« first day (3778 days earlier)      last day (1080 days later) » 

4:00 AM
@Underslash Reputation is a good system. However, like most similar systems, it is prone to "grinding". Badges are the same. If you take a look at my rep graph, you'll notice a stupid-steep climb in Feb/March/April, because I wanted to reach 30k. Now that I have, I interact with the site in a much more moderator way, and I pick my golfs, rather than golfing everything in sight
 
> Pretty soon, I found that I made a little calculation before starting to answer a question: "Will this make enough of a difference in my number to be worth my time?"
 
I think an intrinsic system would be the same though. Just replace rep with time.
 
@cairdcoinheringaahing True, that is a part of a rep based system
@hyper-neutrino In this case, I would argue that that person shouldn't be helping if they dont at least at all want to help someone
 
This is definitely something I've caught myself slipping into a few times when I start losing motivation (either for code-golf specifically or in general) and I've definitely found old, unanswered, and exceptionally difficult challenges to go solve just to remind myself that it's fun to problem-solve for the sake of it and that it's fine if I don't get numbers, but just the satisfaction of doing it.
 
4:02 AM
Once I noticed that I was <100 edits away from 1) the Copy Editor badge and 2) being the #4 editor, I started editing like crazy (currently #3, closing in on #2)
 
I think the combination of having a strict site + rep system does weed out people that try to do low-effort grinding, which means at the very least if someone doesnt want to help, but still does, the answer wont be abysmal
 
I've stopped editing as unnecessarily now, but for a while, a substantial number of my edits were trivial
Damn you hyper :/
 
12 seconds
 
That was fast
 
Missed out on a flag tho :/ :P
 
4:04 AM
HN needs to be slowed down a bit, people will start asking questions and they'll find out HN's a bot
 
so, im curious (not trying to bypass the sus rule): Did that person just get suspended / banned?
 
Does flagging a spam count as helpful?
 
uh, wdym, like my flag?
 
Unless a mod (probably hyper) suspended them. Otherwise, they just lost 100 reputation (that they didn't have)
 
4:05 AM
ah
 
@Bubbler If marked helpful by a mod or 5 others, yes
 
Oh, OK
 
If a post is flagged for spam and then deleted via vote, the flags will be marked helpful and the post's content will be hidden, but the user does not get any penalty. At least, that's what I remember / have heard. It's a bit weird.
 
Are there any other challenges on the site similar to the "build tetris in GOL"?
 
In what way?
 
4:06 AM
There's "Build a digital clock in GoL"
 
@cairdcoinheringaahing Badges definitely do drive people a lot (prime example: me accelerating posts through the sandbox much too fast for what you're supposed to do in order to rep-cap :P), which is why there are a couple of tags I somewhat disapprove of, like Peer Pressure.
 
So... if SD autoflags using your account then you get a "helpful" count automatically?
 
Not really the GOL part
 
Arguably more impressive, as only one user did it
@Bubbler Yes, I believe so
 
@Bubbler Yep. Setting up autoflagging technically lets you passively farm helpful flags on random sites :P
 
4:07 AM
more "esoteric challenges that are incredibly difficult, yet people have done so" if you get what I mean
 
The only "downside" for that is that SD is network wide, so you don't really get a lot on one site
 
Anyone else here not really like the tetris question? It's seriously overrated imo.
 
Charcoal is only like ~100 or so flags away from Marshal
 
the question itself kinda is, yeah. TBH it's kinda not very well specified. the answers, on the other hand...
 
The answer is insane, but the question itself really does not deserve 1k upvotes.
 
