7
$\begingroup$

I have been thinking about this for sometime. As you know, SE has changed the Community Wiki policy and now only moderators can make a question a CW. Community Wiki has 2 main effects:

  1. The question becomes owned by community, the original OP does not gain reputation,
  2. The required reputation to edit drops to 100,

and a side effect:

  1. The answers posted after a questions becomes a Community Wiki become Community Wiki automatically, i.e. the answers are owned by community, (but answers posted before the question becoming CW don't).

These three effects are not always correlated and there are cases that one wants to have one of these three but not the other two, e.g. the owner does not want to earn reputation but wants the users answering the question to earn reputation for their answers. So it seems to me that SE is mixing too many things in one concept and that is the main reason for many issues we had and have with CW policy.

Robin brought up the idea that we should make the non-CW answers to a CW question CW also. I have a problem with answers becoming CW when the question becomes CW because sometimes the answers deserve to earn reputation, but in the current situation it makes sense, if the answers need to be CW then the previous ones also need to be CW, on the other hand a user might not have answered a question if it was a CW question so I don't know what is the right thing to do here.

I see three clear reasons for making a question a CW (and it would be nice to add other reasons to this list):

  1. OP doesn't want to earn reputation (for some reason),
  2. OP wants to allow others to edit,
  3. the question is a ranking list.

For example I think a question like this and this should be CW but the answers should not be and the users answering them should earn reputation for their answers.

Another issue is that the new policy is not clear on what we should do when we think a question or answer needs to be CW, I used to flag some posts for moderator attention but don't do it any more because I now think it is not very nice to suggest someone else question or answer should be CW to the moderators, so I have decided to go with a softer way of adding comments like "I would upvote it if it becomes CW".

I have an proposal that I think might solve some issues and would like to hear what others think about it:

In general the question or answer should be made CW only on the request of the owner (with some exceptions, see below), other users that think the post should be CW can leave a non-abusive comment like: "nice answer, I would upvote it if it becomes CW" and up-vote only if/when it becomes CW.

Exceptions:

  1. the OP is a drive-by user and does not reply to the answers or comments after considerable time,

  2. some clear cases like: a ranking list. These should be explicitly included in the policy,

  3. other special situations where the moderators have strong reasons for making a question CW.

What do you think about this proposal (agree, disagree, opinions, improvements and suggestions, ...) and CW in general?

ps: Another minor issue about CW is its (lack of) effect on badges, it seems strange that a user does not earn reputation but does earn badges for CW posts, e.g. I should not have earned a badges for these answers. Right now many of higher level badges are awarded for CW questions/answers, which does not seem a good thing to me.

Official FAQ for Theoretical Computer Science

Material to supplement the FAQ

MO policy

SE policy

Old discussions on cstheory meta:

$\endgroup$
8
  • $\begingroup$ Something I just noticed, not sure about it but it seems that a CW answer is mainly owned by the user who has written the largest portion of the text in it, and that user probably get a badge also. $\endgroup$
    – Kaveh
    Dec 6, 2010 at 2:18
  • $\begingroup$ @kaveh, I think that is correct. $\endgroup$
    – Suresh Venkat
    Dec 6, 2010 at 4:12
  • $\begingroup$ @kaveh: If I had known this, I would have offered my bounty to some other answer. Sigh :( $\endgroup$
    – Sadeq Dousti
    Dec 6, 2010 at 13:27
  • 1
    $\begingroup$ ps: This answer on SO meta explains when a question becomes CW by system automatically: 1. 5 users edit the post, 2. OP edits the post 10 times, 3. answers posted to CW question, 4. the question and all answers turn into CW when there are more than 30 answers. In addition: OP can make their own answer CW; a moderator can make a question CW. $\endgroup$
    – Kaveh
    Dec 6, 2010 at 22:22
  • $\begingroup$ I have a post that I've edited 11 times (cstheory.stackexchange.com/posts/2379/revisions), and it's not CW yet. I expect this feature may have been changed when the CW policy was changed. $\endgroup$
    – Peter Shor
    Dec 16, 2010 at 2:57
  • $\begingroup$ @Peter: it might be the case, though I haven't seen anything about it on stackoverflow meta. $\endgroup$
    – Kaveh
    Dec 16, 2010 at 5:27
  • 1
    $\begingroup$ @Kaveh: I couldn't resist testing to see whether it would be CW if I edited it once more (making it 10 real edits to the body). And the question did need editing, so as to put the results of the bounty and a comment on my answer near the top. The result: 10 real edits to the body of the text by the OP make a question community wiki. Changing the tags and the title don't seem to count. $\endgroup$
    – Peter Shor
    Jan 6, 2011 at 14:43
  • $\begingroup$ @Peter: thanks for the update, I think it makes sense not to count title and tag edits. $\endgroup$
    – Kaveh
    Jan 6, 2011 at 22:59

