string
ascii-art
kolmogorev-complexity
math
graphical-output or image-processing
array-manipulation
restricted-source
geometry
grid
game
random
combinatorics
quine
decision-problem
date
sorting
parsing
base-conversion
graph-theory
binary
path-finding
permutations
source-layout
interpreter
optimization
puzzle-solver
regular-expression
chess
printable-ascii
palindrom
natural-language
fibonacci
cryptography
code-generation
hello-world
fractal
compression
music
balanced-string
polyglot
matrix
internet

where did you get these?

these have more "life" to them

All of the tags from the first 3 pages that weren't language specific, a question type, or a near-duplicate of another tag

@ГригорийПерельман Alrighty

We need to specify a date and time when we each post, after ensuring everyone who might be intrested has seen this.

Eggs

I was thinking all posts would be simultaneous and use the same 2 tags. But they could be spread out with different random tag requirements

That could work, to prevent spam. First one person posts, then two new tags are generated, then next person posts, etc.

Then everyone should get two-three tag combinations to chose from in case Someone gets audio and regular-expression while someone else gets binary and restricted-source

moi

what is this exciting endeavour, and will it take less than 3 and a half weeks?

A competition to write a challenge, and likely.

well that sounds splendid, where do I sign up?

Here

We don't have a list of people yet

As you can maybe tell

why 3.5 weeks specifically?

We could do a trial run today.

my deadlines stop pending and start piling up in 3.5 weeks

either everything will be a disaster and I'll have to focus on my work

or I'll have no work to do and no motivation to do anything

it's 2:35 in the morning where I am, 21 hours sounds like a good trial-run timeframe

@HelkaHomba I don't have time for that, but feel free to.

this could be cool

@ConorO'Brien what are the proposals so far?

you can look at the results I think

docs.google.com/forms/d/e/…

well I have no reason to be asleep, is there a plan?

@PhiNotPi We scrapped that for the tag idea (I think)

I suppose I can merge the two, but I'm not accepting "Dragons" as a valid theme.

Actually, everything there is either silly or already covered by a tag

I didn't know we had a goat tag?

ah, I may have missed the context, is this intended to be an exclusive competition?

I like the idea of assigning everyone 2 random tags... maybe we can also randomize the order of posting, or require everyone to pre-commit a decent outline of their challenge spec before any of them are posted.

Well, the idea was that we don't tell too many people, and hopefully not get a spam of posts that way.

But I suppose that's not really fair.

fairness isn't really my department, do whatever makes sense, nothing wrong with demanding record of achievement. really depends on what you are trying to achieve with the competition

e.g. must have codegolf.stackexchange.com/help/badges/91?page=1

Course, we can't stop people from writing challenges anyway

seems like a generous lower bound ;)

Ok. Must have codegolf.stackexchange.com/help/badges/93/socratic :P

I don't even have inquisitve ;(

If we do the thing with randomized tags for everyone and posting at different times, we should compile a list of participants.

Me

I haven't even asked 5 questions, so write me off

@VisualMelon you could help generate the random tags

Just remember everyone should have the same amount of time to post once seeing theirs

ah, the impartial bystander, my favourite role

@HelkaHomba wow, that's actually my "Next badge"

57/100 though

Information pertaining to the contest is here

Feel free to edit as necessary

Off to meta SE?

You'd better stick a "Spring 2017" on the end of the name, don't want to create confusion the next time you run it

@HelkaHomba you probably want to add your name to the HackMD ^^^^^

I feel like semantic edits (like rule clarifications) should be allowed--we do want quality questions on this site, after all

Fortunatly, SE saves revision histories.

We can "Freeze" a post and still allow said semantic edits.

Lots of rules added hackmd.io/…

Is there a way to give all participants random tags such that the arbiter/referee guy can't fix them?

use an online dice roller

Not that I don't trust @VisualMelon, just curious if that's possible. Like where we can all be sure it was done fairly

I wouldn't trust me.

