11

Like 634 others, I follow @StackProgrammer on Twitter to read and retweet hot questions. However, it's been nagging at me recently that I perceived "a lot" of these tweets -- tweets ostensibly meant to focus attention on those questions and bring both veteran and new users into the site -- were leading to questions already closed or deleted by the time I read them (and my tweetstream is always on).

I looked at the last 100 tweets of hot questions, which only takes us back to Feb 21st, and 14 were closed and 1 was deleted.

I am not questioning whether or not the questions should've been closed, or even the speed with which they are closed (I agree with the community in all of the instances), but:

  • Should we be concerned that 15% of our tweeted hot questions will lead users to closed or deleted questions?
  • If we should be concerned, and knowing that the moderators & community members are doing their jobs, is there any tuning that can be done to what I assume is an automated process at the SE level?

Bear in mind that I only looked at the last 100 because I only noticed it happening recently -- it could very well be that this isn't an issue at all and is just an off week. But if it continues to happen at a similar pace, I think from a community promotion standpoint it might be a problem we should consider trying to solve.

Edited to add links to most of the 14 closed questions, per @Rachel's request in a comment:

11
  • 2
    How many of those questions were closed by star mods?
    – Jim G.
    Commented Mar 5, 2012 at 2:42
  • @JimG. Of the 14 closures, 10 were by a star mod.
    – jcmeloni
    Commented Mar 5, 2012 at 3:04
  • 1
    Wow. @Rachel - Should we be concerned about this trend?
    – Jim G.
    Commented Mar 5, 2012 at 3:11
  • 1
    @JimG. Why you are asking me? lol I've already stated my concerns, and it's up to SE to do something about it if they agree, and so far they don't seem to agree. I still think we either need to rename the site or adjust the FAQ, and since I can't change the site name I plan on attempting to propose a revised FAQ, so we allow some programmer-specific questions without going back to the "what kind of pickles are good for programmers" type of questions. And providing I can find the time for it
    – Rachel
    Commented Mar 5, 2012 at 5:05
  • 1
    @jcmeloni Out of curiosity, would you be able to provide a list of which questions were tweeted and got closed? I'm just curious the sort of questions they were.
    – Rachel
    Commented Mar 5, 2012 at 5:06
  • 1
    @Rachel Done, although there's one that slipped past my list-making and I don't have it in me to go looking through all the links again to find it. I think you get the idea, though. :)
    – jcmeloni
    Commented Mar 5, 2012 at 13:14
  • Pro tip: Just add the url to the question, no need to add the title and link to it, the title is fetched automagically.
    – yannis Mod
    Commented Mar 5, 2012 at 16:02
  • @YannisRizos Thanks. It seems I always do things the long way before I've had my coffee.
    – jcmeloni
    Commented Mar 5, 2012 at 16:21
  • @Rachel I still don't understand: if you want a programmer-related site, why don't you make an Area51 proposal?
    – Dynamic
    Commented Mar 9, 2012 at 20:25
  • @Jae Because this site stole the domain name, and the last programmer-related site proposal got closed as a duplicate to P.SE. Change the site name to something like softwaredevelopment.se and I'll forever drop the topic of site scope on here.
    – Rachel
    Commented Mar 9, 2012 at 20:27
  • @Rachel What about aboutprogrammers or programmersarepeople or... :P. Ya I get it, but I still agree with the fact that those types of questions aren't fit for StackExchange.
    – Dynamic
    Commented Mar 9, 2012 at 20:32

2 Answers 2

14

I am concerned as it seems to affect all sites, not just ours. I don't like the fact that questions that aren't the best advertisement for the site being Tweeted or posted to Google+.

This has been raised by moderators several times in the past and we have be told that changes to the way questions are Tweeted are in the pipeline. I thought that there was supposed to be a change where questions with pending close votes weren't Tweeted - but obviously if a question is closed by a single moderator vote that rule won't apply. I will double check what the situation is and report back.

I'm fairly sure that closed questions aren't Tweeted, but that doesn't help if the question is closed after it has been publicised.

The other request that has been made is to give diamond moderators some control over what questions get Tweeted. At the very least this would give a pair of human eyes on what goes out for a sanity check.

I would like to point out that I rarely close questions without there being at least one moderator flag being raised. My job isn't to trawl the site looking for things to do, but to react to flags and questions here.

UPDATE

A search of Meta (not by me - my meta searches suck) has turned up this:

Avoid tweeting bad questions

It appears that the bot shouldn't be Tweeting closed questions. The only thing I can think of is that the question has had enough activity to be Tweeted (the rules are here), but the question has also picked up flags as well. We see the flags and close the question but it's out there in the Twitterverse.

The sequence will be something like (assuming it doesn't look at flags):

  1. Question gets posted.
  2. Question gets flagged.
  3. Question gets enough activity to be Tweeted.
  4. Question gets close votes.
  5. Question gets closed by a moderator.

So it appears that the bot is pimping closed questions. Therefore, it would useful if the bot could look at any existing flags and not Tweet it if there are any. Obviously it won't help if the question gets flagged after it's Tweeted, but questions that don't fit tend to get flagged quite quickly.

UPDATE 2

I've reported this problem on Meta Stack Overflow as its clearly Tweeting posts with flags.

