Open-ended & discussion questions - another take

Question

On "List Questions - Another Take?", in a comment thread, I asked why we should strive for one correct answer. Jouke van der Maas answered,

Because that's what the site is for. If you don't like that, you should go to some discussion forum. Trying to get the single best answer to a question is what this is all about.

Here at meta.ui, we're trying to decide what the site is for! Saying "that's the way it is because that's the way it is" is begging the question. So rather than working from the assumption that we should strive for a single answer, we should decide on policy (and a FAQ entry describing this) that best suits the users. Relying on notions of "site integrity" or citing the precedent of StackOverflow doesn't serve the users, and UX is all about user-focused design. Precedent is a useful tool -- but it can be harmful, too.

UX questions are subjective

---

The current FAQ says:

Avoid asking questions that are subjective, argumentative, or require extended discussion. This is not a discussion board, this is a place for questions that can be answered!

The vast majority of questions on here (and about UX in general), however cannot be answered. For a little evidence of this, here are the (at the time of original posting) top 15 questions by votes, categorized:

Objective Questions

These are the type of questions that are ideal for StackOverflow, and the FAQ and other StackExchange sites are geared towards.

• Should the OK/Cancel buttons be aligned right or centered?

Discussion Questions

These questions are ones that could have a correct answer, but that answer (like most things in UX design) is "it depends". Most of these questions have become discussion threads.

• For websites, is it better to have a variable width layout or a fixed width layout?
• Are radial contextual menus better than vertical list menus?
• Whats more user friendly local authentication or remote authentication (OpenID, OAuth)
• Do you need a search button with a search box?
• Is it better to prevent a forbidden action or display an error/explanation message?
• How to visualize the possibility of drag’n’drop?
• What can be done to make a long, multi-step wizard more user-friendly?
• What is an acceptable response time for my ajax ui?

List Questions

These are the open-ended/list questions that "List Questions - Another Take?" was about -- questions that are of general interest but cannot have a single answer, and do not have a direct application.

• Must-read User Interface Book?
• Which are the best UI related blogs/sites?
• Common web app usability gotchas?
• Are there any good resources about designing touch screen interfaces?
• What systematic methods there are for designing user interfaces?
• What resources do you use to find good color combinations?

I'd say this is strong evidence that the one-question-one-answer model of StackOverflow will not work here. This site is being used for its intended purpose -- ex I was able to get an answer to an actual design question I had within a day. So are the users wrong for using the site as a discussion forum instead of a Q&A site? Or are the FAQ and zealots wrong in assuming it's a Q&A site.

_{(there's some blur between the categories, and you may disagree where a question should go -- I'm making this CW so y'all can re-catgorize them if needed)}

Chat is not a solution

---

I've seen a few people suggesting that the right place for these questions is chat. I'd like to counter that argument right away, because I think chat is the worst possible place for them.

• Not persistent -- every time the user has the question, it must be asked (and answered) again
• Not searchable
• Not discoverable -- there's no way to find out what or how to ask
• Answers cannot be updated and refined
• In a live environment, people will spend less time researching/considering their answers
• Fewer users
• Badly organized -- comments are not tied to an answer
• Noise in environment
• No voting -- no way to recognize what answers are better
• No linking -- results can't be shared or blogged about
• No bookmarking
• Doesn't help our Google rank

If you're still not convinced, here's a scenario. Joe wants to know a good book to read on UX. So he goes to chat.

JoeThePlumber: what are some must-read books on UI design?

John_MC: About Face 3 by Alan Cooper is my favorite. It focuses on desktop apps but is applicable to web design too. But it doesn't have enough pictures, so it's hard to visualize.

SarahP: u shuld reed jurasic park it proves dinosurs lived with ppl and has a lot to do with user interfaze

AfricanPrince419: hi i am the son of a nigerian dictator and have 32 million dollars but need $1000 transfer fee you can get half the money

Joe has received two answers now, rather than the 20+ he could have. Both answers are of lower quality. He has no way to decide which answer is better. Then, the next day, Barrack wonders the same thing...

What can be done?

---

If this site is going to be successful we need to choose a different model than StackOverflow -- and we need to do it soon (before we leave beta). Here are some ways:

• Update the FAQ and remove the above-quoted passage, ASAP
• Stop the "one-answer-only" police from attacking harmless discussion questions.
• Give reputation for community wiki posts/answers (especially if no one else has edited them), to encourage more stuff to be CW
• Since fewer questions will have a single answer, don't nag users to accept an answer
• Don't show a user's accept rate

The tribe has spoken. Focused, discussion-centric questions can work on a Stack Exchange site! Perhaps this was not the original intention, and perhaps somewhere Joel Spolsky is muttering to himself, plotting his revenge. But we, the users, can make this site the ultimate community for user-experience experts. We will stand <strong>strong</strong>. We will persevere. And we will be victorious. Viva La Revolución!