You'd need a multiplayer dice roller

they exist

alternatively, take the hash approach: blogs.msdn.microsoft.com/oldnewthing/20060523-17/?p=31093

everyone writes their own obfuscated tag table, which maps intergers to tags

they declare the hash

someone else produces an integer

then they declare the table, and determine the tags

that sounds complicated...

yeah, but it doesn't involve trusting anyone except the hash designer

Guys

Write a script to get a random element from the tag list

Use the user's SE id as the seed

Put it on TIO.

I can approve of that

@ГригорийПерельман But then couldn't they figure out their tags too soon and have extra time?

I'm assuming the minute I know the 2 tags I then have exactly X hours to write and post the challenge.

~23% of the "theme" suggestions pertain to goats

SE id + some other constant, which is determined when your time starts?

that could be crafted

(have to trust the person that produces that constant)

Hmmm

Does this work -> rolz.org/dr?room=PPCG

Blocked by a web filter for me, but probably.

It's a chatroom like this but you can roll arbitrary dice and more (no reg required)

Convenient

here is a totally legit random tag generator

... oh, TIO link is too long for chat

tinyurl

tinyurl.com/mmuyjnq

definitely something that needs reviewing

Maybe

TIO link for seeding people via username

you've probably already found a solution

Well, my username gets "polyglot quine" which maybe has been done to death

Just now, and we've also realized the problem with the solution.

Thats ok, I'm not sure Imma participate anyways, what with my 1 challenge

But anyone can do that, and find out their tags early.

Oh yeah

oh.

puzzle-solver ascii-art seems pretty straightforward

Welp, I'm screwed.

Is this official?

no

Still debating over best way to assign tags

I feel like it should be random, seeded by username, so that even if you have the program, you can't get the tags

@MistahFiggins But then you could if you knew the username

rolz.org/dr?room=PPCG is also an option where we can all see shared random results

seed it with their rep

that won't encourage gaming the system atall

what are we trying to do? make random assignments but keep them secret?

Oh ye, hmm

I'm sure there's some fancy cryptographic solution to this.

@PhiNotPi Basically. It's fine if they're generated truly random on the fly and not secret

i.e. dice rolls when someone is about to start their X hours

So people can't determine what the assignments are before their time is called.

The thing it, we can't just say "Hey guys I rolled a d20 and got a 17" due to the problem that we don't want to trust anyone.

we could just seed it with time on TIO, and trust that we are too imprecise to game it

@MistahFiggins it is seeded by username

can we trust caching?

No, you can reset that.

figures; drat

The dice rolling thing was ok, I think. what was the problem?

we could do something like "the time you respond to this message is your seed"

we don't know the microseconds on that

@VisualMelon But if someone says "I'll start my challenge at time T" they may have already checked that that seed gives easy tags

yeah, but we seed it with sub-second precision

i.e. even if you script up the request, you can't be sure when it will submit

because nothing is predictable enough

@MistahFiggins There isn't one, we're just wanting to see if something is more convenient.

you're sure you don't want to take the hash approach? ;)

@VisualMelon Maybe. But how do we trust the tags someone says they got?

it won't work because caching can't be trusted

(as Pawel said)

@VisualMelon You just need 1 hash, and everybody picks (openly) 2 different numbers after the secret hash is created

Then, the trusted person reveals the hash

no no no

@VisualMelon that's looking less complicated...

:P

if you are going to the effort of hashes

then you have no trusted parties

you rely on openness and identity for everything

for example: Helka picks a challenge name (one from PPCG), hashes it, and sends us the hash

then someone generates a random number

doesn't matter who, we just agree when it has been created

then Helka declares the challenge name, and from it and the random number we seed an RNG to produce the tags

We could do the random thing, but the seed is the xor of numbers given by at least five people

we verify that Helka didn't lie by hashing the challenge name to check it matches the hash provided

yeah, just take it that no-one has the computing power to generate solutions in real time

I see

Jeez, let it never be said cryptography is an easy field.