4:08 AM
Sure, but that's what comes with an extremely impressive answer
 
I see some value in the question because there wouldn't have been such great answer without it in the first place
 
@RedwolfPrograms I think it's wildly under specified, should be closed as such and is not an interesting question. The answers however are the complete opposite
 
I'd personally like to see a lot more of those type of challenges, since open-ended is usually only bad when that openendedness leads to trivial solutions
 
4:09 AM
@Underslash I disagree with that last part
 
the thing is there's massively disproportionate effort between the question and the answers
 
I'd like to see more difficult challenges, but not underspecified ones
 
Honestly, I'd be in favour of locking and closing it, aside from the fact that the answers deserve upvotes
 
yeah, unfortunately history locks must freeze the entire thread
 
I wouldnt phrase it as underspecified, just that answers can vary a lot, yet all be equally as impressive
 
4:11 AM
I would disagree. What is "tetris"?
just that alone makes it a bit unclear
 
Correct me if I'm wrong, but can't you brute force it?
 
brute force what?
 
do you need one of each tetromino in succession or fully random? do you need to display preview? do t-spins work? which spin variant do the pieces follow? etc. there's a lot that should be specified that is totally up to implementation detail
 
wouldn't be a terrible idea to just slap on a disclaimer of "this question isn't up to our community standards but we're also keeping it open"
 
0
A: Sandbox for Proposed Challenges

dingledooperSum of set bits from 1 to n code-golfbinary Given an integer \$ n \$ \$ (n \ge 1) \$, return/output the total sum of set bits between \$ 1 \$ and \$ n \$ inclusive. To make the problem more interesting, your solution must run with a time complexity of \$ \mathcal{O}(\log n) \$ or better. This is ...

 
4:12 AM
because it really is incredibly vague
 
I agree with that
 
oh true. we can just edit it lol, unlike the other question i was trying to find an appropriate post notice for
 
@Bubbler Given that there is specifically a starting configuration that works (see the answer), you can brute force your way to that config
 
@hyper-neutrino Well, the answer with even the most basic implementation is still going to be incredibly impressive, which is why I wouldn't mind those sorts of things
might just be me though
 
But it still shouldn't be answered when the post author could just say "nah, this isn't what I meant"
There's a reason good challenges have to be well specified
 
4:14 AM
@cairdcoinheringaahing I don't think so. You can't judge a program's behavior with a program, if the language is Turing-complete
and GoL is TC
 
But the board is of a limited size
 
Maybe we should close the tetris challenge with a custom close reason
 
So the GoL used isn't really TC
 
Well, considering that post specifically, its hard to misinterpret creating tetris in GOL
 
@RedwolfPrograms I wouldn't be opposed to a mod-close, a la the 100+ score pop-cons
 
4:15 AM
No, you still need to run for infinitely many generations to see if it performs a task
 
and if the author had claimed it didn't fit, then they shouldn't have worded it that loosly
 
well that's the point - they shouldn't have worded it that loosely
 
It shouldn't have any new answers imo, and closing wouldn't prevent voting
 
closing might discourage clicking though? idk
 
It probably doesn't get much random traffic anymore anyway
 
4:16 AM
@Bubbler Yeah, I guess I mean that you can "roughly" brute force it, to severely limit the possible starting configs
 
they shouldn't have worded it loosely, but imo, only if they would have been disatisfied with the answers
 
@hyper-neutrino It's got 1k+, it doesn't need random clicks :P
Might be worth taking to meta tho
 
also, sadly, literally none of the people involved in this are active anymore :(
 
the old gen
 
@cairdcoinheringaahing You can't even roughly do it, do you think you can find "nth triangular number" program by brute-forcing all Jelly programs up to 10 bytes? How would you know the given program is correct?
 
4:19 AM
doesn't this require solving the halting problem lmao
 
Exactly ^
 
I'd be really interested to see the smallest possible GoL program which implements something like the collatz conjecture
 
If I had to put it into words, a good question should be exactly as vague as the asker wants to be, ie, they would accept any reasonable answer that falls under ambiguity. If they want vague answers, they should be allowed to ask for them, maybe with some sort of tag, but if they fail that, then its a bad question
 
@Bubbler Yes, you could brute force that in Jelly - I've done it. You jus have to use some smart constraints (e.g. for GoL you limit it at 10k generations, and only investigate those that show promise after that many)
 
@Underslash But the problem with that is...what if the people working on this had spent months on it, but then the author just replied with "no, this is the wrong tetris. sorry."
 
