« first day (3783 days earlier)      last day (1062 days later) » 

12:00 AM
CMC: Given an integer n > 2, output the result of [1×2×n, 3×4×n, 5×6×n, ...]. The final element will be n×n if n is odd, and (n-1)×n×n if even. Essentially, break the range to n into pieces of 2, get their products and multiply each by n
 
I think I can already see the Jelly program that resulted in this CMC
 
wait yeah no there's no cumulative ƒ
I always forget there's no cumulative ƒ
:/
 
@UnrelatedString \?
 
12:03 AM
@cairdcoinheringaahing Boring Scala answer: n=>1 to n grouped 2 map(_.product*n)
 
Ah, with starting value. Hm.
 
Can you prepend/append to pretend to have a starting value?
 
I think you have to do ⁸;<link>\
 
and then there's no exploiting the weird ƒ behavior you just ran into
 
12:05 AM
Even K (which has a very limited vocabulary) has cumulative fold with start value.
 
ƒ is one of the newest quicks tbh - it was added right before Dennis went dormant
This bug exists as well :/
 
wait what was i thinking
i lterally just forgot to make it 2-wise
yep, the classic
 
I hate that that input works
 
@cairdcoinheringaahing What should it give?
 
What does it give in APL?
 
12:08 AM
0
 
that seems like a sensible compromise
Hang on, is that special cased for +, or does ×/ also give 0?
 
compromise between erroring and what
 
@cairdcoinheringaahing Special-cased for 0. ×/ gives 1.
 
@cairdcoinheringaahing That likely gives 1, the identity
 
12:10 AM
So what does <dyad>/ do for an empty list?
 
In mathematics, an identity element, or neutral element, is a special type of element of a set with respect to a binary operation on that set, which leaves any element of the set unchanged when combined with it. This concept is used in algebraic structures such as groups and rings. The term identity element is often shortened to identity (as in the case of additive identity and multiplicative identity), when there is no possibility of confusion, but the identity implicitly depends on the binary operation it is associated with. == Definitions == Let (S, ∗) be a set S equipped with a binary operation...
 
Dyalog APL doesn't seem to let you define your own identities though
 
so is everything equipped with an associated identity/default or is it
okay yeah so it is that
nice
 
@cairdcoinheringaahing Exactly, that was what my follow-up question was going to be.
 
@UnrelatedString idk about other APLs but those probably don't either (although it would be very nice)
 
12:11 AM
I wouldn't mind it returning the empty list tbh, especially as Jelly has builtins for S and P
 
@user I know of one that does.
 
Either that or 0
 
@Adám Which one is it? (also, is Dyalog APL planning to add that sometime in the future?)
 
one of the twenty things keeping my hung up in developing perhaps is that by ripping off brachylog employing backtracking it's equipped to actually make failures useful, but it's also hard not to think about cool ways to cover corner cases :P
 
@UnrelatedString Why not have both? (with one option being opt-in)
 
12:13 AM
true
vyxal flag moment
 
@user NARS2000. Not for now. We don't like the way NARS2000 has done it.
 
I was thinking more of a quick/operator/modifier/whatever, but that works too
@Adám :( That's a pity, I always found it hard to use
 
maybe i could even have two different names
the one that fails more is Perhaps
the one with more corner cases filled in is Yeah Probably
 
lol
 
This is for Maybe right?
 
12:15 AM
Actually, it's for Perhaps (Seriously)
 
Ah, close enough :P
 
(idk what it's actually for)
 
i was going to call it Maybe if it wouldn't have made naming stuff weird considering the Maybe type in Haskell
 
Maybe you could rename Maybe (the Haskell type) Mebbe?
 
which the operator parser (read: everything so far) thoroughly abuses
true
 
12:16 AM
@UnrelatedString Abuse is good, that's the code golf way
For legal reasons, that's a joke. Kids, don't abuse anything, please.
 
:D Just got the silver [math] badge
 
Congrats!
 
