18

In the SO Review page https://stackoverflow.com/review, I see the following stats for questions awaiting review:

  • Close votes: 51.8k (and rising)
  • Suggested edits: 10 (fluctuates, but rarely goes above 20)
  • Everything else: 0 (occasionally see some if you're really lucky)

My question is: Why?

Why are there so many close votes waiting for review? Close votes should expire after four days. And even if there's a good reason for them hanging on longer (ie they're not being viewed), there surely has to be a cut-off eventually after which they can be safely dropped?

And yes, I saw this question, which is asking basically the same thing. But that was several months ago, and it seems that the queue has hardly moved since then. It seems to me that there's a massive backlog that will never be cleared, and apparently will never expire either. That's not good.

And on the flip side, why are there hardly ever any new user posts or low quality posts to review -- there are always new users posting things, and I recall before this new review system was launched that the list of low quality posts to review was always pretty big.

7
  • Reviewing is easy, closing is hard. Apr 25, 2013 at 11:59
  • @0A0D - I think that's an over-simplification, but there's definitely some kind of imbalance here.
    – Spudley
    Apr 25, 2013 at 12:11
  • My review page only says 50.5k close votes - hence my estimate of zeroing it in 2018. If it really is 51.8k now, that estimate would have to be revised :-)
    – Monolo
    Apr 25, 2013 at 12:24
  • @Monolo: Mine says 51.8k Apr 25, 2013 at 12:31
  • The other problem is that the close vote will still exist even if you chose the option "Leave It Open". Apr 25, 2013 at 12:32
  • @0A0D I guess it subtracts the reviews (including skips) that one has done, although it doesn't take the user's filters into account. That would explain the difference, and be in line with the other queues, now that I think about it. At any rate, too many.
    – Monolo
    Apr 25, 2013 at 13:16
  • Reviewing takes 3000 rep. The others take almost no rep. I help with the other queues whenever I open the page and it has something to do (which is a few times a day, maybe 10% of the time I check.) I'd help close, too, if I could, but I can't, because I don't have 3000 rep. So, to edit what 0A0D said: my experience, reviewing is easy... answering questions before someone else already answered it better, that's what's hard. I worked hard for my ~900 rep, but it's been pretty slow.
    – neminem
    Nov 25, 2013 at 18:21

2 Answers 2

11

The reason that close votes over 4 days old are still in the queue is explained nicely by Martijn Pieters in this answer to the question you linked. To summarise his answer:

  • Close votes only expire if the question has had 100 views or more
  • Users without the the ability to cast close votes must use flags instead. Although these don't appear as close votes when reviewing, they still mean that it gets entered into the queue.

The backlog for close votes is gradually decreasing over time and will eventually be cleared many years from now 8)

To address your query on why so few first posts are in that particular queue, there have been over 1500 of them reviewed so far today. There have been over 550 low quality posts reviewed so far today also.

Reviews are just handled very quickly by the active community as there are so many people participating. Anyone with over 125 reputation can review first posts and late answers.

4
  • Can you rephrase: "without the cast close votes privilege flag instead".. my brain is having trouble comprehending that statement Apr 25, 2013 at 11:58
  • @0A0D is that better? :)
    – Ren
    Apr 25, 2013 at 12:00
  • 1
    Now it is...... :) Apr 25, 2013 at 12:02
  • 2
    The backlog for close votes is gradually decreasing over time and will eventually be cleared many years from now 8) It's over 90k now. It appears to be growing to me
    – Cruncher
    Nov 18, 2013 at 16:50
5

If you look at this question you'll notice that it was at 55k. This was approximately 6 months ago.

As noted by the OP, it is now at 51.8k, meaning that we can (coarsely, but generously) estimate that the close vote review queue diminishes at a net rate of about 7k questions a year. By applying a somewhat simplistic calculation, at the current rate it will get to zero in about seven years, sometime in 2020. Technically, that is not never, although in Internet terms it is a geological age.

We need a virtual earthquake that can eat all the close votes in the queue.

Alternatively, the rules could be amended so questions with no recent activity, few views and a certain number of close votes would be closed automatically. The exact thresholds TBD, of course.

Some people believe that there are gold nuggets hidden in that mountain of old, half forgotten questions, and they are probably right.

My counter-theory is that at the rate of new questions coming in, new gold nuggets will quickly be created to take the place of any that might have been thrown away in an automatic sweep, and I don't think we should be overly worried.

They are only questions, after all, so "Kill them all, God will know His own".

If a question is truly important then someone, somewhere will ask it again. Heck, they do it even if the question is not important; the Stack Overflow home page is my witness.

4
  • 7
    The queue isn't the problem. We could shrink that number dramatically just by ignoring close votes that aren't recent (and might end up doing this just because folks get so hung up on it). The point of reviewing them manually is that for every terrible question in that queue there's another that isn't really a problem but managed to attract a vote or flag somewhere along the line just the same. Don't hesitate to use Skip and Leave Open when appropriate - reviewing goes much, much faster when you focus on stuff you know and don't stress about closing stuff you don't feel strongly has a problem.
    – Shog9
    Apr 25, 2013 at 15:13
  • 1
    @Shog9 Don't worry, skipping and filtering is what I do. What I do object against is the expectation that the members of this community are expected to go through 50k questions up to 5 times. I am doing my part also so I have a high horse to complain from :-), but I would never expect it from other community members. It is simply too pointless compared to the effort. That, and the religious belief that if just a single person out there gets some help from a post, then it is useful and needs nurturing. The cost/benefit of that doesn't hold up.
    – Monolo
    Apr 25, 2013 at 15:53
  • @Shog9 And apart from my rant above, somehow cutting that review queue down to a manageable size sounds like a good idea.
    – Monolo
    Apr 25, 2013 at 15:57
  • 2
    it only takes three reviewers if a post doesn't need to be closed. Another option is to break up the queue into more specialized chunks - for instance, duplicates require considerably more effort to review (in my experience), while flags from low-rep users often have a lousy accuracy rate (so perhaps easier to dismiss when reviewed by high-rep users).
    – Shog9
    Apr 25, 2013 at 16:16

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