(the benefit of the hash approach is that we just have to agree a protocol, we don't have to trust any service provider (other than SE for identity and openness))

And we can't really not rely on SE

exactly ;)

@VisualMelon Another way is, when someone is ready to start their challenge, roll 1D49's (1 die with 49 faces, as there are 49 tags) until 2 unique numbers come up. Index the tag list with those numbers to get the tags that that person must use.

#1D49 works for rolling dice everyone can see here: rolz.org/dr?room=PPCG

that works too, if we trust Rolz

Well, @ГригорийПерельман can't see...

that's fine, we just lie to him and everyone agrees it will be funny

I can on my other computer i.e. in about an hour

"you got random aproximation"

@VisualMelon Unless one of us runs it...

in that case we couldn't trust it

(we can trust TIO, because we can independently verify the results)

I think that's a level of paranoia we don't need to accommodate.

(with seeding, that is)

@VisualMelon Dennis runs it

He can do bad things

doesn't matter

And half the languages there were made by us :P

if are seeding stuff then we can run the script on our own machine

Tru

use C# then, I promise I'm not Anders what'shisname

But really, I'd trust melons method or mine just fine for this

Ye

just use Rolz and TIO

It doesn't have to be the same method for everyone, right?

no

If someone can't access Rolz, try a different one.

alternatively... there are standard sources of randomness

(timestamped randomness)

@ГригорийПерельман Weird how "no" or "yes" or "right" would all be affirmative responses to this

Yeah

we'd better decide how to randomly assigned random tag assignment methods...

can't have people analysing them and selecting favourable protocols

Let's use Rolz

Roll a d29

Map numbers to tags in the order they appear on the hackmd

There are 49 tags. (Though I'm not against culling a few awkward ones)

Like answer-chaining definitely

that's too tricky for this kind of thing

If anyone want to make changes, they can.

I must say I'm a tad confused about this whole thing: you want more good questions, so you are taking people that already produce good questions, and imposing constraints to make it more likely they will produce not so good questions?

or is this about greater tag diversity?

I just wanted to (try to) beat someone at writing a good challenge ;)

Now that makes sense to me ;)

It's turned into a hashathon

By the way, feel free to add tags.

Can I participate?

Sure

Hackmd pinned to starboard

Sign up there

@ГригорийПерельман if someone deletes all that is there a backup

@HelkaHomba HackMD saves revision histories

@MendeleevLemon Sigh First it's "DownX" and then it's "XLemon", what's next

@MistahFiggins Famous scientists

I think that happened a while ago

like 2 weeks

I like this idea. I don't know that we need to go to great lengths to cryptographically secure our "random" tags. In general, I trust people here enough for that. It's not like we're exchanging money.

@MistahFiggins DownXLemon

Do you like my idea?

@HelkaHomba Well I would win since I've made all recent challenges from scratch and there has been 4~5 successful ones

:P

This is kinda cool. I would like to watch you do this and think it will be cool

@HelkaHomba do you want to write up a meta post annoucing the competition? I would but I'm bad at making meta posts >_>

Actually, nvm, I'll do it.

Out of curiosity, what's the point of "You must post within +/-5 minutes of your deadline."?

I don't think there is one.

We collectively decided it was a bad idea and no one removed it.

One sec

Also why is everyone talking about good challenge writers and no invited (or even mentioned) @Zgarb? :(

He doesn't hang out in chat, so we have no idea how to invite him. We can't really ping him

Hopefully he'll see the meta post.

> and X hours to write a challenge

We haven't decided all of the rules.

and what if that happens on night?

oh ok

Your time starts whenever you're available though.

It's not simultaneous, we don't want fifteen new posts within a few minutes of each other on the front page.

@MartinEnder By the way, want to sign up?

but thanks

@HyperNeutrino Probably not.