4:21 AM
Then thats when it becomes a bad question
ik im dealing in idealist scenarios
 
But allowing that in the first place is a bad idea
 
yeah, which is why there should be a distinction for that type of question
 
I think requiring challenges to be well specified is the best option
 
you can brute force a program that returns the correct outputs for a set of test inputs, and if your test data is good enough, that should be fine, but programmatically proving your program is algorithmically correct is... probably very difficult
 
Sure, solving the halting problem is impossible. But if you put sensible limitations on your brute forcing, that's not something you have to consider because you look for potentially correct behaviour in a limited space
 
4:23 AM
@cairdcoinheringaahing A bruteforcer can only get so close. It cannot prove that a given Jelly program calculates nth triangular number for all possible inputs (which is infinitely many).
 
like, if the author of the tetris question had said that the answer didn't properly meet the requirements, then thats a bad question
 
The bruteforcer itself cannot be a valid answer on a code golf challenge
 
but, as they didn't, and as they wouldn't have, then the question is fine
 
Any valid Tetris program should keep working for 100k iterations of GoL (as it should keep working infinitely). Therefore, if you iterate all possible GoL starting configs with a max number of on cells to begin, and only keep those that have 100k generations, you can do some manual searching to find any potentially correct solutions
 
Okay, but then how does one assess if a question is good or bad? It should be independent of the existence of answers.
Think about it this way - I only want to answer good questions. How do I select whether to attempt a challenge like that?
 
4:26 AM
Thats why there would have to be some sort of other distinction
Yeah, I get that part
you can't really tell before hand
 
And if the OP decides the answer is invalid, the question becomes bad, and then what do we do, delete the answer and then close and remove the question? At which point answering those types of questions was a waste of time.
 
@Bubbler Depends on how well it can analyse programs. If it can equal the concept of RS to nth triangle number, then it knows that RS works
 
thats why its idealic and not actually practical
 
@cairdcoinheringaahing Well, then you need to see the output of the bruteforcer and validate yourself that it works. And that means a bruteforcer should run fast enough to let you find a potential answer, which is not the case here (and for most challenges)
 
The point is that once the challenge is posted, an answerer should be able to determine if an answer is valid or not without OP checking it (barring like formatting or other minor technicalities; those are also technically unclear but it's pretty much impossible to write a challenge that can be clear in all possible cases as long as only edge cases need handling)
 
4:27 AM
@hyper-neutrino Score; only worth answering if it has a score of 10 or more and no Jwlly answer
 
of course
 
if you were guarenteed that a question was good, then you would always be able to answer it in good faith, regardless of how vague it was
so in order to make all questions good, just ban everyone that makes bad questions, problem solved
 
Well, then there's no one left :/
 
@Bubbler thats when the good questions start rolling out
 
4:29 AM
@Bubbler Yeah. It isn't solvable 100% by a forcer. However, with some cleverness by a person, you can limit the forcer to potentially correct solutions, and have the rest done manually. The buoy of the challenge is done by a forcer
 
@cairdcoinheringaahing hey, yeah it's been a while
 
@Underslash I would say I can make good questions. That doesn't mean all of my questions are good
 
@Underslash no, that's when the spam starts rolling in, because we've just banned pretty much everyone and nobody is here to maintain the site :P
 
@Stephen You've been pretty active in the review queues from what I remember right? Nice to see active reviewers :)
 
@cairdcoinheringaahing in this hypothetical system, any bad question would get you banned
 
4:31 AM
yeah I browse the site, do review queues, and answer tag:sequence questions
don't actively participate much anymore though, not since high school
 
@Underslash I'm one of 10 (I think, might be 12) users with the Socratic badge, meaning I'm good at asking challenges. If one bad challenge got me banned, I think your system is broke :P
 
obviously, if you ever make a bad question, it means you obviously cannot be trusted to ask good questions, which obviously means you should get banned
:)
 
@Stephen How's cQuents going btw?
 
my highest-voted challenge ever was almost not posted because it was a duplicate of a CMC
 
socratic has been awarded 12 times, though two went to calvin
 
4:34 AM
@cairdcoinheringaahing it exists, about a 0% chance I'll update it though
 
what is the socratic badge, at any rate?
 
so 10 users have exactly one socratic, and socratic has been awarded 12 times, so neither of your guesses were right :D
 
most room for more commands are 2 bytes+, I don't want to expand the char set past ASCII
 
