What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 133 Stack Exchange communities.

Over the past few months, I'm noticing that (relatively speaking) my questions are

  • Getting low views
  • Not getting upvoted
  • Sometimes getting downvoted, en mass and quickly... then +1 by other people balancing it out
  • Ignored or not answered

As someone who likes to find problems and fix them, I need to ask... is there a problem I'm unaware of? Is it me, or has the usage (or expectations of Stack Overflow) changed in the past year?

When I think about the SO community I wonder if they:

  • Stopped upvoting as much
  • Shifted technologies (MSFT and .NET aren't as popular anymore)
  • Got jobs that prevent them answering questions (thanks http://careers.stackoverflow.com/)
  • Or perhaps we have more trolls on the site who like to downvote
  • Site usage has increased / increasing the noise

Either way, my feelings aren't hurt by downvoting, it just makes Stack Overflow a little less effective in sharing my solution, or fixing a problem.

So what do I need to do to be more effective on Stack Overflow?

I hope irony will not bite me, and make this a low-view, downvoted question. ;)

share|improve this question
6  
This is an opinion, but I believe that everyone is getting fewer views because of the increased volume of questions posted to Stack Overflow. –  Gilbert Le Blanc Nov 15 '11 at 13:17
    
Drop 500 rep bounties on your questions and you'll get some stellar answers... –  Adam Davis Nov 15 '11 at 19:41
    
@AdamDavis related: a 50 point Bounty has been completely ineffective... stackoverflow.com/q/2562162/328397 –  makerofthings7 Nov 15 '11 at 19:45
    
@makerofthings7 There are some questions that are simply not going to get an answer here. For instance, stackoverflow.com/questions/5747829/… . I don't know what I'd do with your example, it seems to be just as much a niche subject as kernel programming. –  Adam Davis Nov 15 '11 at 19:49
    
Here is an example of a question that was downvoted by "Herd thinking" and reversed course after I added a bounty –  makerofthings7 Nov 15 '11 at 19:53
    
Here is a second example of a question downvoted by "Herd thinking" note that a similar question in the comments was up voted. These recent examples in rapid succession are atypical for my yearlong experience with SO –  makerofthings7 Nov 15 '11 at 19:55
    
@makerofthings7 I'd say that your first example was too verbose, and it starts off looking like a "please help me find a similar control, or port this one" rather than a "I'm halfway through porting this control over, and I'm running into this specific error/problem - how should I fix it?" The second is a legal question, not an english question. –  Adam Davis Nov 15 '11 at 20:53

2 Answers 2

up vote 22 down vote accepted

Well, a number of them seem to be "how should I setup my repository", which is a bit open-ended and subjective, but perhaps more importantly: it isn't a trivial thing. Your last touched on OData, AppFabric, Entity Framework, caching, repositories and service layers (plus also some ASP.NET MVC and ViewModel thrown in) - that's quite a specific and complex combination, that isn't something people can easily setup locally to help you! It was also a 3-part question, which is not usually a good sign.

I would perhaps say: try to distill questions a bit, removing all parts that aren't central to the question you want answering.

You have, however, still had a reasonable number of views and answers, even recently. Ultimately, the site isn't going to answer every question you ever have...


This directly mirrors the feedback on your last-but-one question (from comments):

I don't say you should provide the whole source code but some code snippet related to the issue. Nobody will go to download any specific SDK or sample to diagnose your issue.

aka: help us to help you, by representing the problem in as small a sample as possible. In fact, in most cases the very act of reducing a problem to the minimum possible repro shows you exactly where the problem is, and you don't need to ask the question any more!

share|improve this answer
    
What are your thoughts about this question ... I wasn't expecting it to be subjective... –  makerofthings7 Nov 15 '11 at 13:13
    
@makerofthings7 my thoughts would be: performance is tricky! if you want results that apply to your scenario, you are going to have to measure your scenario. No one approach "is faster". Personally, I would look to tools like mvc-mini-profiler before doing much, to see if the performance gap is actually in the code you think it is (in truth: it is almost always in unexpected and unpredictable places). –  Marc Gravell Nov 15 '11 at 13:20
    
