Following that logic, I asked a question where I described that I'd like to use the community here to form a resource that is better than the other ones I found, because I think the democratic voting system we have is superior to either anonymous wiki format (one of the links I found) or editorial (the other one, in which case the author is no longer maintaining the list).

For that reason I strongly subscribe to the notion of lists in situations where they produce valuable resources that people around the internet can rely on for accuracy, timeliness and relevance. Best/worst questions are a grey area.

Answer 3

I don't know what the test for "acceptable" list questions is, but here are some thoughts:

It might be that for some questions, the single "right" answer is an aggregate of many items; no individual is likely to have all of the items; and those items must be collected and sorted by several members of the community. (Hence the unwritten laws that list questions should CW, there should be one item per answer, and it's okay to post multiple answers.)

Another theory is that a question may have several answers that are not mutually exclusive. While you may have to choose between jQuery, Dojo, and Ext JS, you'll never be forced to select just one of several Hidden Secrets of JavaScript.

Also I think the answers should be useful to others. Entertaining answers are good as long as they clearly have practical applications. But that one applies to normal (non-list) questions too.

The book question passes all of these tests.

Answer 4

I just tried to use chat.meta.stackoverflow.com (third place). I hated it.

I love StackOverflows UI for discussions (not just specific answers) and can't stand this arbitrary rule.

Solution: They should just add a checkbox to mark the question as open-ended and let people filter it if they want. The UI shouldn't be different. It works perfectly for discussions AS IS.

< analogy >

This issue is like Same-Sex Marriage. If you say it can only be one way (Man/Woman), then you are happy with your one way. If someone else LIKES MARRIAGE but wants it another way (Man/Man or Woman/Woman) then they are S.O.L. They don’t want some messed up “third way”, they want to be married! Stop being control freaks and let people do what they want to do!

< /analogy >

Answer 5

I still disagree with you. The new 'third place' was invented for these things. I think the main site should remain both serious and concrete; no questions without answers that are helpful.

Now this last point is debatable in the case of this question, but 'list of' questions in principle ask for multiple answers, whereas we should try to ask for one correct answer. That answer may not exist due to multiple viewpoints, but it should be what we strive for.

Answer 6

To ensure that your question is not closed preemptively, it may be beneficial to phrase it in a way that is concise.

While this won't increase the relevancy of your question, it will decrease the possibility of misinterpretation.