@Underslash It means you've asked a question on 100 separate days
 
wow
thats a lot of questions
 
4:36 AM
s/question/good question
 
Maintaining a "positive" question record (I.e. not closed/downvoted)
 
Not exactly "good" but at least "not too bad"
 
damn, guess im ineligable lol
 
@Underslash someone (Calvin, I'd link but mobile is awful) has done it twice
 
i need to get working on socratic, lol :)
 
4:37 AM
^ Them
 
those are a lot of good questions
 
they also have 22x famous question, lol
 
They have more Good Question badges than I have questions :P
2 off tho IIRC
 
i have a rather low amount of badges tbh. and just content quality. my content quality is pretty average here i'd say
i've never had anything really blow up the same way stuff like the covfefe challenge did :p
 
@hyper-neutrino you have more than me at least :)
 
4:39 AM
@hyper-neutrino 3 more months and you'll get Constable :P
 
I did one thing and I'm proud of it and I'm never doing anything like it again, I'd rather get paid to think about stuff for a day
 
Maybe Sheriff, I always get those mixed up
 
sheriff
:P
 
code-golf detective
 
i am only #26 in edits apparently, lol
 
4:41 AM
@hyper-neutrino "This site ain't big enough for the both of us" *un-diamonds WW, JoKing and Doorknob*
 
@hyper-neutrino Closing in on #2 here :P
 
Is it just me, or does the monospace font in the ask question page feel bigger?
Might have to make a user script...
 
i could edit more aggressively lol, i could probably find stuff if i hunted for it
i consider myself a rather passive user though; i mostly act on things once they're brought to my attention (except when they demand a 12 second nuke :P)
 
Apparently, I'm 3 edits away from overtaking Razetime on Meta
I've made 109 meta edits since April 1st O.o
 
4:44 AM
I am, however, only 33 reputation from overtaking Fatalize and claiming the bottom-most spot on the all-time reputation leaderboard :P which I don't care about much but validates the amount of time i waste spend here :P
oh i'm 5th in meta edits
probably from status-tagging posts :)
 
how do you check your ranking in that?
 
I'm ranked so high I'm not even on the leaderboard
 
@lyxal Check the latest PR on Vyxal.
 
0
Q: Smallest slice of array containing a number of items

Redwolf ProgramsTake this array: [1, 2, 6, 4, 4, 1, 0, 0, 2, 0, 4, 1, 4, 2, 4, 8, 1, 2, 4] There are quite a few places where [1, ..., 2, ..., 4] appears, where ... is any number of items. These include [1, 2, 6, 4], [1, 0, 0, 2, 0, 4], [1, 0, 0, 2, 0, 4, 1, 4], and [1, 2, 4]. The shortest of these is simply [...

 
4:56 AM
10/10 descriptive title :p
This challenge is in no way inspired by me trying to find the optimal paragraph to turn into "Dutch will be the m..." earlier ;p
 
5:15 AM
olololol
 
> Try it online!
ry it online!
Try it onlin!
Try t online!
e!
Gotta enjoy some wacky TIO link texts
 
5:30 AM
@cairdcoinheringaahing oh I did not know i was number 1 there
musta been the spam from Husk LotM
 
@Razetime What is the answer for length 3, thickness 1 and the grid of
#  #
 ##
 ###
  #
 #
?
 
@Bubbler so, sheep can't intersect
it'd be 1???
yeah the max i can recognize there is 1
 
I can recognize 2 ways to get 2
 
ok yeah there's 2
 
obligatory shitty freehand diagram
 
7:13 AM
Defining "intersect" that way introduces a special case which I find unnecessary
Just an opinion though
 
7:38 AM
@hyper-neutrino I was very confused by where intersection comes in until I realized you drew the second hash too far to the left
 
7:48 AM
@lyxal I was messing around and trying to fix the bug, but it broke.
 
I did it. I finally did it.
Now I can take a break.
 
Did what?
 
Check the main page.
I answered something.
 
8:05 AM
@Ausername which bug?
 
8:21 AM
@StackOverflow I approve
The streams have crossed
 
8:45 AM
@lyxal The | in double-char strings bug.
 
9:04 AM
@lyxal P.S. I updated the docs. /
Ugh I keep getting reviews in which an older question is flagged as a dupe of a newer one.
 