@MartinEnder So you decline?

yes

"You are allowed 1 major edit to the challenge within 1 hour of it being posted. A major edit is any edit that either adds/fixes test cases or clarifies some information. Any number of minor formatting/grammar tweaks are allowed during the first hour. After that, the answer is considered frozen. Further edits will not be taken into consideration during judging." How is this supposed to work? I thought the score is just the number of votes at the end?

Or do edits forfeit the challenge for the contest? If so that seems like a really bad idea for challenge quality.

> You get 3 sets of 2 random tags

we have to use both?

Quick question: when this contest will start?

I like the idea of this competition. I would only suggest that the number X of hours be rather large, of the order of a few days (unless that goes agains the spirit of the challenge). People may have commitments or deadlines that prevent them from preparing the challenge in a short time

@Zgarb hey, do you want to participate in this? it's a challenge writing competition

@betseg Yes

you've written good ones before, so we thought we'd invite you

@LuisMendo It's not simultaneous, everyone starts when convenient.

@MartinEnder Too many edits would mean you forfeit since I think there needs to be some sort edit of limit. Or maybe we do something like: every edit by author counts as -5 upvotes for this contest after the second edit.

It still discourages people from improving their challenge.

I'd say the quality of an actual challenge on main should be more important than participation in some unofficial challenge-writing contest.

Competitors here should realize that if they know their challenge is bad, it would be better to forfeit and not post it in the first place, so as to avoid lots of bad new challenges. If they know it's bad they probably have a slim chance of winning anyway.

there's a difference between a challenge being bad and having to fix some issues in a great challenge that weren't obvious from the start

in fact, since people might be trying to break new ground for the contest, it's even more likely that those challenges might need some touching up after they were posted

@HelkaHomba sounds like something to be pinned IMO

@MartinEnder true

I also think that if a challenge is great and very popular after a few edits, the original idea must have been good enough to warrant a good score in the contest.

@ГригорийПерельман tis gon be hard

And the case of someone posting utter crap and then editing it into a great challenge later to extend their deadline or whatever is prevented by forfeiting the challenge the moment it gets closed.

@betseg That is the point

@MartinEnder I was thinking it only matters if it's closed when we record the scores

Like, if it gets closed as unclear for an hour or two while more testcases are added thats ok

but idk, I'm ok with an "if it ever gets closed, you're out" rule

@HelkaHomba that's a good point actually

@LuisMendo You choose when you start so you could schedule to have X hours free. I feel X = 2 hours is a good limit. X = 3 could also work but seems slightly too generous

I'm kinda ambivalent on the edits and closing. I mean, part of the challenge here is to write a great challenge on the fly, yes? That's kinda the point of the competition? On the other hand, getting tags like chess fibonacci as my two kinda stretches the boundaries of a "good" challenge.

@AdmBorkBork CMC: place a queen in the bottom right corner of every fibonacci square. Given N squares, how pairs of queens are attacking one another?

Sure, I'm #justsayin

Not actually sure it that's hard or trivial, but ideas can be plentiful :D

@AdmBorkBork You get multiple sets of tags to chose from so that doesn't happen.

Oh, I missed that

^^ that too

That makes it different

With that, we should be stricter on the edits/closure rules, IMO.

I agree that that makes sense for the competition, but the effect will be that people will keep their challenges in a suboptimal state to remain in the contest, which is bad for the overall challenge quality of the site.

Well, get it right the first time. :D

@Riker It would be fun, but I don't think I have time for it now.

@Zgarb hm, okay

@HelkaHomba CMC: knight's tour on a Fibonacci square?

@Zgarb When will you have time? Time slots will stretch over a long period

These rules seem to encourage poorly planned and rushed HNQ-bait challenges

We're accepting suggestions to improve them.

Requiring challenge writers not to get feedback or make multiple edits tests their skills but makes challenges suffer

How do you suggest improving the rules?

