5
$\begingroup$

A while back in this meta question Lembik suggested in his answer a Sandbox for new puzzle-challenge-questions.

I think this is an idea worth discussing and potentially voting on in a specific thread for this.


Edit: I've overlooked that this posting is asking the same, sorry. Still, let's re-discuss with new effort and come to a conclusion here.


If you have something to propose and potentially "vote" on, please use answer not comment in this thread. Voting on the main post indicates your general opinion of having such a feature (in any form.) Use comments for clarification of negative votes please.

The three associated questions

  • Is it technically feasible and what features would it have or could it have?

  • Do we want such a feature why/why not?

  • In which way should the feature be handled? (Proposals please)

$\endgroup$
  • $\begingroup$ Worth reading, from PPCG: How does the Sandbox work? How do I use it? $\endgroup$ – Aza Dec 4 '14 at 8:19
  • $\begingroup$ There's an important difference between PPCG and PSE - code golf challenges require a lot more effort to write, and people are more likely to know about the Sandbox. Anyone can cobble together a bad riddle and post it without even knowing about the Sandbox. I'd expect almost every single user to do this at least once, unless there's some very very very very obvious instruction to them $\endgroup$ – Joe Dec 4 '14 at 9:51
  • $\begingroup$ Related (possibly as duplicate): meta.puzzling.stackexchange.com/questions/1579/… $\endgroup$ – Victor Stafusa Dec 4 '14 at 11:08
  • $\begingroup$ @Victor I cross-linked in both postings (edited here) and will delete the comments here now. $\endgroup$ – BmyGuest Dec 4 '14 at 11:21
  • 1
    $\begingroup$ @Joe but it could improve the quality of puzzles that the authors want to improve. If it doesn't improve the quality of all puzzles but does improve the quality of some puzzles then overall that's a net improvement. IMO we're going to need to do a number of different things, this could be one of them. $\endgroup$ – A E Dec 4 '14 at 17:49
5
$\begingroup$

I would say no.

This is a good idea. We do need some way to revise puzzles before they are released as the quality of puzzles has deteriorated in the past weeks. The "First Posts" review queue is not enough, as many new users post multiple riddles with negative feedback without changing their style.

There is are some problems though. Problems that are important enough to make the sandbox nearly useless.

How can we give feedback without knowing the answer?

The sandbox should be a place for the community to suggest edits for a question and make it better before it is actually asked, but without the answer, we cannot give feedback reliably. The reason this works over on PPCG is because there is no right answer. The OP there does not have to provide the correct answer, simply because the question doesn't depend on the answer. He might post a basic, ungolfed solution alongside his question just to show the answerers what to do, but the community doesn't need the best solution just to give feedback.

I don't see how this would work with Puzzling, however. I have seen many puzzles that seem unsolvable at first glance, but when I read the solution, it was just a well-crafted puzzle. I think the majority of the time, users would not be able to help improve the question effectively if they don't know the difficulty of the question - after all, if the puzzle is very well-crafted, then the solution may not be easy to find.

How can we ensure that the people who view the question in the sandbox don't have an advantage over others?

This results from the fact that there should be a single, distinct solution to each and every puzzle, assuming it is a challenge question. Using the PPCG example once again, a user may think they have a good solution. Once the question is released, they can post it whenever they want. Unless they happen to find the perfect solution somehow, there will probably be a better solution, whether it is one that is more creative, or one that is shorter.

Here on Puzzling, once someone finds the correct solution in the sandbox, they will try to post it as soon as they can when the question is officially asked. For some users who don't visit Puzzling Meta, they will not have those extra few days to work on the puzzle.


Again, I do feel like we need a way to revise most of the puzzles that come into the site. I just don't think a sandbox is the way to do it, unless we have a way to prevent these problems I mentioned.

$\endgroup$
  • $\begingroup$ +1 Totally agree, esp. with the last line. I was thinking of maybe a sandbox that only mods can access. The user will reveal the answer to them and they will be expected not to publicly reveal it. $\endgroup$ – ghosts_in_the_code Dec 27 '15 at 15:13
0
$\begingroup$

I think a Sandbox would be a great idea which could improve site-quality a lot without giving a "moderators kill all creativity" impression.