Now, time to be 40% of the way towards the gold badge :P
 
@user *screams* because that's what kids do :P
 
12:19 AM
@Wezl As long as you're not doing drugs, go ahead
 
@cairdcoinheringaahing how many?
 
for legal reasons I'm a joke
6
 
@Wezl sigh. no, you're not a joke
You're Wezl
How don't you know this?
 
@Wezl I hate to ask, but what are the laws where you live?
 
Any Feed Back please?
 
12:23 AM
@lyxal well thanks for the emotional support but I'm afraid you'll have to die now
 
@Wezl we don't do that here
 
@lyxal What? Die?
 
@user Dying isn't something that can be done where I am
 
@Bubbler Using your new tag to its full extent, eh? :)
 
@lyxal 7, but 6 are invisible
 
12:24 AM
@lyxal Oh I forgot hell was eternal
:P
 
@cairdcoinheringaahing well done 👏👏👏
@user nah, it's illegal
 
@user Well, adding it to the existing challenges will beat already :P
Meanwhile, inside my brain: "hey, I have 42609 rep now! so funny! ... well, not quite so..."
Challenge idea: calculate how funny your rep is
 
@Bubbler that's funny enough for me. Nice
 
12:50 AM
the KOTH seems to be pulling in new users again, which is nice
 
1:38 AM
@Bubbler I guess your next best shot is 69420 now :P
Or we could downvote you until you get to 42069, of course ;P
 
a 500 and 50 bounty + an upvote works
 
@hyper-neutrino ?
Oh, I see, I thought you meant someone had to give Bubbler two bounties
 
1:56 AM
lol
 
except that it doesn't save any bytes this time
 
huh. nice
 
How does the second one work?
 
like the first one but ranking instead of grading down, apparently
 
yes
 
2:18 AM
Thoughts on this answer? I generally don't get involved with discussions, so it'd be nice to get some opinions
 
Looks good to me
 
i feel like if you have a filename that encodes information and is included in the score it's against the spirit of things for it not to be output but i suppose that would be exceptional behavior
the edge case you laid out is probably the ideal case to actually consider
 
3:21 AM
Looks like there's support for a Hello, world! feed.
 
CMC: Given two equal-length non-increasing lists a and b, find the largest N such that there is some i where a[i] <= b[i + N]. This is CCC 1996 S5.
this could go on main but i'm not a fan of just copying the problem over even though i'm permitted to by license. also the main interesting part of the challenge is time complexity
 
Should I go ahead add the HW feed?
 
Sounds like a good idea. No new answers have come up, and no one's downvoted HN's
 
@ROs @mods Any opposition?
 
I think you're good to go ahead with it.
 
3:24 AM
 
CMC make a 16x16 profile picture for a Hello World feed
 
You'll need to assign a user I think
It can probably just be named "Hello, world!"
 
i named it "New Hello World Answers"
that might be too long though
 
I think I prefer "Hello, world!" :p
 
should the profile picture be "HW", a globe, or what?
 
Hmm, not sure.
What does NLP's look like?
 
 
I can make a pixel art globe really quickly if y'all want
 
It's probably been made already
 
3:28 AM
We're not stealing some other person's pixel art :p
 
:(
If you're making your own, can you also make it a valid Piet Hello World program?
:P
(actually, can you just copy the standard Piet hw? as long as it's cited)
16
A: "Hello, World!"

M LPiet, 15x15(225) 13x15(195) codels I think Piet program size is counted by codels. If someone wants to know the size in bytes anyway, I could shrink it down to 495 bytes(15x15 codels) and 476 bytes(13x15 codels) using PNGGauntlet. Codel size 10 for better visibility. 13x15 version: First version...

 
I've got a pretty cool looking globe, it'll take me a little while to finish it though because I have to go do something.
We can probably wait half an hour though, not that big of a deal
 
i'm just waiting for the bot to do its typical feed spam lol
5 messages moved to ­Trash
 
3:44 AM
Okay, back
Sorry about that
Opinions on this one?
I think the ice caps needs to be made a bit more visible
 
i think the ice caps either need to be more visible or excluded entirely
a vast majority of globe clipart i find with a quick search doesn't even include the ice caps
 
I kinda like them though :p
 
so IMO if you want to include them they should be more prominent to look more deliberate and not like a mistake (especially at a small scale)
 
I'll try darkening them
If it doesn't look good I'll just omit them
 
@hyper-neutrino why the choice of picolisp recently?
Asking out of curiosity
 
3:54 AM
@lyxal did you see the new challenge about closing brackets?
 