Answer 1

I agree with Robert's POV, disagree with Jouke's, and would like to add that systems (any system: political, evolutionary, mathematical, web-based) evolve because the components of that system push against their boundaries.

Hence, I support us stretching the definition as much as possible of what this site should be. It's what made StackOverflow grow in the early days with all the heated discussions about what was or wasn't allowed, and it's what will make UI grow as well, given that Robert's right that UI questions aren't objectively answerable in the same way programming questions are.

Where I work, I evangelise to my coworkers the following ideology: if you want to get something done, do it and if someone has a problem with it, deal with them then. Don't ask first, because if you get "no" then you're starting on the wrong foot. I think we should apply that ideology to UI as well: let's keep trying new things and if they break, we reevaluate. If they don't, and people find them useful, who are we to say they shouldn't be done just because StackOverflow didn't do it? It's a different subject, in a different context, with a different community.

Answer 2

From my perspective, I started out pretty enthusiastic about this exchange, but the more questions I read where moderation intervened ("This is for stack overflow" especially), the less interested I am in participating.

I'm not offering a solution, just surfacing my own reaction to the moderation. I would prefer to to err on the side of a relaxed policy versus a strict. Keep the trolls out, and help the newbies learn how to write better questions, but don't sweat the questions where some subject boundary may be gray, or even crossed occasionally.

Answer 3

I agree with you on all of these points except some about what the chat is good for. I've raised this concern in the chat feedback and I was told that using chat for largely off-topic questions was already being discussed internally.

I don't disagree with list/subjective/whatever questions entirely, especially for this site, but I think that some of them are still off-topic, and among those is the "best/worst x you've ever y" questions, which were in part the subject of the meta question you referenced. I don't know about the searchability of the chat but I know starred questions can sort of work like votes, and the threads will be preserved unless there is little content in them. Every question you cited here, I personally agree belongs on the site. I just think some would be better suited for a more discussiony place.

All in all, the SE engine was designed for professionals to get answers relating to their profession. If a question is answerable, and somewhat professional (ie. what to do when x? Which y works the best here?) then it belongs. But if it's just looking for a fun list (worst UI you have ever seen? stupidest client ever? etc) I think they are off topic and the chat merely provides a place we can talk about that, informally, anyway.

In any case, what if that "worst ui ever" thread was titled something like "what are some serious UI decisions that can make/break an interface" (or something) and then there was a chat thread attached to that question where people would put things like, "Don't be like Lotus Notes! [picture]". These are just suggestions, I'm not necessarily stuck to this model, but I would rather not avoid it entirely -- utilizing the chat to better facilitate chatty, discussiony questions seems like a good idea, worth a shot.

Answer 4

Love your idea:

Give reputation for community wiki posts/answers (especially if no one else has edited them), to encourage more stuff to be CW

This is a really thoughtful post, thank you for taking the time.

Community Wiki Posts Need More VIsibility / Clearer UX

As a Stack Overflow addict and user of several SE sites and a self-titled "User Advocate" and web developer, I was super excited to see a UX forum on SE. For one simple, selfish reason: I think the community wiki feature needs some serious UX love and I assume, perhaps naively, that a burgeoning UX audience would feel compelled to address this.

There's no obvious way to search in the wiki posts or to generally browse one type over the other. Such a function may seem divergent from SE's core strategy to be a Q&A site, however, if u build it (awesome), the (awesome) people will come and there's a wealth of excellent posts and experts to be leveraged.

I understand why it is the way it is now and that wiki vs. q&a is a major diversion, but I also feel it's a major lost opportunity. SE has big mind-share right now - they could dominate not just in Q&A but in general reference for their sites as well.

Subjective Topics

Agreed, UX is subjective. Yet Q&A is applicable as folks have many Qs about it. Since there is often more than one (subjectively) acceptable answer, maybe the ability to accept more than one answer would be a great solution. This would be a SE platform feature request, but something they could enable only for sites that require multiple good but different (divergent even) solutions / perspectives.

Answer 5

This issue represents the number 1 problem with StackOverflow's model. At this point, I am totally confused with meta vs. normal and where to ask what. My head is spinning.

UPDATE:

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.

I totally 100% agree with the solution presented in the question. For this site, I think changing the rules makes total sense and the three categories presented are all perfectly good questions.

< 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 >