6 I do not think author’s solution needs to be visible, and certainly not to those who have not solved it. I would prefer a time to be attached to the solution/approval. The reviewer should not be stopped at 15 minutes though.
7 I know this too is low priority, but this could be checked on submission. Confirm the range of outputs is >2.
8 & 9 I understand these are not urgent concerns but I would love to see them addressed eventually.
If you require a solution to approve, some of the tougher community puzzles might get stuck in process longer than you like. You might consider just passing with a single approval if the person has a high reputation (like the chat moderator pool, or maybe the top X in CP).
What are those “contribution comments” you’re talking about?
Here is my response to each point:
Clash difficulty
Extremely hard and extremely easy coc puzzles 100% should be refused.
The difficulty is tied to fun and you should reject non-fun puzzles.
That’s what you used to argue against Boulet’s extremely easy puzzles.
I can create extremely hard puzzles to prove my point if needed.
In classic puzzle form, it can live of course because players can choose what they play. In coc you shouldn’t be given a problem so easy that it’s ‘fastest typer.’
The problem is that we already have refusals with reasoning like ‘easy’/‘too hard’ and we will have that even if guidelines would change.
But the guidelines can give power to the contributors that read them. Do you want to give them the power to create puzzles like Boulet’s?
Yes, we can skip all of that by putting work on the bot but the bot is already working and I still get those puzzles that take 15 minutes to solve by <300rank players.
Also, do you have data from the bot on how many puzzles are filtered because of difficulty and how many clashes it took? If it takes 15 or fewer frustrated players to filter out too easy/hard puzzles then I may agree with your solution.
Language-specific puzzles
Ok, I don’t know what are those “contribution comments” so I’ll not answer that point yet.
I have to add tho that there are obvious language-specific coc puzzles but there is also stuff like ‘True’/‘False’ and ‘true’/‘false’. And it needs to be clear that the second type should be avoided too if possible.
Complexity of statement
My point here is that some coc puzzle statements are hard to understand and I think that’s a more important issue than the hardness of the problem itself.
Take a look at the https://www.codingame.com/contribute/view/5788f7727cdcf84fc6c6baf79ce9105eb10d
It’s a troll puzzle but it shows the problem: it shouldn’t be a game of parsing texts for meaning, it’s a programming challenge so reading shouldn’t take more than 1-2 minutes max.
More serious examples: https://www.codingame.com/contribute/view/591461e2a3f2f8c60372447b5130fbef0c51 https://www.codingame.com/contribute/view/5791bf233f3db9e2b474a5405089960c068b (I don’t like this one especially. We played it on zapakh stream with a lot of <200rank people and only two of them managed to get 100%. One is zapakh who approved it, the second one is TheCrutial) https://www.codingame.com/contribute/view/553356953badd7a8b669b51cb79d86ade0d9
All of them have the same issue of having a wordy statement. But once you get it it’s not hard at all and it looks fine. Reviewers usually exclude statement reading from the puzzle-solving time just because it’s hard to include it when you don’t see the timer. Having a separate time-cap on statement reading should solve that I think.
Reviewers should solve the puzzle
Looks like all of the suggestions under this accumulated in “make a beta for coc with pending puzzles and ability to approve when if you solved it 100% under 15 mins.”
That would be very cool but sound hard to implement.
What is not hard to implement is hiding the solution/validator by default a showing it on mouse click
I don’t know how long you’d like them to be stuck.
IMHO a puzzle that’s unsolvable definitely shouldn’t pass. It doesn’t seem reasonable to me to allow judging whether it’s solvable with no solution.
When I was young and naive I’d have added “for anything but the simplest puzzles”.
But the simplest puzzles can have the stupidest bugs. We’ve observed them come up with them repeatedly.
I would just hate to see great ones like Space Maze have trouble getting approved, and discourage people from making more like that. If someone like you solves and approves it, there really is no need to look for two more opinions.
That reminds me.
Maybe Contribution author name should be hidden as well.
“I like you so I’ve approved” is present enough for JBM to start satirizing it. Yet no one else is talking about it?
I guess contributions like that are often approved quickly so not a lot of reviewers see it but
I would also suggest that all new contributions be required to be a WIP for at least X days (to be determined). I see many clashes that are accepted the same day they are first published. Most people do not check new contributions everyday, therefore these clashes receive very little feedback and the moderation is not representative of the community.
At the very least it would allow more experienced clashers to provide feedback on how to actually improve the puzzle before it gets accepted by people who don’t even try solving it. Some puzzle ideas are great but they need a little more polishing before getting accepted.
Perhaps more comments would help less experienced moderators in their decision (since you cannot expect all senior clashers to always be there to accept or reject in time).
But I admit that as a standalone suggestion, without anything discussed earlier, it would not do much.
I disagree with “it would not do much.” WIPs are getting a lot of likes/dislikes sometimes so someone is definitely checking them out.
There also can be some sort of after-approval period when it’s approved but still placed on the ‘pending’ page and not in the pull yet.
It would force a contribution to be hanging on the wall for some time to gatekeepers to get in and reject it if it’s too bad. Also, a third approving person will have some time to revaluate.
Anonymity will reduce bias. Of course it’ll not prevent bias entirely but at least people that want to be as objective as possible will be happy and it’ll be just a little bit harder for sockpuppets.
I guess license prevents absence of credit tho
I get a big “people are bad and anything but force wouldn’t work” vibe from you.
I mean, you can be right but I haven’t seen any data on the badness of codin gamers yet. Are you talking out of expirience? You’ve seen features like that be implemented and failing?
I’ve made an update of the guidelines for Clash of Code:
Let me know what you think.
I’m aware of the current issues with the moderation process itself, some of which are probably consequences of the quest map updates.
I believe the moderation process has its downsides but globally works. I see plenty of members playing by the rules, helping each other and trying their best to make great contributions for the community. Perfect is the enemy of good.
However I do agree that some sub-par contributions get abusively accepted, from time to time. I don’t think new guidelines will prevent these abuses.
I’m currently looking into the data to check the extend of the issue and how we could smartly prevent most of theses abuses.
Shouldn’t #4 be two separate points? I understand that it’s logical to put it together because it both about output. But it shouldn’t both be about output I think. The language-specific output format is the most common form of language unbalance but it’s not the only one. Also, It has little to do with the binary output. Yes, ‘true’/‘false’ and ‘True’/‘False’ are common and both binary and language-specific, but it means nothing, you should still treat those problems separately. And mixing them into one rule may lead to some confusion.
Still, very cool addition. Clarifies which vision of CoC you support, yet doesn’t break anything.
I am a problem setter at SPOJ PL. Would it be okay if I copied one or two puzzles here that I originally created at SPOJ? I am talking about problems I authored. Your guidelines say “must be original” but that statement can taste differently depending how you feel about it.
[Unofficial answer] If your OJ does not object to your multi-publishing I think it is fine.
In most cases you cannot just copy. You may have to alter the in/out style and requirement and problems’ difficulty level (solvable by slow languages) to fit into this site.