@hyper-neutrino yep
 
it's about a picolisp feature
so it's only fitting there's a picolisp answer :P
 
Phew
 
i don't have plans to use this language ever again fwiw
if i want to use a lisp-like language i'll probably use racket
 
For a moment then I thought gpt3 had access to time travel: when I plugged in a list of the top golfing languages, it autogenerated picolisp
Despite having no way of knowing about it at all
 
3:56 AM
that doesn't imply gpt3 has time travel
it just means the hivemind has upgraded to the same model of gpt3
i cannot disclose which algorithm we use though
 
@hyper-neutrino well there's only 4 models it could be
 
right but what if we are not on GPT3
 
@hyper-neutrino you just said you are
 
I swapped the ice caps around
I think this one's better (I also made canada look a little nicer, though maybe a bit less accurate)
3
 
4:04 AM
@lyxal what i was trying to say is that if gpt3 generating picolisp had any correlation with my language choice, it is not necessarily gpt3's time travel, but is more likely that the hivemind changed
however, gpt3 generating picolisp was purely coincidental
in fact, you may want to ask Wzl, because they are the one who came up with the challenge in the first place
 
@RedwolfPrograms where's Australia?
 
Too far west to be seen
 
Sad.
 
I can try to make it work, but it wouldn't look good :p
 
To make it fair for everyone, can we use "Hello" part instead of "World" part (something like a handwave emoji)?
 
4:13 AM
Fair as in which regions are shown?
 
Yes :P
 
I don't think it's too significant, the blobs are vague enough you could probably argue what was supposed to be south america is australia :p
 
It was just a random idea, I'm fine with the globe
 
I went with the americas and europe/africa because it's much more recognizable as earth, whereas europe, asia, and australia would look more like a giant blob of green :p
 
idea: run a script to keep updating its pfp so it rotates :D
4
 
4:16 AM
That'd be really cool, but probably infeasible due to caching :p
 
pfft wdym, when has caching ever caused issues on SE, especially in chat
 
gif pfp when
 
I wonder if that would work
 
doubt
 
I'll test
 
4:17 AM
i haven't heard stories of anyone exploiting it or being annoying with it so the feature probably doesn't exist
hey there @JoKing!
 
heyo
 
If it can be done on discord, what's saying it won't work here?
 
gif pfps work on discord?
 
No, you can rotate server icons
Which probably means you can rotate pfps
 
like via scripting or smth?
 
4:18 AM
Using a bot account
 
That's horribly cursed
 
That's just the old and new ones, in a gif
 
yeah but discord doesn't cache 130% of things (number may be off due to caching)
also thanks, I hate it
 
I'll try that as a pfp now
 
4:19 AM
Oh yeah gif pfps don't work here
 
That reminds me of 2012 somehow
 
If gif pfps worked on SE y'all'd be hating me rn
 
3 mins ago, by hyper-neutrino
i haven't heard stories of anyone exploiting it or being annoying with it so the feature probably doesn't exist
 
If I replace the src in the image the gif works, does SE just strip the animation I guess?
 
@RedwolfPrograms it does
 
4:20 AM
Aw :(
 
Just uses the first frame
Source: I once upon a time tried to make my pfp a rickroll gid
*gif
If it worked, you would be seeing it rn
 
I've cursed TNB's transcript forever with jiggly africa
And americas
 
it's not too late to atone for your sins
 
And I think UK keeps appearing and disappearing
 
I think that was meant to be the scandinavian countries, idk
 
4:23 AM
oh this once again reminds me
chat still has glaring security holes due to arbitrary images and SE has ignored it for 11 years
@RedwolfPrograms I think you mentioned making a reworked chat client; am I remembering wrong / are you still working on that?
if so, you should block images except i.stack.imgur :p
 
I'm starting on it once I get the library working
I'll make sure to do that, yeah
 
Is it a userscript or something?
 
