10

For nearly 5-6 hours now, I am not getting the usual sorted list of questions on SO's interesting questions page. Here are two screenshot(with awesome red-circles)(which are becoming more common in my questions :():

Image 1

Image 2

As it can be noticed, all the timestamps are placed randomly. Is it newest feature/change in SO? If so, please revert it. :/ Or is it one of the two options mentioned in the question title?

  • 3
    Not sure what the problem is? You're in the "interesting" tab, the questions aren't supposed to be sorted chronologically. – yannis Apr 2 '13 at 2:10
  • @Yannis AFAIR, they always appeared in the most recent to oldest order. – hjpotter92 Apr 2 '13 at 2:12
  • Or rather, in the most recent to oldest changes to them – hjpotter92 Apr 2 '13 at 2:12
  • You're right, I guess I never looked at that tab carefully. – yannis Apr 2 '13 at 2:23
  • In the latest podcast they discussed mixing in more popular questions from the last few days, so that newer visitors get a taste of something other than new-and-unanswered questions. – Jeremy Banks Apr 2 '13 at 11:19
13

This is intentional at the moment, we're testing some improvements to interesting.

  • I'll be staying awake then. :) – hjpotter92 Apr 2 '13 at 2:40
  • 5
    Damn, guys, April Fools is over already >:( – Omega Apr 2 '13 at 3:16
  • -1, I hope those so called improvements won't make it or I will be a very angry user: meta.stackexchange.com/questions/175543/… – Darin Dimitrov Apr 7 '13 at 9:33
  • @Darin you don't even know what's being tested here, obvious by the question you asked. Being very upset at a few questions (only on the home page at that) is a premature over-reaction at this point, we're collecting data...that's the point of any experiment. You're making the assumption the data is even for the home page... – Nick Craver Apr 7 '13 at 12:10
  • 2
    @NickCraver, I am not making any assumptions, all I do is see a change to the behavior of the main page on StackOverflow that I am protesting against because it seriously impact the usability of the site from my standpoint. In your answer you mentioned that this is intentional because you seem to be testing some improvements, all I did in my comment was to explicitly state that I hope that those improvements at their current state are something that I absolutely do not wish seeing. The sooner those changes are reverted back, the better. – Darin Dimitrov Apr 7 '13 at 12:34
  • @Darin It's not the main page of Stack Overflow, it's the home page (the main page is the question page, by a very wide margin). We're testing improvements which are constantly being tweaked and changed as we go and run successive experiments, you don't even know which set you're seeing. We're not taking only vocal feedback off meta into account when saying whether or not a change to interesting (and other things this is really for) makes it...we're collecting actual numbers for hard data, that's the point. – Nick Craver Apr 7 '13 at 12:57
  • 3
    Alright, home page, main page, whatever, you know what I am talking about. Let me put it that way (hopefully it will avoid any confusion): if the thing you are testing (call it feature or whatever) actually gets implemented I will simply request my account being deleted from StackOverflow and stop contributing because apparently you don't take into account user requests. You are experimenting with things without first asking the users. – Darin Dimitrov Apr 7 '13 at 13:00
  • 1
    @Darin - thanks for clarifying, now it's clearly apparent you're not overreacting at all. The testing is completely random, we're not asking for a reason...that defeats the entire point of an experiment, especially one that may affect anonymous users. I said we were not taking only vocal feedback into account, nowhere did I say we not taking it into account, we certainly are and always have. Quit trying to twist my words into something I didn't write, you've been around long enough to know that's not true. – Nick Craver Apr 7 '13 at 13:20
  • 2
    Alright, if you are taking user feedback into account, I expect that my original question either gets closed as duplicate to this one or an answer gets posted explaining precisely what is being tested, how is it being tested, why is it being tested, what are the possible outcomes of those tests and most importantly when to expect them the results of those tests and under what form. I would also like an explanation of on what basis are the affected users chosen. – Darin Dimitrov Apr 7 '13 at 16:22
  • @Darin are you telling me you don't know what the word random means? I find that hard to believe. Also, I'm not running these tests...I have other pressing infrastructure changes on my plate. The guys actually doing the research here will likely give more info next week. – Nick Craver Apr 7 '13 at 19:13
  • @NickCraver, I understand what the random word means. But unfortunately this answers only the last question of my previous comment. It doesn't answer the parts in bold. I am of course not expecting an answer from you personally if you are not the one involved in the experiment. What I expect is just an official statement from someone on the SO team that would answer my questions. It is important to note also that while you are conducting those experiments I find the usability of the site greatly reduced and really hope that my request will be heard. – Darin Dimitrov Apr 7 '13 at 20:49
  • @Nick I see it's live now for all users on Stack Overflow in a "weaker" form of just taking over some questions in the top 10. Is there some formal announcement? Also, it caused the homepage to load in over two seconds where it used to load in less than a second before, about 500% slower - hope you'll improve this over time. – ShaWiz Aug 18 '13 at 13:22
  • @ShaWizDowArd it's actually a different experiment running (of which you may be a subject) - 2 seconds is absolutely unacceptable though, I'll have Ben C take a look tomorrow. – Nick Craver Aug 18 '13 at 13:38
  • 8
    I don't think this is an improvement. I have the same question at the top for most of the day. I keep clicking on it although there is no activity on it. Quite annoying. – Mischa Aug 18 '13 at 14:08
  • 1
    @Nick any updates? Still getting slow response 100% of the times and I'm 100% sure it was much faster before the change. – ShaWiz Aug 19 '13 at 11:31
5

