# Has underhanded been getting too much like code-trolling?

Some time after the Great Purge, a huge wave of questions emerged.

Are some of the questions too broad and hence facing similar issues to those of ?

An example of an question which I found too broad was the now-deleted one called "Print an unexpected obscenity". The task was to write a program to print any insult, provided that the insult-printing is sufficiently obfuscated and undetectable.

Is this particular question too broad, and too much like ? It attracted a number of answers whose output was well-documented. For example, the output of this much-upvoted Bash/otherUnixShell solution is easy to predict, and it isn't very subtle. (I'm aware that I'm not really approaching this with a sense of humour here. I did find it funny though.)

You think you deserve an entry in the manual of manliness? Let's see:

man you


Output:

No manual entry for you!


Clearly, this isn't very .

TL;DR: Is it too broad to ask for any insult? And how unexpected does it need to be to qualify?

# Better Underhanded question

Write a program that makes 2 + 2 = 5 was a very fun question. I think it is objective enough. The task is to write a program which appears to calculate 2+2 (not too broad; asking someone to make a specific calculation is about as specific as it gets), but misuse some language feature to make it output 5. My answer is here.

TL;DR: Some questions can be as objective as gets. Scoring can't be done with any tag other than AFAIK as it is a contest for creativity.

The only question doesn't meet the criteria for .

# TL;DR

Some of the underhanded questions are too broad and too much like code-trolling. To an extent, underhanded is the new code-trolling. I wouldn't kill all of it with fire though. Some of the questions are good and objective. But others are too broad and hence too much like code-trolling.

• Well noted - I was about to post something like this as well. We should definitely learn from past mistakes, and here's an opportunity to use what experience we have from killing code-trolling and apply it to improve the site. – Doorknob May 31 '14 at 18:14
• I also agree that underhanded questions should be subject to particular scrutiny as they are easy to derail into code trolling. I think they can be very specific though. Well, as specific as popcons get - we're not going to convince anyone who is opposed to all popcons in general. But if you're fine with popcons, I think underhanded challenges can be quite interesting ones (especially if they specify in what way the challenge should fail). So do close them quickly if they are bad, but think about it carefully first. – Martin Ender May 31 '14 at 23:23
• I agree. Some are rather broad, rather hard to become interesting. Others are pretty good, though. – Isiah Meadows Jun 1 '14 at 21:37
• Hmm, I wonder if the trio of -1's on my Pi question was due to people thinking it was code-trolling-ish. – Kyle Kanos Jun 2 '14 at 19:34
• @KyleKanos Maybe... the problem with downvotes is that people tend not to give reasons why they downvoted. Great question though – user16402 Jun 2 '14 at 20:05
• @professorfish: Thanks, I wasn't sure how it'd do on the main page given the +4/-1 it received in the Sandbox. I'm not worried about the 3 DVs (especially since there's 22 UVs), but I am shocked at the 6k+ views. – Kyle Kanos Jun 2 '14 at 20:17
• I can go the rest of my life knowing I made a very fun question. I'll admit it is close to code-trolling but this site has always been about creative answers and it offers solutions that aren't as broad. – qwr Jun 14 '14 at 5:58

Short answer: yes, and the tag wiki is partly to blame.

Longer answer: although doesn't inherently hit Too Broad in the same way as code-trolling, it does suffer to an extent from a lack of objective specifications. The current tag wiki excerpt says:

An underhanded challenge is a challenge to write a program that looks as if it is doing one thing, but does something else.

But what a program looks like it's doing is subjective, and so a question which tries to stick to that description will generally deserve to be closed as Unclear what you're asking.

To pre-empt one anticipated objection: making the questions doesn't solve anything, because that's a winning criterion and doesn't fix the lack of an objective acceptance criterion.

(As a side-note: the tag wiki currently excludes some of the oldest questions in the tag, so that's a second reason to change it. However, I don't at present have specific suggestions for new text).

I would argue that we should treat more as a win condition - sort of an auto-popularity contest where users are expected to vote based on how sneaky the code is. Challenges should have an explicit specification, which should be used as the acceptance criterion. This does allow completely aboveboard answers to be "valid", but really it's no different than a answer that meets specs but is completely un-golfed.