I want to test my idea to sneakily get Announcer with the chat exploit, but it'd involve annoying a bunch of people with posting a broken "image" then deleting it.
That's something SE might care a bit more about if it works
@Bubbler Probably a userscript, yeah. I'll have a terminal based one as well.
 
Cool
 
i'd be willing to let you test that in a chat room if we can organize enough users to help test it at once. can't say if SE would appreciate that though, I can ask TL or smth so we don't get in trouble later :p
 
4:28 AM
We'd need 25 users, so finding enough willing to test might be difficult
Maybe asking here and in the tavern might be enough, idk
 
oh why'd i think it was 10 :/
 
Maybe let people enter a separate chatroom
 
It'd definitely be in a separate room, yeah
 
If the offending message is recent enough in the room, everyone entering the room will trigger the link anyway
 
It's not nearly as harmful as some of the things that are possible with the exploit, but it's still probably not a good thing to test without people's explicit permission
 
4:31 AM
That's true
 
I could just get Announcer by using 25 different VPNs :P
What did you do to the Mediterranean, Redwolf?
And where's Portugal?
 
Europe is very inaccurate :p
 
No kidding :P
 
You try :p
 
Can I make it 32x32?
16x16 is a mess
 
4:35 AM
I don't want to cheat and make it higher resolution than all of the others
 
Ok, I'll try
 
another interesting concern is whether or not the ability to embed/onebox foreign images is even GDPR compliant
 
Almost certainly not
 
Resizing a big copy failed :P
 
4:45 AM
That one's nice too, although I think the colors could use a bit of tweaking
The vulnerability works on main too.
 
What, so deleting an image could break SE chat somehow?
 
This is much worse than chat. A question can get seen by thousands of people, and you can add alt-text which shows up by the broken image link and makes it seem like it's something harmless.
@lyxal No, you can get people to visit arbitrary URLs without their knowledge
 
It actually redirects them?
 
no, but it makes a request to the link
 
No, it does it in the background
 
4:53 AM
because SE doesn't force all URLs to be hosted on their own domain
 
this would be really easy to fix by just not rendering images that aren't on i.stack.imgur
 
That rickrolls the browser internally, just by visiting the page
 
Oh
I see
So they can get artificial traffic
 
i was able to steal my own IP (using a vpn for testing, since redwolf's school blocked my site) in chat
 
4:54 AM
@lyxal And worse
 
so if I pasted that image into a chat room
i now get everyone's IP
 
Or a question
Thousands of IPs if it hits HNQs and has a baity title
 
well i mean
at that point that's not really that bad anymore
the more (anonymous) IPs you get, the less data leaked
 
That's true I guess
 
honestly the problem is mainly in chat - you can paste an image in many places and try to get the intersection of IPs of users in those chat rooms and you can narrow it down like that to get a specific user's Ip
 
4:56 AM
For questions it's less about IPs, and more about things like farming Announcer-type badges, faking traffic, etc.
 
you can also probably exploit this for advertising purposes or smth like that
 
another major issue lies with SVGs. it's possible to freeze someone's browser with SVGs
 
So the image link isn't actually a link to an image?
 
Exactly
 
4:57 AM
and this is especially annoying because it's hard to edit out or delete a post when your browser just dies when you visit it
 
@RedwolfPrograms so I could be making requests to my vyxal interpreter using an image. Aha
 
yep, and if I catch anyone I think is intentionally exploiting this (it's been publicly known for a long time now anyway so I don't care about publically discussing this) i will 11 them :p
 
Q: other than Jelly, which languages have quicks/Postfix operators?
 
Wait so you could easily get Publicist throught this glitch?
 
