# Delete all code-trolling questions

The tag has brought nothing but poorly conceived "challenges" to this site and the new users it has brought are posting poor knock-offs of an already poor idea. The quality of questions has fallen dramatically and the good questions that pop-up amidst the morass of rubbish get buried very quickly.

I propose that all questions be closed and deleted to stop the rot.

Well, lets get some facts. Feel free to downvote me.

First, you say that "the good questions that pop-up amidst the morass of rubbish get buried very quickly". In the first 3 pages of the newest questions, I see only 2 code-trolling questions, one is near the bottom of first page and one in the third page. Don't think that 2 questions far from the top buries the "good questions" very quickly. As usual, most questions are code-golf. From those code-golf questions, some are good code-golfs and some aren't, i.e. exactly in the way it always has been before. So your statement is not true. You might argue that this happens because we are getting more people posting here and good questions are buried, but this too means that new good questions are popping and that this site is showing some signals that we are finally getting out of beta, and I think that this is good. A proof of that is this question: Produce the number 2014 without any numbers in your source code

Three of the most voted questions in only a month, including the first and the second place. Surely my question was the most explosive ever in the site, but the third one (a very good code-golf) is the second one, currently is the hottest question in StackExchange and was posted only 4 days ago. Do you not think that this question would not attract a lot of inexperienced users too?

And, you are looking only to the questions, not to the answers. Truely, some answers are poor, exactly how it always has been before the day that I invented code-trolling, but some answers are very good ones. Personally, I learnt a lot in this site coding something that does something interesting and looking for creative and original answers that features creative and original ways to do something than looking byte-crushed illegible golfscript answers that serves for no purpose and that teaches me nothing. And this is the main reason why I think that code-trolling is valuable, I could learn for fun and laugh, and a lot of people too, and I think that learning while getting fun is the real purpose of this site. Even some very poorly conceived code-trolling questions got some really creative and original answers.

But now, we have a lot new users, thanks mostly to those three questions. Surely, there are some bad new users, but there are good new users too, and if you gets new users in any StackExchange site some are good and some aren't no matter what. Bad users will not get much upvotes, and will either go away quickly or will learn the way to become good users, and most good users stay. Sometime before I posted my question we had around 4 questions per day in area 51 and were declining. Now we reached 7, which I think is excellent. You could argue that new users should come more slowly (and if you does, I will strongly disagree), but this is no reason to delete code-trolling questions.

In fact, you are using code-trolling as a scapegoat and not addressing the real "problem": We simply have more users, a lot more users. Deleting code-trolling questions will not solve your "problem". In fact I don't see this as a problem after all.

And about the the "dramatically fall in the quality of the questions", again, only 2 questions in the first 3 pages are code-trolling, so code-trolling is not the problem here. In fact, I think that this is a misperception caused by the wave of new users. What I see is that some new users post some poorly conceived questions frequently missing a clear winning criteria, and code-trolling has nothing to do with this. Deleting code-trolling questions will not change this either. Surely, some people post low quality questions (as they always did, whatever the tag is), but this is the reason that we have the downvote button.

We have only 16 code-trolling non-deleted questions now, only 8 with more than 1000 views. All of them pretty quiet now. Don't think that deleting them will magically improve the quality of whatever is the overall quality of this site.

• There were a few times this week that roughly two thirds of the questions on the new list were code-trolling. It seems to have died down a bit now though. – Tim Seguine Jan 5 '14 at 23:25
• @Tim I did not saw this, but I think that this could be a side-effect of mass editing the questions, which bumped them to the top altogether, specially to add the disclaimer to do not take them seriously and to reformat the question titles. – Victor Stafusa Jan 7 '14 at 13:24
• "Feel free to downvote me". No, thank you. Have a cookie :) – Cole Johnson May 9 '14 at 16:37

I'd be more than happy to kill them with fire, but sometime early in 2013, I stopped trying to enforce some of our content rules as it had become apparent that something new was needed or the site would never get off the ground. (That, and I've gotten tired...I never intended to remain a moderator here after the beta.)

I'm neither happy nor particularly surprised to see that allowing popularity contests and other subjective criteria contests generated only a passing increase in interest. I think that the phenomena is more of the same, but on a more compressed time scale.

