What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 149 Stack Exchange communities.

Stack Exchange is not completely lacking in wiki features. We have "Community Wiki" posts. We have "tag wikis", we have collaborative editing. However, for some reason none of these uncover the true wiki power IMO.

I want to see Stack Exchange littered with canonical answers. Answers which are the de facto dupe, which give a user everything they possibly need to answer the question. Answers like these, and these, and these. Whilst Server Fault and pockets of the Stack Overflow community have made efforts to achieve this, it's always seemed (to me) an underused aspect of the system, and almost a hack rather than the system supporting such behaviour.

Enter a ...

This is in four parts (and here's the TLDR);

  1. Give users the ability to mark a post as a "Collaboration".
  2. Allow users to find Collaborations and add their wealth of information to them.
  3. Allow users to mark their improvements as needing review.
  4. Add a review queue which is just to vet the high quality content we want this system to have.

Now, you're sat there wondering whether I'm high on something, because we already have collaborative editing and community wiki posts; however:

  1. We are actively discouraged from adding huge realms of new information to posts via collaborative editing — "hey dude, you're putting words in the OP's mouth there; that's a radical change and a no-no".

  2. Community Wikis don't really get used in a wiki style. Someone marks their post as "Community Wiki" in some half lame attempt to say "hey dude, if you happen to come across this post, and fancy adding some info to it, feel free, but I'm not bothered either way."

    ... except no-one does. It still feels like you're encroaching on someone else's post/property. What if the information you want to add is wrong? What if the edit isn't up to some of Stack Exchange's wacky high standards that are hard to discover by yourself?

So, allow me to continue:

  1. A user marks their post as a "Collaboration". Either via the existing CW, or via an additional button.

    This detaches the post from the user. The initial user is no longer promoted, to diffuse the feeling of ownership and property.

    screenshot of proposed layout

  2. A review queue (or something similar) is created, which lets users find "Collaboration" posts they are interested in. Yes, this is basically a search, but it promotes this kind of action to people.

    • Either via the "filters" that already exist in the Review Queues
    • Or the system intelligently finds posts, like this

    The options are "Improve Post" or "Skip".

  3. If the user chooses "Improve Post", they are allowed to improve the post. Add information, remove information and collaborate wiki style.

    Users can optionally flag their post as needing review. Edits made by users < x rep get forced into needing review.

    screenshot of proposed review-needed checkmarks

  4. A queue (the same queue?) exists which allows people to review improvements to these collaboration posts. This will be separate to the suggested edit queue, as radical changes are OK, and we want the content to be vetted, not just the edit.

    Again, maybe the content shown to users in the queue can be filtered like in #3. People can "Approve", "Improve" or "Reject" the proposed changes.

I propose that tag wiki/ excerpts become part of this new review system. The "needs review" properties are relevant, and it is more wiki style than the simple questions & answers.

Now, I really envisage this being high quality content. Higher than anything Stack Overflow or the Internet has seen before. I'd like to see the bar at being allowed to submit content without needing review, and the ability to review content, being set extremely high; 10k, 20k privilege (and hey, the 20k privileges are pretty lacking at the moment).

I appreciate nothing I've really said is radically different from the current SE, but I really feel like it instills a new wiki style of editing to the Stack Exchange network.

Comment away.

share|improve this question

5 Answers 5

One of the things that sets Stack Overflow and friends apart from other sites is the question and answer focus. I think this is why it is more useful than most wikis, and draws out so many detailed answers. Moving towards Wikidom reduces that which distinguishes this site. Great answers are rewarded with reputation and badges. Your examples all are well rewarded, resulting in hundreds of points of reputation and one or more badges each.

Your suggestion also deviates from the Stack Overflow ethos in that there is no rewards system involved. The dangling of a numerical carrot quantifying prestige is also a big part of what makes Stack Exchange sites tick much better than most wiki sites (with a few exceptions, such as Wikipedia).

I agree that the community wiki feature seems ineffective, but I would rather see it just disappear, and let the Stack Exchange folks focus on tuning the Q&A rewards & moderation systems. Adding more wiki mimicking features seems like feature creep to me.

For example, if there were a once a day or once a week "super vote" or micro-bounty that folks could award to really good answers that would be more interesting than wikifying things.

share|improve this answer
your last paragraph almost made me not up vote, but you were only comparing those to "wikifying" things more, so I won't be harsh! ;) –  Andrew Barber Jul 24 '13 at 3:54
User A posts a simple question that's been asked a billion times. Users B, C, and D post a quick answer in the hopes of easy reputation. If there was a canonical post that could be linked to, User A would get better quality help. The goal should be to help User A, not preserve meaningless reputation boosts to users B, C, and D. –  jmac Jul 24 '13 at 4:28
The wiki system I'm proposing will still retain the Q&A style of SE; it will be just answers that will be "wikified", rather than articles such as MDN/ wikipedia. All of the answers I linked to are CW, and therefore no reputation has been awarded (or badges?); which to me shows that not all people are driven by reputation/ badges. I'm not completely opposed to rewards been dangled though; maybe +X rep for an edit, and badges for X edits in total. –  Matt Jul 26 '13 at 11:13
@jmac: And yet, finding originals for duplicates (finding the right, useful original) is frequently non-trivial. So enhancements making doing that a rep-building exercise (and there have been some suggestions periodically) are worth looking at. Because people crave acknowledgement, and SE does that with rep. –  T.J. Crowder Jun 1 '14 at 18:02

I was thinking the same thing myself a few months ago. It's really a charitable act, since tag wiki edits garner two points and community wiki edits no points. Highlighting outside usage of excellent solutions, such as this:

would help motivate users to contribute.

share|improve this answer

Here's a further refinement: anyone who wishes can edit the post and then sign up to get notifications on comments/revisions/edit suggestions to the post just like a post owner. If you add that in, you would, I think, fix one of the most puzzling things about current CW: only the original poster gets all the notifications for that post!

In effect, then, a Collaboration would have multiple owners in just about every respect — not merely past drive-by contributors, but those who are actively interested in maintaining that post.

For edit suggestions specifically, I'm envisioning Collaboration owners that have CW privileges but don't have edit privileges getting the ability to review suggestions in the usual multiple-reviewer queue, as though (for that post) they were an ordinary 2k reviewer. And owners that have edit privileges would be able to unilaterally approve/reject changes just like any normal post owner. (Collaboration owners without even CW privileges would still get other notifications, but couldn't approve or reject edits and would [probably?] still have to make suggestions.)

share|improve this answer

I think questions should be titled in a way so they're not too broad but not too specific to only relate to the exact instance of the problem either, in consideration of people having similar issues but different scenarios, so as to ease and promote searching for an answer in someone else's question before posting a new question.

If I were to ask a question, "How do I properly target a list element nested in a parent container, I'd bet that the first answers are going to be nothing more or little more than a block of code that solves the issue which is of little use to someone if they don't understand jQuery to begin with.

If I were to answer and explain how to traverse up the tree, and into the parent container, and rephrase the question to read, "How do I traverse up the tree and into a container with jQuery" then the title still relates well enough to the question as to catch the attention of anyone capable of answering but by renaming it and broadening the scope of the answers it becomes a better and easier resource for people to find when searching in the future.

Otherwise you end up with questions How do I target this, how do I target that, how do I target this other thing, ad infinitum...

share|improve this answer

I think this is quite hopeless: StackExchange is nothing like a wiki.

Maybe some specific implementations could be improved, judge yourself by looking at the list of differences.

share|improve this answer

You must log in to answer this question.

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