# Is this interactive puzzle appropriate on Puzzling.SE?

There have been proposals before for semi-interactive puzzles [1, 2]. These are puzzles where more information is revealed by the OP as parts of the puzzle are solved. This is not deemed a good fit.

I am toying with the idea of posting a puzzle that is fully interactive, in the sense that it includes a Stack Snippet 1 which allows puzzle solvers to ask questions and get answers immediately. This has been proposed before, but no consensus was reached. (The proposal is rather … succinct, to be fair.)

Would the below question, polished and tweaked, be a good fit for Puzzling.SE?
Closely related: how should the question be tagged? I include proposed tags below, but am wondering if it makes any sense to include a new tag for interactive puzzles.

What can be found in the land of No Idea?

The land of No Idea is a curious land, where

• many objects exist that we have in our world;
• some objects and creatures exist that we do not have in our world;
• many objects do not exist that we do have in our world.

After much studying, I have found a pattern in what can and cannot be found in the land of No Idea. Challenge yourself and see if you can discover the pattern too! Feel free to ask me about any object, and I will tell you whether it can be found in the land of No Idea.

Remarks.
• The code of the snippet I link to has been obfuscated and minimized to avoid accidental spoilers.
• Play fair, don't spoil this challenge for yourself by reverse engineering the code in the linked snippet.

________
1 Unfortunately, Stack Snippets apparently don't work on Puzzling. I can see why. Lacking them, I can link to a runnable snippet as I've done above (the snippet, like the question, will still be tweaked).
I have included the code in the question as a comment for prosperity.

As you've already discovered, Snippets don't operate on Puzzling.

There's a strong preference to keep all puzzles self-contained on site. This is to keep content within the puzzle posting itself, so the puzzle does not become broken should the external link break, move, expire, or be deleted. I'm not as familiar with jsfiddle so don't know if it reaps content not accessed after some interval, but several similar sites do; we wouldn't want the meat of the puzzle to be inaccessible to future visitors because nobody happened to visit the page for a year or three.

In any event, there's nothing that will prevent a sufficiently enterprising person from reverse-engineering the code to get to a valid answer. Arguably, you're dealing here with an audience able and primed to do exactly that :)

In the past, when someone proposed an interactive puzzle (as in, posted in bits over time as earlier parts were solved), a number of concerns were pointed out but the final suggestion was to try it and see how it went. And it went more or less as expected, which is to say, not well. Your proposal is for a puzzle that is present in its entirety at post-time, so as long as the content remains in place and the would-be solvers don't go spelunking through the code, it's much more likely to be successful than past attempts at something semi-interactive. In that vein ...

• My chief concern is with not keeping the puzzle self-contained and on site.
• I'm also slightly dubious that this format will allow for a quality puzzle that doesn't rely on the gimmick of having an script verify if your guess is valid. A well-clued puzzle doesn't rely on the poster (or the poster's script) being there to answer questions along the way. See, for example, the many puzzles under "word-property" (which seem similar to what you're looking to do), where words that both possess and do not possess some specific property are listed and the property they (do or don't) share is discerned from studying the static list. You're creating effectively the same puzzle but with open-ended lists built by the users through submission of guesses to some script, which seems like an unnecessary step.
• Publishing a snippet that embodies the property in code is just welcoming people to reverse engineer it, no matter how many ways you say they shouldn't. You'll never know for sure if it was solved, or if it was cracked.
• I would really hate for this to become a trend. Even a well designed puzzle that avoids all the potential pitfalls, will likely invite imitators that don't, and that makes the whole class of similar puzzles (including the original) look bad.

Having said all that, and speaking as myself and not as a moderator, I personally would be cautiously ok with a trial run. I've wondered about doing something similar in the past, and even dabbled a bit with interactive MathJax constructs, but never got far enough to make a usable puzzle out of it. I'm curious how others here feel.

• I pretty much agree with everything you've said here, except I'm even more cautious than you seem to be from your last paragraph (and am leaning towards "against"). – Deusovi May 30 '18 at 21:36
• I also generally agree, but am (with Deus) leaning towards the side of more caution, and this is coming from someone who's sitting on ideas that would require semi-interactive content (I've even got one partially prototyped javascript-based interactive puzzle). Additional concerns that make me nervous, over/above what you've listed: 1) where does the boundary lie between "puzzle" and "game"? 2) it subverts one of the unintended norms of PSE, that you can generally see the entirety of the puzzle and its depth upfront (I already dislike the majority of "imgur mazes" for that reason). – Alconja May 30 '18 at 23:17
• 3) there's the issue of trusting someone else's deliberately obfuscated and unreviewed code (which admittedly stack snippets would help assuage, and at the very least, if we allowed anything else, we should restrict it to browser based javascript as a minimum bar of security) – Alconja May 30 '18 at 23:27
• Thank you for the elaborate answer. For what it's worth, I also (like other commenters) pretty much agree with everything you say. Hence why I asked here, instead of directly posting. Some remarks: (1) to keep everything together, I want to include the code in the question, even if only in a comment (as in this Meta question); (2) this puzzle could indeed very well work if posted as a static list of words. I could add some more words as hints, if it isn't solved after some time. I was just wondering if the interactive aspect would add something. – Just a student May 31 '18 at 6:05
• @Alconja, I think you make a very good remark about the boundary between puzzle and game, I had not considered that. Essentially, the way I came up with this is that this puzzle is told in the way I wrote it down, and you can keep people guessing for some time. – Just a student May 31 '18 at 6:06
• In my mind the easiest way to make an "Interactive" puzzle is to post each section as its own puzzle. – dcfyj Jun 6 '18 at 13:16

No, this is not an appropriate puzzle for Puzzling.SE.
And I will for that reason refrain from posting it.

1. As already observed, Stack Snippets won't run on this site. Even with the code included in the question as a <!-- comment -->, this makes the puzzle depend on a third party.
2. The script is a gimmick that does not add much to the puzzle.
3. This blurs the line between puzzles and games. This site is for puzzles.
4. One of the unintended norms of Puzzling.SE is that you can see all of the puzzle upfront. Having a script deprives puzzle solvers of that feature, possibly relying on them entering exactly the right question into the script. This is unfair.
5. Finally, having the script somewhere will invite people to reverse engineer it. It will always be unclear if an answerer solved the puzzle, or cracked it.

Thanks to Rubio, Deusovi and Alconja for their feedback [1, 2, 3, 4].