4
  • I was wondering the extent to which it affected other sites, because the two other sites I spend the most time are low-traffic enough for me not to notice it if it happens. I also don't think that the tweets happen after a closure, but I can't tell (I'll keep an eye on it), and I +1 the idea of human eyes first -- that would be a good start if it were worked into the process.
    – jcmeloni
    Commented Mar 5, 2012 at 13:17
  • 2
    +1 for a great answer, and I really appreciate you pointing out that moderators should be exception handlers, not enforcers.
    – Rachel
    Commented Mar 5, 2012 at 16:35
  • 2
    @ChrisF Thanks for your continued attention!
    – jcmeloni
    Commented Mar 8, 2012 at 13:08
  • FWIW recently criteria for tweeting Programmers questions was tightened to address concerns like that: "raised the threshold score to 3 for this particular category". Questions with score less than 3 won't get tweeted
    – gnat
    Commented Jan 20, 2015 at 6:09
5

Yes it is a bit concerning, although 15% is not that much. I went through some of the questions and there is a minor pattern there, several of the closed questions that made it to Twitter were incorrectly migrated from Stack Overflow and they carried upvotes and upvoted answers from there.

The problem is in the "hot questions" algorithm and as ChrisF mentioned it affects all sites, not just Programmers. Some things we could do:

  1. Close questions quicker

    There seems to be a deep rooted perception that closing questions is a bad thing. It's not, and it should happen fast. Closing is just a peer review mechanism, a signal that a question is a bit problematic and in need of improvement.

  2. Re-open questions

    Don't dismiss closed questions, they are not dead. If you feel a question can be re-opened, please try to get it re-opened. Bring it up on chat or Meta, find others that can help. Not every closed question will be re-opened, obviously, but some will.

  3. Vote sensibly

    Upvote questions that show research effort, are clear and useful. Don't upvote questions that you feel are interesting / fun but blatantly off topic. Similarly with downvotes, they are a comment to the question, not the person asking the question.

Further reading:

Lastly, we prefer to be called diamond moderators ;)


As for the list of questions, there where flags on all questions, including the three questions that where closed by a single moderator vote:

  1. https://softwareengineering.stackexchange.com/questions/136610/version-number-soup-in-ror-is-grails-better-django

    Typical "which technology is better" question, it could have been closed as off topic. Closed by a user and a moderator.

  2. Online Code Editor

    Blatantly off topic question, but borderline on Stack Overflow. Closed by a moderator, and I trust that it was never migrated to SO because the SO mods rejected it.

  3. Is it a good idea to generate code with the help of your IDE?

    Closed as a duplicate by a moderator. This shouldn't really be an issue for someone coming to the site from Twitter.

  4. What do you do when you can't seem to understand a certain part of programming?

    Extremely vague and open ended question, gentle reminder to everyone: Programmers Stack Exchange is not a free for all discussion forum. Closed by a user and a moderator, and it makes me sad that it got as many upvotes as it did (question & answers).

  5. Is Object Oriented stuff really that important?

    Extremely open ended and argumentative question, comment wars on the question and almost every answer (and a particularly nasty "discussion", that has since been removed). Closed by a user and a moderator.

  6. why do some job posts say "high pressure environment"?

    Blatantly off topic, typical example of a populist question. Closed by five users.

  7. A Class named Class?

    I closed this one because it was getting a bit out of hand, and there wasn't any need for further answers. It had three re-open votes, and I've just re-opened it with a notice for longer answers. If it starts gathering poor answers again, I won't hesitate to close it.

  8. Offshoring a Software Project -- Conflict Resolution

    Very open ended. Closed by two users and a moderator. Potentially salvageable, but it requires a lot of work, and OP seems satisfied with the existing answers.

  9. https://softwareengineering.stackexchange.com/questions/137864/how-long-did-it-take-you-to-get-kernel-programming-down

    Open ended and quite difficult to understand what it's actually about. Closed by five users. I don't understand why this was tweeted, it's a zero scored question with no answers.

  10. https://softwareengineering.stackexchange.com/questions/137896/software-engineer-already-in-late-thirties-would-i-have-trouble-finding-a-new

    Duplicate, closed by four users and a moderator. Same as (3).

  11. Does Object Oriented Programming Really Model The Real World?

    I closed this one because it started gathering answers that re-iterated earlier answers. It's not a question that will be deleted soon (if ever). It deserves the community's attention (wink).

  12. https://softwareengineering.stackexchange.com/questions/138132/list-of-the-really-big-and-really-small-discrepancies-with-js-in-the-big-5-brows

    Vague, overly broad, argumentative and mostly off topic. Closed by five users, I wonder if the fact that the single answer was negatively voted should have prevented tweeting it.

  13. Python is slowly replacing C in universities. Does this move degrade the quality of CS students?

    Vague, speculative, argumentative, wrong. Close by five users, I suspect it got twitted because of the highly upvoted answers.

It is worth reiterating that there where flags on all questions, including ones closed by a single moderator. I will keep an eye on the one I re-opened, hopefully answerers will not fail to read the notice for yet another time...

2
  • 1
    I wasn't even sure that I thought 15% was too much, though in the end I would probably say that it is. At least, that was the threshold where I started to notice. Also, your point about SO migrations is a really good one -- if the hotness algorithm could take that into account (maybe wiping clean the hot-making action on anything that's moved, so that "hot" is relative to the new site) it would probably help. Finally, I agree completely with your three points, but if a user isn't already wired to know #2, to them it just looks like they were led to something they can't do (bad UX).
    – jcmeloni
    Commented Mar 5, 2012 at 13:25
  • To be honest 15% not being that much was based on what I've seen on other site's tweets, strictly anecdotal. Of course there's room for improvement, but right now we can't do anything without the team's help. Now that you've posted the list the community can review the questions and re-open them, if improved, thanks!
    – yannis Mod
    Commented Mar 5, 2012 at 13:40

You must log in to answer this question.

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