6

So to support the goal of waiting a bit to close questions, and under the assumption that once questions end up in the review queue they get closed quickly, I had made this feature request on MSE.

As it turns out, it's already an option! Shog writes:

This already exists - in fact, it's been implemented for somewhere around four years now.

The default delay is... 15 minutes. No review tasks for the Close or Reopen queues will be created until the oldest pending close vote or flag or reopen vote or qualifying edit is at least 15 minutes old. No review tasks for the Late Answer or Low Quality review queues will be created until the post itself is at least 15 minutes old.

The rationale for this delay (for those queues only!) is pretty much what you describe: give folks a chance to provide feedback organically and give the author a chance to fix things before we drag a bunch of other people in.

It's possible to increase this delay, both globally and on a per-site basis... While I see little value in increasing this value across the board, if folks on a given site believe their topic or audience warrants a longer delay, then this is trivial to increase - just propose the change on the proper meta site(s) and let the people who'll be affected by it discuss the pros and cons of various values. If there's consensus, I'll make the change.

This means, if we can reach a consensus on it, we can have the amount of time between the first close vote and its appearance in the review queue delayed.

The feature exists already exactly for the reasons I proposed: To give people a chance to clear up unclear questions and such before the question is put on hold, while at the same time letting the community delay close votes without the need for e.g. bookmarking review lists offline or periodically performing mass cleanups. To me this seems like exactly what we want here.

However: There is a caveat here: This delay also applies to the reopen queue, both for reopen votes and for edited questions. This is the flip side of this coin. Striking a balance may require diligence in other areas, which is probably a conversation for another post. Let's see where this discussion leads us, though.

So then, what do you guys think about a longer delay? Perhaps a day or so? The purpose of this discussion is to determine if we want a different delay and, if so, how long. We could probably talk about how we can effectively deal with a longer reopen delay here too, but there's a lot of approaches to that and it may be difficult to discuss in comments.

| |
3

Anecdotally, I'd say a week is more reasonable.

It should give people enough time to add missing information.

I support this change because it would offer a far better alternative to our current modus operandi for cleaning up poor questions.

| |
  • 3
    The caveat here is that this applies to all four of the queues mentioned, which may not be desirable for things like reopen. The other caveat is that close votes and flags age away if corroborating votes are not forthcoming, often in less than a week... As a practical matter, I would consider 48-hours the longest you'd want to wait, with 24 or less more reasonable. – Shog9 Sep 19 '16 at 22:16
  • Ah, we're going to have to find a balance, then. I think making a delay effective will rely on us using close votes more liberally, via the questions list, for the other close reasons, and also being diligent about reopens. This means we still need to make some changes to the underlying philosophy (which I'm all for, as everybody is already aware of, for better or for worse). It's too bad the reopen + close timings aren't separate. As for the timing... I'm going to see if I can gather some evidence about how long it usually takes to clear up a question in comments here. Will be hard though. – Jason C Sep 19 '16 at 22:20
  • @JasonC, you're setting out on if not totally, nearly impossible task. It will totally depend on the person and their work setting. It will also depend on your definition of cleared up. – anonymous2 Sep 20 '16 at 12:05
  • @Shog9 - is there any chance that the time delay on reopen could be split from the delay on close? Seems like they are such different aspects of the UX of the site that it would be a reasonable thing to do (or at least to allow). – dlu Sep 20 '16 at 18:00
  • I wouldn't push for it unless it happened that a long window was very useful, @dlu. – Shog9 Sep 20 '16 at 18:20
  • @dlu Given what Shog just said; maybe one thing we could do, as an experiment, is if we got a longer queue delay time, we could actively hunt for questions to reopen to make sure they were reopened quickly. Even though it's a lot of extra work and so defeats one of the ultimate goals here, we could treat it as a temporary short-term experiment to see if a long close delay + short reopen delay served us well. Then, if it does, we can use that to make a case to push for separate close / reopen delay times, and our experiment ends. – Jason C Sep 20 '16 at 18:22
  • @JasonC, maybe I misunderstood, but can we even see that a question has been nominated for reopening before the delay expires. I'm all for experiments and favor working with users on questions that need refinement. – dlu Sep 20 '16 at 18:26
  • @dlu You would be able to see it by looking at the question itself if you have >= 500 rep; underneath it will say "reopen (1)" or whatever. I could theoretically write a web app that maintained a list of questions that had a reopen vote for us to use during an experiment like this, but in general if we've got enough active users actually reading the questions, in an ideal world such a list wouldn't even be necessary (the relative small size of this site ideally means that most people view most questions on their own). – Jason C Sep 20 '16 at 18:32
  • @JasonC, cool, and then there is a way to "manually" reopen the question without having to go through the review process right? I need to read up on all this stuff... – dlu Sep 20 '16 at 18:34
  • @dlu Yeah it's just like closing a question. You never have to go through the queue. You click "close" or "reopen" or whatever under the question and cast your vote, and that's a legit, official, real vote. The queue is solely for drawing attention to questions. – Jason C Sep 20 '16 at 18:40
  • @dlu PS: My hypothesis (I have not yet found a way to query data to support this) is that the behavior on this site is that we're intentionally hesitant about casting the first vote, but the rest of them still seem to come in quickly once attention is drawn in the queue; hence why I think the delay would get us what we want (let us do this but also make it so we don't have to constantly clean up forgotten questions, and make it more agreeable to use the system and cast the first vote). – Jason C Sep 20 '16 at 18:43
  • 2
    If you have 2K rep here right now, you have access to a handy list of pending reopen votes... @dlu – Shog9 Sep 20 '16 at 21:05
2