I'd suggest rethinking the whole premise to emphasize quality as the end product, rather than a trial by fire for the writers

We can't really allow sandboxing the posts while also remaining fair to everyone.

maybe they're allowed to privately talk to 1 user?

I don't really like that.

I think the solution does end up being in changing the edit restriction

HackMDs are globally editable, write in anything you think is best

What if each answerer scored how much they liked golfing the challenge, and average those scores?

What if the challenge you create is some incredibly difficult task, that eventually gets a single answer after several weeks?

I think these would reflect quality more than voters, who I expect will vote more on humor or presentation than the value it brings as a challenge

I think that I'll remove the restriction on edits entirely.

@xnor We could have a panel of judges from high-profile members of the community that aren't participating

ehhh

Like Martin Ender and others

kinda subjective

It'll be subjective regardless.

The criterion is "Best challenge"

Well ... does it need to be a contest?

What if we just have it as an event?

A way to have challenge writers think outside the box and write something different for the site.

Me right now

@AdmBorkBork This is a great idea

I mean, we've been hemming and hawing about creating some sort of recurring event for challenge writing, like the fortnight challenges way back when, but it seems we haven't been able to come up with anything specific.

@HelkaHomba Ah, that makes more sense. But it's probably not the sort of competition for me. I may find it hard to find 2 or 3 hours in advance

I'm not sure if a competition would change the number of challenges I set, although it might help to give me ideas

I would say that I think the competition would work better if anyone could just join in without formally registering

maybe produce a formula that selects the sets of tags based on PPCG user ID and some sort of keyed hash, release the hash at the start of the tournament

that way people can just drop in and decide to participate

@ais523 The "Formal Signup" thing is just so we know to ping you when it starts. Anyone can enter whenever. I want to point out that each person doesn't go simultaneously, we all go at seperate times with seperate tags

The problem with X hours is that some of us may not be awake to participate, See: Timezones.

so can I just join?

@TacoLemon there are no fixed X hours. The X hours start (for you) whenever you're ready.

All is well then.

cool

A vague idea to remove the quality issues while keeping the contest nature: one per week each user can generate a tag pair (or several) and post a corresponding challenge within the week. Sandboxing, edits and everything allowed. And then we simply keep track of a global leaderboard of how many positively-scoring challenges you've written in this format.

Maybe keep track of streaks or penalise discarding tag pairs to get a new one.

simplest possible idea: a random challenge inspiration generator that anyone can use at any time

without any formal victory conditions for using it

But I think removing the incentive to get a lot of votes (while disallowing sandboxing) would go a long way to avoid HNQ-bait results.

@ais523 I guess the difference is, that is just a way to look for ideas until you like one whereas here the idea is more that you need to think outside the box a bit to work with the tags you're given.

what's the opinion on HNQ-bait, anyway?

sometimes you can do it intentionally, sometimes it's 100% unavoidable

maybe we should have judges?

HNQ-Bait?

and sometimes a question's unlikely to make HNQ (likely due to being too difficult to attract a flood of quick answers)

@TacoLemon: a question which will almost inevitably make the "hot network questions" sidebar (shown on almost every page of every Stack Exchange site)

the formula for it is broken, to the extent that it's easy to get there intentionally if you want to

PPCG currently has a question (this one) at the #2 spot on the list that the questions are drawn from (meaning that it's considered the second most worthy to be there)

it's not uncommon for PPCG to take something like 3 spots in the top 10 (especially if people from PPCG are trying to get there)

there's various debates about how much of a good thing and/or bad thing PPCG exposure is

I'm just amazed that Excel is winning.

it just happens to have a builtin that solves the problem

that tends to cause languages to win

I'm surprised nothing else does.

(except sometimes Mathematica, which has so many builtins that they often need fairly long names, allowing golfing languages to reimplement the builtin in fewer characters than its name)

Until someone bothers to figure out how mthmtca works.

wait dammit wrong room

sorry