9:31 AM
0
A: Sandbox for Proposed Challenges

A usernameProve there's another prime A simple proof that there are infinite prime numbers goes like so: Suppose there's a finite number of primes. Then we can have a list of all primes \$[p_1,p_2,p_3,...p_n]\$ \$P\$. Let us construct the product of all of these \$M\$, and add 1. This new number cannot be ...

 
@Ausername I know. I saw that you said that
I also fixed it
 
I know. I was messing around with making the SCCs higher-priority but it was erroring when I attempted to close a loop
Btw I fixed the docs up.
#21 hadn't actually been resolved.
 
 
1 hour later…
10:51 AM
Hey.
I have the upload button now.
 
Hoi
 
And I also have 15 unread inbox messages.
 
Lol
 
And my internet connection is a bit sloppy, so I can't talk as often.
 
sdrawkcab etirw :CMC
 
10:56 AM
ko
 
yhw
 
?ton yhw
!nuf s'tI
 
gniht erus
yddub
 
(: sgnip sdrawkcab on :desilaer tsuj I
 
11:15 AM
uʍop ǝpᴉsdn ƃɾᴉʞlɐʇ ɯᴉ ʇslᴉɥʍ spɹɐʍʞɔɐq ƃuᴉʞlɐʇ llɐ,ʎ
I re-installed the upside down keyboard I once bought from the play store just for that message
 
@lyxal .I ma oS
 
@Ausername ɥƃnoɥʇ uʍop ǝpᴉsdn ʇ,uᴉɐ,ɥʇ
 
11:30 AM
daeh ym no gnidnats m'I esuaceb ,oN
 
11:42 AM
0
A: Sandbox for Proposed Challenges

Old ManGiven a positive number \$n\$ you have to find \$\bf another\$ positive number \$m\$(if there exist one) such that if we insert same digits in both n and m the resulting fractional value is same $$m/n = m_{transformed}/n_{transformed}$$ $$or$$ $$m*n_{transformed} = m_{transformed}*n$$ their might...

 
@cairdcoinheringaahing you forgot to include cache invalidation and naming things, so you were off by one in the wrong direction
 
I'm slowly eliminating all the (non-competings) for language creation date means./
Sorry about the frontpage.
 
?sdrawkcab gniklat pots ew did nehw ecniS
 
spoO
 
.tsol uoy fi erus toN
?segami dnes ot dedeen noitatuper eht s'tahW
 
12:11 PM
Idk, maybe 100?
 
?sdrawkcab gniklat pots I od nehW
?tey I naC
 
Whenever you want, I guess
It is kinda annoying
 
OK, I think we can turn that joke around now.
 
Sure. I'll stop and get back to golfing.
When's the best time to post sandboxed challenges?
 
Thanks.
@StackOverflow An hour from now.
 
@StackOverflow But Mondays and Wednesdays are much better than Thursdays.
 
@Ausername It sounds like there's a vacuum in the background.
 
That was for Adám, for the 'turn that around' joke.
 
@Adám Monday it is then.
 
0
Q: 'Plane' the numbers in any base!

Geza KerecsenyiInspired by this Numberphile video. What is planing? In order to 'plane' a sequence of digits, you need to: Identify the lengths of the 'runs' (adjacently repeated digits) in the sequence. Non-repeated digits count as runs of one. Decrease the length of each run by one. Runs of length 1 become ...

 
12:43 PM
I'm trying to find simple challenges to answer because I'm bored.
Any suggestions?
 
Definitely NOT the most recent one.
 
advent of code
ramps up from trivial to brutal over the course of 25 tasks, then resets in the next season.
 
I got Strunk & White!
Well I haven't yet but will when the thing updates...
 
12:59 PM
@Ausername That was pretty simple.
@StackOverflow Do they have to be on this site?
 
@Adám For C maybe not.
 
@Ausername The 10 phase 1 problems from here are quite simple.
 
They're in APL tho :P
Bai
 
gm
 
@Ausername Uh, no, the problems are not in APL.
OK, maybe problem 2 is, but it is very simple to explain.
 