I do not know about the technical restrictions/possibilities so I'm just outlining how I think it could work:

What is Sandboxed and for how long?

  • There is a Question-about-Puzzles tag which by-passes the Sandbox and gets posted immediately.

Any question tagged like this which is not a question about puzzles but a challenge of some sort gets immediately closed. This feature would preserve the function of this site as a simple Q&A about puzzle-making. (Also, people who hate challenge-questions can out-filter, but that's secondary.) In general I think such "meta-tags" are not a good idea, but in this special case I don't see another way to do it. Do you?

  • All other questions get automatically sandboxed for 48 (?) hours. This countdown is renewed with each edit of the question.

At least 48 hrs to allow at least one full edit cycle for the whole world.

  • After the countdown runs out, the question is put onto the main-site.
  • During sandboxing people (with some reputation limit) can "open-vote" the question to release it immediately.

This rule should allow really good puzzles propagate faster. The automatism should prevent too long queues. It also gives the OP the "final say". If he does not want to edit, he does not have to. It's his own "risk" if his question then gets down-voted to hell for poor quality...

Now the important bit:

What happens during and after Sandboxing ?

  • While a question is sandboxed it can not receive answers.
  • A sandboxed question can receive just comments, votes and edits to both.

The whole purpose of the process is, that there is a "feedback" round which allows to make the puzzle fit the site format as good as possible. After Sandboxing, the puzzle should be 'fit' for the site as best as possible.

Note: It is not always possible to judge the quality of puzzle without the solution. Some puzzle-challenges will not really get changed (Hence also the automatism for releasing without edits.) It's really up to the OP to decided if the puzzle just can't use edits.

The sandboxing is giving the OP a chance to receive general, a-priori feedback.

It will hopefully weed out the really obvious bad questions, and it will give serious puzzle-smiths a chance to make the puzzle shine.

  • When a puzzle is released from sandboxing, it's vote-count is automatically set to zero.

I'm not 100% sure of this, but I think it's a good thing to start with a clean slate. Not all negative-voters during sandboxing might find the time to undo their vote if the OP has edited problematic things.

$\endgroup$
  • $\begingroup$ All other questions get automatically sandboxed for 48 (?) hours. This countdown is renewed with each edit of the question. --- So, if somebody would like to attack a question, (s)he would just edit it before the time ends. If this ends in some sort of edit-war, this is a point to the attacker. Further, people might innocently reset the counter when genuinely trying to improve the question. Some people would not edit it, and rather choose to keep any problems, to just not reset the counter. So I think that this do not works. $\endgroup$ – Victor Stafusa Dec 4 '14 at 9:27
  • $\begingroup$ @Victor Good point. I was thinking of OP edits only and that only the OP can edit in Sandbox. However, I see that others editing could be beneficial. Do you have an idea? $\endgroup$ – BmyGuest Dec 4 '14 at 9:42
  • $\begingroup$ Should a sandboxed challenge include a solution? $\endgroup$ – xnor Dec 4 '14 at 9:55
  • $\begingroup$ I don't think so. Otherwise we have "experienced" users use Sandbox as "main" site and just transfer all problems there... that's not what we want. (I think.) In some cases some comments in Sandbox may ask the OP to provide a solution for verification, but that shouldn't be the norm, I'd guess. $\endgroup$ – BmyGuest Dec 4 '14 at 9:57
  • 2
    $\begingroup$ One downside to this... What if five people solve the challenge before they can answer Then, the one that gets credit is just the one that watches the clock best. That's no fun. $\endgroup$ – Travis Don Kindred Dec 4 '14 at 17:22
  • $\begingroup$ It sounds like what you're proposing requires software support; large changes to the basic functionality of the engine. Such changes are extraordinarily unlikely. It might be possible to cobble this system together with a very strong policy and (mis)use of existing features, but that would require constant moderator vigilance. Starting every question off in a closed state isn't possible on an SE site. $\endgroup$ – jscs Dec 4 '14 at 23:03
  • $\begingroup$ PPCG uses a dummy question on its Meta to house its sandboxed posts, which are added as answers. Critics then use comments to talk about the proposals. $\endgroup$ – jscs Dec 4 '14 at 23:34

You must log in to answer this question.

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