Boring

I find that Code Review is getting pretty boring and unprofessional. The reason why I think this way is that many good questions about real-world problems don't get very much attention. Instead, the hundredth implementation of linked-lists, heaps, project-euler, factorials, fibonacci, sortings or other programming challenge garbage get countless upvotes and answers. I'm not posting any links because I'm pretty sure you know what I mean.

Poor quality

Consequently instead of attracting more professionals to come over to Code Review to share their code and use-cases, and who could also take a look at some other interesting problems and share their valuable knowledge, we seem to favor simple and recurring questions by not voting for the interesing ones and not encouraging their authors for providing as much context as possible so that everyone can learn from them - OP, reviewer and especially other readers.

This has also another negative side effect: there seem to be very few active users who write reviews or share their solutions to intriguing every-day matters.

Too forgiving

I also think we are often too forgiving with questions that lack context and don't get closed. This gives a signal that poor questions are acceptable and the only beneficiary is the OP because only they know enough to be able to apply any suggestions. This makes Code Review not very useful to others.


I wish there would be some additional categorizing to easier distinguish informative questions you can learn from, from those that are pure waste of time. But since there is no such thing I usually see Code Review like this: many upvotes(7+) and answers (3+) = very likely not worth the time; few upvotes(<4) and few answers (0-2) = highly likely something interesting.

What's next?

Is there anything we can do about? I really liked Code Review for it's every-day riddles and learnt a lot of new stuff myself either by trying to crack some of them or just by reading and seeing tools, APIs or libraries I've never seen or tried to used in a certain way before, but recently it's been somehow mostly only unexciting junk. I don't like Code Review as I used to.


This has bothered me for quite some time already and can at the same time be seen as another answer to To serve, what is it to ask for review? that tries to promote poor quality input. I 100% disagree with it. Both, well written questions and answers should be CR's target. Not questions with vague description and unclear purpose of the code.

Quite contrary to the other question I suggest we be a lot stricter about the rules and encourage people to take more time and present their code in a more attractive and informative fashion because I see Code Review primarily as a platform that should be educational to more people than just to the OP.


Question

In other words: what can we do about under-supporting/promoting good questions and over-promoting the not so good, attractive or simply bad ones?

  • 2
    "Is there anything we can do about?" About what specifically? I agree with your sentiment, but it's hard to answer the question if the question is this broad. Are you simply looking for any and all suggestions to make it less boring? Improve quality? Do you simply want the community to be less forgiving? All of the above? Again, I agree with the sentiment, but can you summarize your goal in a single question so the community knows what kind of answers you're looking for? – Mast Nov 4 at 9:47
  • @Mast oh, sorry; in other words I mean what can we do about not supporting/promoting good questions and over-promoting the not so good, attractive or simply bad ones? Being less forgiving for the second category could be one measure... but maybe there are other suggestions? – t3chb0t Nov 4 at 9:52
  • Hard to say. More participation would be a great start, but I don't have the foggiest how we'd achieve that. There are a lot of things we've already tried. Bounties, community-challenges, RoboSanta. But those don't solve the core problem of repeating questions by beginners and waning participation from people with experience. – Mast Nov 4 at 13:20
  • 1
    Well I don't write so many reviews like I used to write because of two reasons. One reason is that my work-load got higher, and well the other reason is you because we are targeting the same language and you are doing a pretty good job. – Heslacher Nov 5 at 7:04
  • 1
    @t3chb0t Let's make that better. Feel invited: codereview.meta.stackexchange.com/questions/9002/… – πάντα ῥεῖ Nov 5 at 18:23
  • 1
    I generally agree about the programming puzzle questions, I've never really been interested in "programming magic tricks", I'm more into hunting for memory leaks and hidden crashes. Just out of curiosity, were my questions interesting to you? – jrh Nov 5 at 23:07
  • @jrh yes, I've voted for every one of them and I even got two green marks ;-) – t3chb0t Nov 6 at 7:02
  • I'm on the same page as Heslacher: if I see a review which mentions all of the points I would make, I just upvote the answer and go on. Most of the regulars (the ones I encounter on C++ tag) usually make very dense and thorough reviews, leaving no opening for another review. – Incomputable Nov 6 at 19:02
  • 1
    @t3chb0t I can post more if you want, I figured there was low interest in the kind of code I made, so I stopped. Though strangely enough a whole lot of readers are apparently interested in FTP upload queues. – jrh Nov 6 at 20:16
  • @jrh yeah, this exactly what I mean and it shows how important it is to post questions that have high practical value! ;-) If you have more interesting stuff, freaky frameworks or crazy experiments please do post. I miss Dmitry Nogin's small utilities - they seem really odd at first but at the same time brilliant. – t3chb0t Nov 7 at 20:16
  • @DmitryNogin you're being mentioned - have you build anything new lately? ;-] – t3chb0t Nov 7 at 20:18
  • @t3chb0t sure, but keep in mind that posts with 1 score are dangerously close to getting garbage collected. – jrh Nov 7 at 21:33