4 Answers 4

Reset to default
5
$\begingroup$

We have provided some additional guidance at the blog:

http://blog.stackoverflow.com/2011/08/the-future-of-community-wiki/

TL;DR version

Most of the time, you should be asking yourself “How can I improve this post so that community wiki isn’t needed?” Community wiki is like a cheese knife: it is a specialized tool to be used sparingly, and only in very specific circumstances.

$\endgroup$
4
$\begingroup$

I'd suggest that we follow an extremely simple CW policy:

  • "Big-list" or poll -> CW.
  • Anything else -> no need to CW.

If you think a (non-CW) question or answer is worthless, don't upvote.

$\endgroup$
1
  • $\begingroup$ If it's a "big-list", does that mean that only the question needs to be marked CW, or do all the answers need to be marked CW too? P.S. Surely polls are not allowed. (I realize this is old, but it's the closest I can find to an expression of current policy.) $\endgroup$
    – D.W.
    Oct 8, 2014 at 6:41
3
$\begingroup$

Good point, Kaveh!

I agree with most of your points. Specially, I liked the way you distinguished between the main and side effects.

However, I beg to differ in one case: There are many good CW questions, specially the big-list ones. I learned a lot from those posts. Unfortunately, such posts do not generate any reputation for the asker/answerer. So, what is the motivation to ask/answer them?

One answer is to offer badges for such cases. IMHO, this is the least thing that the system can do! As a matter of fact, if I were to decide, I'd offer more incentives for high-voted CW questions/answers.

$\endgroup$
2
  • $\begingroup$ you can earn badges from CW questions/anwers. you just don't earn rep. $\endgroup$
    – Suresh Venkat
    Dec 6, 2010 at 16:30
  • $\begingroup$ @Suresh: I know that, but I think I put it badly. I meant the system is doing the least to motivate asking/answering CWs. So, I'm objecting Kaveh's point about not offering badges to CWs. $\endgroup$
    – Sadeq Dousti
    Dec 7, 2010 at 2:08
2
$\begingroup$

This may sound radical, but do we ever need CW, especially for questions? Things will become much easier if we just pretend that questions cannot be made CW, and I suspect that we will not lose much by doing so.

Actually I have never understood when to use community wiki (CW). As you wrote, making a question CW has several effects, and I cannot see when we want to (1) give up reputation points, (2) make it easy to edit and (3) propagate the same effect to answers at the same time. For example, I have never understood why big-list questions should be CW. (I admit that I had flagged several big-list questions for moderator attention to make them CW, following the existing protocol without thinking why.)

Added: As you pointed out in a comment, it is sometimes explained that big-list questions should be CW because votes on answers reflect how good the items mentioned in the answers are rather than how good the answers themselves are. However, honestly speaking, I doubt that this explanation ultimately makes sense.

Many useful answers (no matter whether the question is a big-list question or not) are simply links to some external items, and in my opinion the answers pointing to good items in a useful way are good. (One of the examples which I always think of is Serge Gaspers’s answer to my first question on CS Theory Stack Exchange. The essential part of the answer is a single reference, and it was extremely useful for me.)