Most likely
It's untested AFAIK
 
5:03 AM
Idea: test by including a broken image on a website guaranteed to get multiple unique views that isn't SE
E.g. An online interpreter for a certain golfing language
 
For we southern hemisphere people
 
5:21 AM
@lyxal we need a stronger definition
 
Technically everything is postfix in a stack-based language
 
i don't think they're operators tho
 
not many golflangs have "operators" in the first place
 
05AB1E seems to use prefix or just lambda blocks like vyxal (but for more things)
 
hello everyone
 
5:27 AM
for example Σ...} to sort by some key
 
hhi wasif
 
@Razetime hi!!
 
i suppose husk has quicks in the form of higher order funcs
 
@hyper-neutrino in which language ` Σ...}` sort by key?
 
well i just call functions that act on objects "functions" and functions that act on those functions "operators"
@Wasif 05AB1E
 
5:28 AM
I see
I am learning APL its very fun
 
@Razetime like +/ for reduce by addition
 
husk has everything prefix
 
i think APL has higher order functions? not sure about the details exactly
 
so that's F+ for reduce by addition from the left
 
and idk if they're prefix. i think they can be infix too
 
5:32 AM
APL is tacit, prefix and infix at the same time, isn't it?
prefix -> monadic, infix -> dyadic
 
eh yeah
 
well tacit and what-fix aren't really the same category so it makes sense for that to be the case
 
@hyper-neutrino the distinct between functions and operators is important in APL
 
these categories would be good to talk about in the blog
 
5:33 AM
functions cannot take other functions as args
 
Just wait until I show y'all my new language
 
operators however, can do that
 
operators have to take functions, right?
(and yeah, functions can't take functions because they'll just chain/train/w/e?)
 
yeah exactly(also functions aren't first class)
 
you know what's even more cursed? my language which puts functions into a stack to evaluate them only once the object stack has accumulated enough items so you can manipulate it to look like any -fix (hence the name anyfix)
 
5:35 AM
this diagram is very helpful
 
interesting
 
@hyper-neutrino ah yes that was confusing
 
@Razetime nice
 
@lyxal it's the only thing perhaps has so far
granted only the unary operators are postfix
but all of them are at an intermediate stage of parsing
 
5:39 AM
someone forked proton
weird
i can't figure out how seriously works but i think it does something like vyxal where functions are objects and just go on the stack themselves
seems to have a lot of overloading too
 
I guess most high-level stack-based langs work like that
 
it would make some sense. operators feel golfier though
putting functions onto the stack is only useful in two cases
 
Which would y'all say is more useful, filter or reduce?
 
one, to call them multiple times. in that case, having separate lines be a link saves bytes and makes it easier
two, to use higher order functions. in that case, operators are golfier usually
 
first class functions are always good for practical languages and usually not so good for golf languages
 
5:43 AM
@RedwolfPrograms i'd say reduce, but why?
 
Upcoming language with restrictive number of operations
 
yeah what metric of usefulness is this
 
(filter can be implemented via reduce assuming other sane things exist but not vice versa)
 
reduce is far more general
^
 
Usually a block is followed by a higher-level function that uses it, so many golflangs have special tokens to save a leading (or trailing) block delimiter
 
5:44 AM
yeah
and then that does give you the freedom to do stuff more dynamically should the occasion arise as well
 
@hyper-neutrino you mean the seriously language?
it even has the same function creation commands
 
@Razetime yes
 
I agree with reduce being more general. Church list is a right fold, and you can do cons, head, tail, map, filter on it
 
vyxal honestly seems a bit closer to seriously than 05ab1e IMO, but i'm not actually too familiar with stack-based languages and their implementation differences
 
which somewhat looks more like Jelly quick than APL operator
 
5:49 AM
@Bubbler :/
that's not a :/ because it's ugly it's a :/ because i thought it was my ugly
 
@hyper-neutrino it's mostly taken from keg with all the readability parts
 
oh yeah reminds me is anyone ever going to make a keg legacy fork
because it feels like it's been maybe three separate languages over its history before vyxal happened
 
well if anyone used keg they might make a fork
 
i am back
does anyone still use Keg/Seriously
 
i'm just mildly annoyed that the shitty quine i wrote doesn't work anymore
(and moreso that the golf to it doesn't either)
 
5:59 AM
@Wasif i highly doubt it
I used keg probably twice
 
how many seriously answers even exist anymore
 

« first day (3783 days earlier)      last day (1062 days later) »