Also, the quote you have isn't a good sample... I actually do post full code samples and repro instructions. That specific question is about a MSFT DLL that operates slowly if the simple presence of a MSFT generated file is there. I need a project workspace to upload and show that code. Instead, I described the issue in the question. Feature request: Can SO host a project for me? (become the new GIT) ;) –  makerofthings7 Nov 15 '11 at 13:21
    
And as always, thanks for the constructive feedback! –  makerofthings7 Nov 15 '11 at 13:25

This isn't a full answer, but it's longer than a comment so I'm posting it as an answer. Sorry. :)

When I look at your profile, I notice your top 4 sites have the following stats:

Site           Questions     Answers    Ratio
IT Security    150 questions 52 answers .34
Stack Overflow 359 questions 76 answers .21
Programmers    32 questions  22 answers .68
Server Fault   75 questions  9 answers  .12

Now I could pull up Stack Exchange's data site, but I'm too lazy. So I'll use a rough figure. According to the StackOverflow questions page, there are 2.29 million questions. The latest question has a post number in 8.14 million. Subtracting the two gives us about 5.85 million non-question (aka answer) posts. This makes the ratio about 2.55. So it would seem the average question gets about 2.5 answers to it. This seems consistent with my experience as well, to be honest.

So if the average question gets 2.5 answers, then logically the average user posts 2.5 as many answers as he does questions. Now this can be incredibly lopsided. Some people have thousands of answers and a dozen or so questions. I personally have 941 answers to 34 questions. So heavily active users are pushing up that average. (Consider Jon Skeet at 17679 to 23, 768:1 ratio, or Marc from above at 8712 to 29, 300:1 ratio!) But even so, I don't think a 1:1 ratio is that bad to aim for.

The people you're trying to "get to like you" no doubt are aware of the fact that you ask many questions but post (in relation) few answers. I would strongly consider taking the time to answer a few questions here and there. Pick a tag you're good at and subscribe to it. Then spend just 5 or 10 minutes in the morning to answer a question or two. It really adds up and builds up good community karma.

The second thing I notice is your accept rate. In those top four sites, your accept rates are, in order, 69%, 70%, 0% (yikes!), and 69%. I'm not going to go through all your questions, because I know that accept rate isn't the be-all-and-end-all of whether you're following up or not. It could be you often don't have a good answer. So let's see the reasons why you have this accept rate, and what we can do about it.

I see two probably reasons.

  • You aren't following up on your questions
    • Solution: follow up on your questions
  • You aren't getting good answers to your questions
    • Solution: determine if your questions need higher quality.
    • Solution: continue to update your post as you, yourself, make progress on your problem. (This naturally bumps up your post in the queue, giving it more publicity too. It's win-win!)

Now, this was all statistics based. I pulled these numbers from public data. I know they're just numbers, and numbers can be played to be misleading. But, I also think if you use them correctly, they can be good indicators of what you're doing and how you're doing it. The numbers that are public are specifically designed to be encouraging of the proper behavior even if you try to game them. (That's not to say they can't be gamed. They can be gamed. Don't game them.)

So I say keep your eye on these statistics and try to improve them. Just make sure that when you do improve them, you're always making the right choice. Don't improve your accept rate by accepting random or wrong answers; improve your accept rate by editing the question and encouraging future posts. Don't improve your question answer ratio by posting crappy one liner answers. Improve your ratio by taking the time to follow things you're good at and helping people who need it.

Also, something I've been doing lately, try following a tag you're interested in but don't know much about. When you see an question, try to solve it. You'll probably not be the first, but soon you'll have picked up a new skill by being provided with a constant source of real world problems to solve with the framework or language you chose. Soon enough, you'll be able to answer some of the questions in it. And who knows when it might come in handy, right?

I'll step off the soapbox now. I'll close by saying with no hard feelings that right now, the community shows disapproval of your posts because they aren't meeting the expectations that we hold to all posts and posters. But you want to change that, which is why you posted here: that's the first step in the right direction. I wish you the best in improving your posts and enjoying SO even more. :)

share|improve this answer
    
I honestly knew this would come up, and I don't mind being a whipping boy for good behaviour. Thanks for articulating this, and now that I have a little more development experience behind me, perhaps I can contribute some valuable content to others –  makerofthings7 Nov 15 '11 at 18:13

You must log in to answer this question.

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