Let me start this by providing / adding some numbers to the mix. The following are the all-time stats for questions on Code Review:

closed | deleted | count
 yes   |  yes    | 21.787
 yes   |  no     |  1.541
 no    |  yes    | *7.629
 no    |  no     | 54.699

* only 1.107 of these have a positive score

Note that our friendly neighborhood background job will automatically delete closed questions (especially those without answers) after enough time. Only a very small amount of questions seems to be deleted by the moderators and power users.

These stats suggest that we closed slightly more than 30% of all questions that were ever asked on Code Review. The question closure statistics (available to 10k+ users) state that in the last 90 days, that percentage was 37%. So at least in the last three months, we closed more questions than we'd usually do.

To some extent that can be explained by the start of a new semester and the accompanying influx of students. Most of these closures (almost two thirds) are concentrated on the "code not working as intended" reason. A quarter of closures is for "Lacks concrete context", and the rest falls between "Authorship of code" and the remaining ways to close a question.

Just looking at the closure statistics paints a rather not so useful picture of what's happening, though... As a moderator I can run some interesting analytics, though most of what I'm about to tell you can also be seen in 202_accepted's Stack Exchange Statistics Explorer

Since around a year ago, the 5-week average of questions has been sinking. Code Review peaked at around 350 questions per week in the 5-week average (caused by two weeks of almost 370 questions). Currently that average is hovering around the 250 questions per week mark.

It's then not really surprising to see answers on a similar decline. As recently as two months ago, the overall Votes also dropped by quite a bit, even though they were rather steady in contrast to the decline of questions and answers. As such I'm not too worried about that right now. It might be worth noting that the 5-week average of votes is currently around a thousand votes per week. The historical maximum in a single week is at almost 5000 votes in a single week. It's probably not a coincidence that happened around christmas ...

The clear picture that these numbers paint is that engagement is somewhat waning. So it seems to be not just you that's not quite as addicted to code review as before.

This is just to help put some numbers onto the feeling that you're having. What these numbers say about the kind of question that's recently popular is a different thing entirely.

  • 6
    "The clear picture that these numbers paint is that engagement is somewhat waning." Yup. I think more regulars left than we got back, others became less regular. We lost some experience on the way too. – Mast Nov 4 at 13:10

I find that Code Review is getting pretty boring and unprofessional. The reason why I think this way is that many good questions about real-world problems don't get very much attention. Instead, the hundredth implementation of linked-lists, heaps, project-euler, factorials, fibonacci, sortings or other programming challenge garbage get countless upvotes and answers.

Then downvote and leave a comment stating why you find the question uninteresting. Preferably guiding them to a guide on how to ask a good question


we seem to favor simple and recurring questions by not voting for the interesting ones and not encouraging their authors for providing as much context as possible so that everyone can learn from them - OP, reviewer and especially other readers.

Share the interesting questions. Tweet them, Facebook them, discuss them in chat.


I also think we are often too forgiving with questions that lack context and don't get closed. This gives a signal that poor questions are acceptable and the only beneficiary is the OP because only they know enough to be able to apply any suggestions. This makes Code Review not very useful to others.

I disagree with you here. You might think that this gives a signal to other askers but you know what? Most people who ask poor questions don't look that much at other questions and therefore don't know anything at all about this signal. You could close all questions about anything and people will still come and ask questions about those things. Additionally, poor questions are acceptable. There is an important distinction between bad question and off-topic question. Poor questions are acceptable - not appreciated, but acceptable.

To fix this problem: Post a friendly comment stating what makes the question poor and why and how it should be improved. See Frequently Posted Comments


I wish there would be some additional categorizing to easier distinguish informative questions you can learn from

What is informative and interesting is highly subjective. My only recommendation is: Have a good title. If you see a title that is about linked lists and you don't find that topic interesting, then ignore it.


In other words: what can we do about under-supporting/promoting good questions and over-promoting the not so good, attractive or simply bad ones?

Share and tweet the interesting questions. Post a comment on the uninteresting ones.