1:05 PM
@user >:|
@user combinators are less readable
 
You can make a function fork f g h = \a b -> g (f a b) (h a b) if you want trains
 
@Ausername jokes on you because I've answered the first 5 in Vyxal
The other 5 I used python for because I was using v0.8.x
 
@Ausername Please don't do that again. We are aware that these exist but we've decided to not go searching for them because of flooding the front page. If you want, either space out your edits or post on meta seeing if there's interest in doing a proper cleanup weekend where we just ignore the flooding and fix everything we can by editing.
 
1:22 PM
@Ausername Actually, the English side is much funnier IMO!
Question: If I invent an encoding for Piet with a one-to-one relationship between bytes and codels, do I need to implement a full Piet interpreter for it to be usable here? Or do I just have to implement an encoding-to-PNG translator and mark answers as "Piet + my_encoding"?
 
Probably just an encoding-to-PNG translator
 
@DLosc The latter.
I already did this for APL: github.com/abrudz/SBCS
 
i'd recommend linking a piet interpreter in the repo or smth just in case there are interpretation specific things since languages are defined by their implementation. but you don't need to code a piet interpreter from scratch lol
 
That's the answer I was hoping to hear. :D There are already some interpreters out there with a lot of neat features--I can't (read: am to lazy to) compiet with that.
 
1:44 PM
This is the first time I've ever seen this and I'm excited
 
Wanna commit voter fraud so we both get there faster? ;)
 
uh
 
I'll take that as a yes. Here, you can upvote these answers of mine, including these too
 
nice and subtle :)
 
0
Q: Ambigious Chemical Formula

l4m2Losing capital of a chemical formula, it may become ambigious: co can be both CO or Co. Given an input consisting of a-z0-9, where 0 can only stand behind 0-9. Check whether it's ambigious, unambigious, or impossible. Return 3 different values for them. Shortest code win. Examples: co - ambig...

 
1:58 PM
@Wezl I'll take any opportunity I can get :P
 
I just posted a lost answer that is beating the Jelly answer at a code-golf challenge. Which is a bit of a weird feeling.
I guess Lost is a golfing lang now.
 
wow
 
neat
And I just halved my byte count for that question!
 
@pxeger outgolfed you by 1 byte :P
 
?
 
2:09 PM
bad joke (wow < neat), sorry
 
oh
@Wezl re-outgolfed!
 
Wow, pxeger just halved their bytecount! Ninja'd
 
a
halved again!
 
 
2:10 PM
 
APL, one byte, actually expresses astonishment
 
2:22 PM
Why A007764 is only computed to 27? Cuz insufficant memory?
 
2:53 PM
Probably. Have you tried running the Python code included there?
 
3:04 PM
Fun fact: you can't paste javascript: at the start of a browser URL bar
Presumably to prevent self-XSS
 
huh, that's interesting
 
@pxeger Typing javascript:alert("foo") does nothing either
 
Ugh the bots are still alive >:|
(Just checking in quickly, can't talk, o/)
 
i'll kill them if you want /s
 
@pxeger bookmarklets are blocked on new tabs too (at least for me in chrome)
 
3:26 PM
huh, I have 3 votes left. been a while since I've hit the cap
@lyxal so what is it lol
i don't think you ever told us
we got sidetracked by redwolf's kpop addiction (no context needed) (also it was totally because of this and we didn't get sidetracked like 3 times before this)
 
@lyxal Charcoal
 
3:48 PM
I have gained access
 
that's ominous
 
0
A: Sandbox for Proposed Challenges

MakonedeWhat's my Longest Win Streak? code-golf array-manipulation Challenge Given a list, string, or integer of two distinct and deterministic elements/characters/digits, one representing a win and one a loss, determine the length of the longest streak of wins. Test cases The following test cases use W...

 
@SandboxPosts Isn't this just "split at losses, lengths, maximum"?
 
well yeah
 
3:55 PM
it doesn't seem like a duplicate but it does seem extremely trivial
well correction - it feels like a duplicate but I haven't found a target
 
I don't have a problem with trivial challenges
 
I think its pretty close to the ;# challenge tbh. The only difference is that this doesn't require counting ;, but just counting all characters
 

« first day (3778 days earlier)      last day (1080 days later) »