On [underhanded], [code-trolling] and [obfuscation]

Question

Required reading:
The state of the underhanded tag
Has underhanded been getting too much like code-trolling?
The state of the popularity contest tag
And the motivation for this discussion of course: Rig The Lottery, but Don't Get Caught

There is a discussion in the comments of the last link about policy adjacent to underhanded, code-trolling, and the post is tagged obfuscation, presumably because the OP couldn't tag it underhanded.

The current policy is clear. underhanded and code-trolling are both banned from the site, and the post shouldn't be allowed. This was vocally supported by at least one member, and pointed out as a matter of fact by the moderation team.

On the other hand, it is quite clear that the post in question, were it to be allowed to stand, as well as the submissions, are good additions to the site. The post reached HNQ, and the vote count of sumbissions and the post itself show that most members find it as such. Downvotes were explained by the downvoter as being policy-based, and not quality-based.

To add to this argument, the last time this was discussed 9 years ago, this comment was made, and appears especially relevant:

Many existing underhanded questions deserve to be revived due to their ingenuity. It is a pity that one single answer here was accepted and referred to as "Community consensus" that was used to justify the butchering of this category.

– Andreï V. Kostyrka Commented Nov 21, 2016 at 11:36

It is my personal view that the community culture that created code-trolling and made underhanded into a mirror of it has since disappeared. I believe that code-trolling would not be as popular as it was then, but I do not think it would be a good idea to reinstate it (if you believe the opposite, feel free to create an additional policy discussion post about this). However, I believe that underhanded can be rehabilitated as shown in the linked post, hence this policy disccussion.

I see 3 options, for which I will create answers for everyone to vote on in order to avoid a repeat of 9 years ago. Feel free to add other options if you believe they are relevant.

Option 1: underhanded stays banned. Nothing changes, the current situation works well and the linked post, while good, cannot stand because it is against the rules (or, it should be locked as historic).

Option 2: underhanded is rehabilitated and can be used like any other tag. Low-quality questions using it are still closed as low quality, but a challenge simply being underhanded is not enough of a reason to close it.

Option 3: underhanded is rehabilitated, but with additional safeguards that place a heavy burden on the poster to make sure the post is a good fit before being posted. I propose mandating use of the sandbox for all underhanded challenges, and such a challenge must have at least a net score of +2 in the sandbox to be allowed on main. Not using the sandbox prior to posting is grounds for immediate removal regardless of quality.

Both option 2 and 3 imply unlocking relevant historic challenges that were locked when the previous policy was enacted.

• Seggan: Can you add an explanation of what underhanded is for us new people?
• Wheat Wizard: Please don't post poll answers. Allow people who actually believe in an option to explain their motivations.
• Themoonisacheese: @WheatWizard the goal was to avoid doing a repeat of last time. do you suggest doing anything this time or is it more for next time?
• Wheat Wizard: There was nothing procedurally wrong last time. Honestly I think the linked comment is simply upset because the result didn't go the way they would have liked. But I don't think there is anything do now.
• Themoonisacheese: @WheatWizard i also don't think something was wrong last time procedurally, but i think it is a real behavioral thing that people will rather comment on a proposal they disagree with isntead of posting the proposal they agree with. still, understandable.
• Themoonisacheese: @Seggan the tag wiki is pretty clear when you extend it: "An underhanded challenge is a challenge to write a program that looks as if it is doing one thing, but does something else."
• mousetail: I don't see how anything changed since last time this was discussed. There where "well received" challenges before the last ban too so I'm not sure how another well received challenge would invalidate that decision.
• Themoonisacheese: @mousetail in my opinion, the community culture has since changed massively. a lot of what I see from that time (Granted, i wasn't there so my perception is warped), especially in [underhanded], is golfing of the rules rather than golfing of the code, and i think this is no longer the case of the current community.
• Steve Bennett: Can someone explain what [code-trolling] is? Linking to the tag doesn't help since it doesn't have a real description.
• Themoonisacheese: @SteveBennett you have to expand the tag info using "learn more" link

Answer 1

Option 2: underhanded is rehabilitated and can be used like any other tag.

Low-quality questions using it are still closed as low quality, but a challenge simply being underhanded is not enough of a reason to close it.

• Albert Renshaw: +1 I'm of course biased here, but this seems like a good path forward, at least experimentally. As it currently stands site activity is in a decline due to a larger trend of StackExchange as a whole, so quality submissions should be encouraged. An excellent moderation team is already in place on this site and it is not as if we are overwhelmed with content needing review. I have full faith in the community's ability to be reasonable when it comes to content quality and, as with all tags, any are of course prone to low quality content, but that's why we have moderation in the first place.
• The Fifth Marshal: Given that I have flags that have been pending since April I'm far from convinced that harping moderator capacity will cure-all here.
• caird coinheringaahing: While, nominally, I like this, I think it'd need to be very closely watched once policy changed. The last thing we want is for people to see it emblazoned across the site that [underhanded] is no longer banned, and take that as tacit permission to start spam posting the kind of low quality nonsense we had before. Ideally, it should be treated like [pop-con] : functionally banned, but the occasional high-quality challenge be allowed, instead of simply being closed because "it's tagged with this specific tag"

Answer 2

I've thought quite a bit about this over the past day. And although it's too late to meaningfully contribute to this conversation I'm going to contribute anyway.