Users should not be seeing these delays anymore. Like Nick said, we were running an experiment. The thing we're testing obviously needs a performance improvement (this is a known issue). What went wrong here is that the experiment was only supposed to be run on a small percentage of the users, and it looks like every user was affected in the latest iteration of the experiment. I've fixed it so this shouldn't happen like this again, and I'll be working on some performance tuning before hammering anyone with it again. The reason we use the homepage for these experiments is that it's the only place we can collect meaningful data. If we added an 'experimental' tab, not enough people would use it, and those that did would be more self-selecting. Same with the other tabs: we could tweak "hot", but our changes would make even less sense there, and we'd still have the same data collection problems. That leaves the homepage, and we try very hard to do these experiments in such a way that exceedingly few people would notice anything different (indeed, nobody yet has pointed out the more obvious signs of the experiment).

Normally we like to get user feedback when we try new features or make substantial changes to the site. In this case, the problem with feedback is that the data we're collecting from the homepage isn't really for the homepage. It's for mobile, and not only that, but the tweaks you've been seeing on the homepage aren't what mobile users are going to see either. If we solicit feedback, we're going to get a lot of "we hate it! it's too slow! it doesn't make sense!", and our response will be "we know! it's not going to stay this way!".

It really was just an experiment: "do we get better user activity when we present question set A over set B?" The data validate or invalidate our experiment, and then we can use that algorithm to help select content to show mobile users. One of the reasons you're seeing questions hang around longer is that we want the feed of questions we provide to mobile users to be stable, and not a constant churn like 'new' or 'hot', or even like the homepage itself. We are getting feedback and bug reports from our mobile testers, and as mobile gets more mature you'll see more and more of the normal type of Meta interaction with our users that you're used to.

  • 5
    What was the experiment supposed to be? I don't see how showing the same questions for hours at a time on the "interesting" page is ever going to be a useful idea. – Martin Smith Aug 19 '13 at 15:21
  • 1
    It'd be nice to be able to opt out of being a lab rat. – j08691 Aug 19 '13 at 15:21
  • 4
    The home page is one of the few useful ways for me to find new questions on a site as large as Stack Overflow. There are far too many questions flooding in at any given moment, on topics that I do not care about, to use the "Newest" or "Unanswered" tabs. Please do not break the home page's current functionality by displaying old, stale questions. I want to be able to refresh it after a while and get a list of new, interesting questions to read/answer. The interesting algorithm works superbly well, but it will all be for naught if the time stamps aren't also factored in. – Cody Gray Aug 19 '13 at 15:27
  • @CodyGray: Is something from only 5 hours ago old and stale? – Ben Collins Aug 19 '13 at 15:35
  • 8
    Well, potentially. If I already looked at it 5 hours ago. Or 2 hours ago. Then it would be stale to me. I suppose it is perfectly fine if you're tracking and excluding questions that I have already viewed/interacted with, but this has not been my experience. – Cody Gray Aug 19 '13 at 15:38
  • 2
    I agree with Cody. Like I also expressed below: you expect new activity on questions at the top, but with this change a question seems to stay on top for hours even though there is no activity on it. You click on it on it anyway expecting new answers or an edit by the asker, but there is nothing. If it stays on top for hours this doesn't just happens once, but multiple times. It's just annoying. If an "old" question would go away after looking at it once I wouldn't mind, but this implementation is not good IMO for the reason above. – Mischa Aug 19 '13 at 16:28
  • 1
    @Ben why not open the discussion about the change in homepage algorithm to everyone and get feedback, like you do with many new features of such a big impact? I fear that having it your way without even trying to get feedback will drive many good users away when implemented. – ShaWiz Aug 20 '13 at 6:23
  • @ShaWizDowArd: moved my comment to my answer. – Ben Collins Aug 20 '13 at 14:11
  • Thanks Ben. By "mobile users" do you mean those using the android app, or anyone browsing in mobile theme? – ShaWiz Aug 20 '13 at 14:18
  • @ShaWizDowArd: Android app (and when we get there, iOS). – Ben Collins Aug 20 '13 at 14:23
  • @j08691: There are no rats. Only unicorns. – Ben Collins Aug 20 '13 at 14:24

You must log in to answer this question.

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