I do think that might as well be custom made to generate a maximum of me-too-ism in terms of both questions and answers. As I have said elsewhere "Let's turn a bad Stack Overflow question into a contest" struck me as a bad idea from the start.

For the nonce I am still looking for new idea that might revitalize this place and am still going to under-enforce our nominal rules so that people can give try new ideas a try.

A bit harsh to say that the idea itself is poor. While many of those who initially tried to join in on the fun didn't bring anything new to the table, I think that a good question can be just as valid of a programming and engineering challenge as plenty of problems on this site. It's just the initial influx of people when the idea isn't entirely developed yet, but I think that there can be value to keeping the good ones.

I'll reference the first code-trolling question as an example:

"I need a program where the user inputs an array of doubles and the program outputs the array sorted".

Writing code that does this correctly and efficiently isn't particularly interesting, and some probably find it trivial if taken literally.

But in , the challenge is getting 100% of the test cases in as few characters as possible. The best answers often use features of a language that others weren't aware of, or tricks such as bit manipulation to squeeze a few bytes out.

Based on the description of , the challenge is writing code that looks entirely correct, but fails upon close analysis. This sounds to me like something that could easily expand as a problem type on this site. However, it appears that the most upvoted answers are satirical in nature and not based on flawed code that looks correct.

To me, both and problems are equally valid types of challenges for this site. I don't believe that should be removed entirely. Instead we should continue to weed out the bad examples, including future answers that do not follow the guidelines.

If it's because the name "code-trolling" doesn't sound serious enough, then we could propose to rename it to something along the lines of "code-flaw-injecting".

• So which ones are the good ones? Which ones are presented and functioning as programming and engineering challenges rather than bullshit-spouting contests? – Peter Taylor Jan 5 '14 at 16:36
• @PeterTaylor Interesting effort in implementing an unusual algorithm, Interesting effort in obfuscation, might be an answer to some other question. And indeed there are much more. It would not be hard to create a non-code-trolling question where those answers would be valid which no more but very minor modifications. – Victor Stafusa Jan 5 '14 at 20:29
• @PeterTaylor I think "bullshit-spouting" isn't appropriate terminology for Stack Exchange. – SimonT Jan 5 '14 at 21:13
• @SimonT You think that, or you've seen a policy to that effect? 'cause if there's no such policy, I'd apply the moniker to your comment. – boothby Jan 5 '14 at 23:14
• @boothby I said "I think". Do you have something more constructive than attacking me for saying that? – SimonT Jan 6 '14 at 4:52
• @SimonT Calm down, I didn't attack anybody. But if my comment was an "attack", it was "in defense" of Peter Taylor's language. Your attempt to censor Peter Taylor was based on a bullshit rule you just made up (bullshit both because it's made up, and because we hates it). Unless, y'know, there's a policy you can cite. But I'd still call the rule bullshit 'cause we hates it. – boothby Jan 6 '14 at 6:04
• I agree that code-trolling should be about producing solutions that look correct on first glance or to an inexperienced coder, but actually do something hilariously wrong. I found that most of the low-quality answers to this effect were the ones that did some extremely obvious wordplay or deflected the question in some way. – Joe Z. Jan 6 '14 at 20:08

I'm sure you know you can click add an ignored tag to get rid of them from your own view of the site.

But I actually wouldn't mind if, like hats and April fools day pranks, they were voted off the site and/or disappeared after a certain date.

The first one was fun. That question got more views, upvotes and answers than any other on the site. There's been a huge new influx of users... most of them bad, but a number look promising.

The subsequent questions on the topic have largely been lame copycats. The answers have as low of quality as the questions. When the site suggested I review 10+ lame answers, I thought "kill it fire" too... but the initial rush seems to have died down. Let's wait a week or two.

If asking or answering these programming challenges (and they are programming challenges) is how some people enjoy spending their recreational programming time, who am I to judge?

I am somewhat concerned about "innocent" (?) people getting duped or sucked into these because of SEO and then taking them seriously. That is unfortunate and leads to more noise. So I might prefer to see the site automatically format questions and all answers with strong header/footer/notices.