I believe there are two main types of people who dislike code-trolling (these are not exclusive categories):

1. There is a contingent who do not like popularity-contest at all.
2. There are those who find code-trolling repetitive and boring.

This is evident in the old meta answers. To quote specifically from Alex A.'s answer:

There are only so many times we can have the same challenge of "do X but make it look like you're doing Y, lol sneaky sneaky" anyway.

The 1st contingent is never going to be satisfied. I have come to accept that popularity-contest is never going away. So I will talk about the second position.

To investigate the second one, I think the core issue is that methods for solving code-trolling are very reusable. There are a couple of clever ways to do X but make it look like it's doing Y, and they work pretty well. And pretty much every way to do X but make it look like it's doing Y, also can do W but make it look like it's doing Z.

This really cool, unless you've seen it before. I think the really frustrating thing about code-trolling is seeing an answer get a bunch of upvotes for using a trick you've known about for nearly a decade (I've been on this site far too long).

Now, I think some people will point out that code-golf also has tricks that can be reused. However I think there are two key differences here that make code-golf fun over time:

1. A code-golf answer typically employs several basic tricks, and not every trick works for every challenge. Success at code-golf is about building a repertoire of tricks and applying them as appropriate. On the other hand most code-trolling answers employ one or maybe two tricks, and these tricks are rarely tailored to the challenge.
2. Applying the standard tips is the starting point for code-golf. If you apply all the applicable standard tips properly you have a passable answer. To have a good answer you need to come up with something clever for the specific challenge. code-trolling usually is just about the one trick you use.

I think with the way I've framed this, what I'm going to say is pretty obvious.

The issue with most code-trolling challenges is that they are duplicates. code-trolling (or underhanded if we want to call it that) should remain on-topic so long as popularity-contest remains on-topic (read as: forever). However, if a question does not provide some substantial deviation from the formula "do X but make it look like it's doing Y", it should be closed as a duplicate.

Closing wide categories of challenges as duplicates is of course precedented. Almost all song lyric kolmogorov-complexity challenges are closed as duplicates of the rickroll challenge. For basically the exact reasons I've outlined above.

• Albert Renshaw: I can respect this perspective... I do wonder if there is some way to reduce the challenges down to a sort of subtype of what the troll is and prevent duplicates within the subtype. (cont.)
• Albert Renshaw: For example in my case it is to obfuscate RNG for a specific input, this could perhaps be reduced to a subtype of 'RNG' and 'X' (where 'X' is some other term to represent the concept of having a specific input be the activation method of the troll, I am assuming there are only a few common names that would need to be made) Thus each challenge can be reduced to A,B where A is the thing being obfuscated and B is the trigger And so a ledger of which A,B challenges have been created already could be used to enforce no repeats. (cont.)
• Albert Renshaw: I am not particularly saying this is a good/feasible idea or not... I have no clue if such a thing would be well received or not, or if its even possible to categorize each challenge into this A,B format (though it seems it may be viable [and some creative people may eventually create new "B"s, which would then open up a ton of new [truly unique, non-repeat] challenges for the existing A's which have been exhausted previously, which could be neat)]) ... There could also be an issue of a rush to make a challenge for each A,B combo first before it gets locked, which may be weird/annoying...
• Albert Renshaw: Well anyways, I could speak back and forth on this endlessly but don't want to flood the comments with an entire novel. Perhaps the best course of action for now is to simply wait and see what happens. If a few more high quality challenges emerge more and more and are explicitly unique perhaps it warrants further discussion, but as of now a single instance after years of stagnation probably isn't worth much further discussion. Thanks for the consideration thus far!

Answer 3

Option 1: underhanded stays banned.

Nothing changes, the current situation works well and the linked post, while good, cannot stand because it is against the rules (or, it should be locked as historic).

Answer 4

Option 3: underhanded is rehabilitated, but with additional safeguards

that place a heavy burden on the poster to make sure the post is a good fit before being posted.

I propose mandating use of the sandbox for all underhanded challenges, and such a challenge must have at least a net score of +2 in the sandbox to be allowed on main. Not using the sandbox prior to posting is grounds for immediate removal regardless of quality.

• Themoonisacheese: this has my vote
• Wheat Wizard: This doesn't seem like a great idea to me. What's going to happen is that users are going to ask an code-trolling question, unaware that there are any rules about them, and then get told "Uhm actually you needed to sandbox this etc. etc." This is much more annoying and frustrating to the user than being told "Nice idea, but we decided we don't do that here".
• Themoonisacheese: @WheatWizard what about it being done the other way? the tag is nominally banned, but good posts in the sandbox may be granted exceptions?
• Wheat Wizard: I think that's a better start, but I think most people are going to assume challenges are not banned when they see them.
• Seggan: @WheatWizard maybe require a "underhanded challenges are allowed on a case by case basis" on the post itself?
• Weird Glyphs: Honestly, using Sandbox as a tool to choose good underhanded questions seems like a bad idea. Only a small fraction of CGCC users are actually going there to check other question ideas, which means that well-established users with ideas based around popular topics have at least a good chance to be approved, while regular users with more obscure topics will have to beg in TNB for people to come judge their challenge, and nobody is going to show up. In other words, the upvotes you are going to receive are not a good indicator for the quality of the challenge itself.
• Steve Bennett: Participation in the sandbox is so low, that forcing this path sort of amounts to an um, underhanded, way of banning it again.