  • I also think one of the contributing factors is the student population. More experienced programmers would not probably have a question about a real world application. On the other hand, a student will likely have a question about code they came across or wrote in class. – FreezePhoenix Nov 5 at 17:47
  • 5
    I'm not sure that downvoting questions for merely being uninteresting or repetitive is the way to go. It would make more sense to upvote and call attention to good questions. Save the downvotes for questions that are actually bad. – mdfst13 Nov 7 at 14:10
  • 1
    @mdfst13 I disagree with you there. The downvote button merely means "I don't like this question." and it could just as well be used for uninteresting or repetitive questions. – John Hamilton Nov 8 at 6:01
  • 4
    @JohnHamilton Downvoting a new user's question purely to say that you personally don't like that question has the impact of telling said new user that you don't like them whether you meant it that way or not. It's quite rude and people should avoid doing it. It is one of the reasons why Stack Exchange is considered an unfriendly place. – mdfst13 Nov 8 at 18:12

If you're finding Code Review boring then it's time to take a break! This is part of the natural cycle of volunteer engagement on a Stack Exchange site — it's fun to start with, but after a while you've seen all the common types of question and you don't want to have to make the same points again and again in your answers. Find something more interesting to do for a while: Code Review will still be here when or if you're ready to come back.

There are systematic reasons for the aspects of the site that you complain about:

  1. Most posts will be from beginner and student programmers. That's because everyone is a beginner at some point, and beginners feel most in need of review.

  2. Beginners are going to be posting beginner exercises. Everyone starts with these kinds of exercise because that's how you learn to program. Another heap implementation is boring to you because you've seen dozens of them, but for the beginner it's something new and complex to learn. (Obligatory xkcd.)

  3. We're not going to see much professional code here, because of confidentiality and copyright. Even disregarding these concerns, professional programmers are better off reviewing within their organization, in order to benefit from shared knowledge and expertise.

  4. We're not going to see much open source code here, because open source projects have other and better forums for review, for example through the GitHub pull request interface. (Also, copyright may be a concern too.)

  5. Challenge problems are more attractive for reviewers, because the problems tend to be self-contained (don't require the reviewer to understand a lot of context), and well-specified (the reviewer doesn't need to engage in a long dialog with the poster to try to extract a specification for the code they posted).

Because these reasons are systematic, you're going to find it frustrating trying to defeat them. The systematic forces operate for everyone 100% of the time, but you only have limited time and energy. Better to spend that energy productively elsewhere.

  • 2
    I cannot take a break, I have a streak of 881 consecutive days - a break would be a disaster - or I'll create a script that keeps it alive ;-P – t3chb0t Nov 8 at 14:00
  • ok, I thought about it and I came to the conclustion that this is not the case. I find my job very interesting and solving new problems everyday is a lot of fun. If this was the reason why I should quit Code Review (temporarily) then I would have to tell my boss to give me a looooong holiday. I'd like to once agian emphasize that I never said that beginner questions were boring. I didn't even use this word in my question. Boring questions have no practical value, they don't challange you to see things in a different or a new way or to make some research etc. [..] – t3chb0t Nov 8 at 17:31
  • [..] Boring questions are about problems that no-one cares about. This means I don't agree with point 1 & 2. I also don't agree with some of the the other points but will replay another time. – t3chb0t Nov 8 at 17:31

In other words: what can we do about under-supporting/promoting good questions and over-promoting the not so good, attractive or simply bad ones?

We could be more aggressive about tagging questions , , and possibly some new tag. Then you could exclude questions with those tags from your view. That would make the remaining good questions stand out more.

You also might consider browsing the Unanswered tab rather than the front page. Because what you're saying is that you prefer questions that have not received much attention.

Or you might try this advanced search. Adjust the parameters as necessary.

  • This is not exactly how I ment it ;-) I don't automatically consider beginner quesitons bad and I see CR primarly as a learning platform - people are comming here to read/post something they can learn from. To me a bad question is one that is not educational to anyone, e.g by severely lacking context (aka: this is my code, how can I make it better? with a bunch of unknown variable types etc). If it's technically ok then it's just fine but it doesn't have to be good yet. Good questions are about interesting practical/strange problems that are very useful and inspirational to other people. – t3chb0t Nov 7 at 20:06

I think you are right, but I don't think it's the only problem, or necessarily a fatal one. For me, the problem is that there are decent questions to answer, but they aren't easy to find and there's no real recognition for answering. Here is a reasonable question that sat around for a month. I provided what I think is a helpful answer, and have received no votes or comments in the month since. Similar story here and here. These questions were not easy to find, and now I'm not sure it was even worth it.

  • Now how have received some ;-] I must have missed them. You can draw more attention to an answer if you use some more formatting or code examples ;-) Text-blobs are often difficult to comprehend on their own so probably that's why the recive less votes. There isn't anything directly visible. On the other hand, only code without comments is bad too like a large piece of code without proper explanation. – t3chb0t 21 hours ago

You must log in to answer this question.

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