I'd sure like to see the delays split between the close and reopen queues. Reopens should hit the review queue quite quickly to provide positive reinforcement for trying to improve a question and resolve the problems that got it closed (and also to make the site feel responsive to users).

On the other hand "normal" closes should be slow – at least here – since we seem to have a culture of working with users to refine their questions and I really like that. A week wouldn't be unreasonable, or as long as it could be without hitting bad interactions with other heuristics. There is no, or at least little, harm in a duplicate or unclear question and some of them pull in new and interesting thinking in the answers.

That leaves the "abnormal" closes. I think a question flagged as abusive should go almost instantly to the close queue. They lower the level of the site and I think the only issue is due process. The posters recourse should be to edit and propose the post for reopening.

How likely is it that we could get the timeframes split?

| |
  • Adding to this: One thing we'd have to count on here is our ability to quickly apply "abnormal" closes (as defined here) via the question list instead of the review queue. That is, count on users to spot these questions in the list and vote quickly there. Can we do this? Do we have the traffic for it? We have 92 users with enough rep to close who have been active in the past 7 days. – Jason C Sep 20 '16 at 18:29
  • @JasonC, will the rep needed change once we finish graduating? Maybe we want to look at those numbers. It does seem that we have some pretty diligent users who are hear on highly frequent (more than once per day) basis. – dlu Sep 20 '16 at 18:32
  • AFAIK the rep thresholds typically change once the site design is complete, staying at the old public beta levels until then. – Jason C Sep 20 '16 at 18:39
  • Yes, that was the impression I got from the graduation announcement. Assuming that we might have to count on diligence to keep the delay working it seemed like it might be worth modeling the behavior based on the post redesign rep thresholds – or do privileges get grandfathered? – dlu Sep 20 '16 at 18:43
  • Privileges are not grandfathered. I think if we do this we should rush to do it before we get a site design -- but I could see if there were an argument for waiting, too (fyi: We've got 27 users active in the past week who would retain close privileges if it went up to 3000 right now). Anyways I would be surprised if we get a design any less than 6-8 weeks from now, hehe. – Jason C Sep 20 '16 at 18:50
  • 1
    I can't think of any reason for waiting, and my hunch is that 27 users might be enough to keep it going. I like the tone of support here for car repair novices and I'm hoping that graduating might help us grow more. So, I'd really like to get our "culture" around closes well established. – dlu Sep 20 '16 at 18:57
  • 1
    The other way to look at this is that, once graduation goes through and privileges change, the entire problem might become a non-issue simply because there'd be that many fewer close votes to spare, @Jason. – Shog9 Sep 20 '16 at 21:06
  • @Shog9 That's true, but it would ideally be a temporary non-issue as the site recovers privileged users. Personally I think it's good to lay groundwork now, but I could see a reasonable argument to almost completely table it for now then see if it still exists later. For various reasons, though, I think it's wiser to think about it early, particularly because the users who will retain privileges are the current power users who drive the philosophies here now, and will continue to set examples in the future. But, honestly, I don't know what the best strategy is, if a strategy is even necessary. – Jason C Sep 20 '16 at 21:16
  • 2
    Arguably this is only even a problem because the site has reached adolescence (so to speak): for a long time, there were only a couple of people actually using it regularly, and they handled most of the moderation at their leisure... Now there are more people using it, more stuff that needs to be moderated, and more people able to moderate... and they haven't quite all gotten used to that yet. Eventually, yes: there will be even more people able to moderate... But by that time, there will likely also be quite a bit more work to do. @Jason. – Shog9 Sep 20 '16 at 21:31
  • @Shog9 - Well said and how true it is. – Pᴀᴜʟsᴛᴇʀ2 Sep 22 '16 at 20:57

You must log in to answer this question.

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