I do not like to talk about reputation points as if they are anything important, but if reputation points are worth talking about, then someone who pointed out a good item in a useful way should gain reputation points in my opinion.

$\endgroup$
10
  • $\begingroup$ I personally think that for questions which ask for a ranking of items like books CW makes sense to some extent, because in these cases we are not voting for the answers but are voting to rank the items mentioned in them. $\endgroup$
    – Kaveh
    Dec 6, 2010 at 2:06
  • $\begingroup$ @Kaveh: I have read that explanation somewhere, but honestly I do not think that it really makes sense. I added to the answer why. $\endgroup$
    – Tsuyoshi Ito
    Dec 6, 2010 at 2:29
  • $\begingroup$ Regarding the addition: first, I agree that reputation is of secondary value to our site, the main objective is have a useful site with quality questions and answers, but it seems to be a reasonable way of maintaining the quality of the answers and questions. About ranking questions, I think ranking questions like those about books can be different (your question is not a ranking question), but I agree, even there an answer can be of high quality (and that was the reason I had "to some extent" in my comment). $\endgroup$
    – Kaveh
    Dec 6, 2010 at 2:34
  • 3
    $\begingroup$ Many of the highly voted questions on this site are CW-type questions of the form 'what is the best X, or what are good references for Y'. Such questions can be asked really by anyone, and to me it seems like pointless reputation grubbing to receive value for questions like this. It also encourages people to ask such questions as a way of gaining reputation. A question should be CW when the question itself does not reflect any signficant quality: this is true for many (if not all) of the big-list questions. $\endgroup$
    – Suresh Venkat
    Dec 6, 2010 at 4:15
  • 1
    $\begingroup$ @Suresh: I do not buy that reasoning. First, people upvote the question because they like it. Asking such questions deserves reputation points. (We cannot deny that the reputation system on SE is populism.) Second, I can imagine that some people upvote CW questions more easily than non-CW questions because they know that voting on CW questions will not affect the reputation points. I suspect that if we abandon the convention of making big-list questions CW, the big-list questions which anyone could ask will not be highly voted. (more) $\endgroup$
    – Tsuyoshi Ito
    Dec 6, 2010 at 4:27
  • $\begingroup$ It's possible for sure. And definitely CW is one of those things that creates endless discussion on all SE platforms, even MO. I'm just a little nervous with a blanket "no CW' policy. $\endgroup$
    – Suresh Venkat
    Dec 6, 2010 at 4:32
  • $\begingroup$ (cont’d) Third, even if I am wrong and big-list questions which anyone could ask are highly voted even after abandoning the current convention, is it that bad? There are already many ways to gain “undeserved” reputation points if one is familiar with the dynamics, but it does not currently harm the website significantly. CW as a tool to prevent “undeserved” reputation points seems to be solving a wrong problem in the first place. $\endgroup$
    – Tsuyoshi Ito
    Dec 6, 2010 at 4:33
  • $\begingroup$ @Suresh: I admit that “no CW questions” is a radical change, and I will understand if it cannot be employed. Still, I hope that this sheds some new light to discussions on CW. $\endgroup$
    – Tsuyoshi Ito
    Dec 6, 2010 at 4:35
  • $\begingroup$ I totally agree with Tsuyoshi on the CW matter (however, i am not sure this is a relevant comment). It is clearly very easy to earn reputation in the site by flooding it with questions/answers, so why not giving rep to someone that asks for the first time a ranking/soft/generic question? $\endgroup$
    – Sylvain Peyronnet
    Dec 6, 2010 at 21:45
  • $\begingroup$ @Sylvain Peyronnet: I think not giving reputation makes it easier to have a real ranking list question because other users can vote answers without concerning themselves about giving reputation to someone. I also agree with Suresh's point. The reputation system is far from perfect and there are ways to game the system, but it does not mean we should make it easier to game the system. (As a side note, there are cases where a post becomes a CW automatically, for example when the question has more than 30 answers, and in that case I think all answers become CW automatically.) $\endgroup$
    – Kaveh
    Dec 6, 2010 at 22:13

You must log in to answer this question.

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