oh maybe i done goof then

i.sstatic.net/3HzIa.png `

what is this monstrosity

please shrink to URL to save my eyes

k

thanks :P

anyway, there is jelly dennis. not as good as it probably would be if I owned photoshop

pls don't ban me

@DestructibleLemon looks more like sunburnt dennis to me :P

Also, long time no see, everyone.

@Avery Haven't seen you around in a while. Hi!

@Avery I put layers in!

@HyperNeutrino heya, how are you?

@Avery I'm alright, you?

Doing fine :)

That's good to hear :)

Just as I was asked "Who is Avery?" Coincidence?

@Adám You were? IRL or on chat?

@HyperNeutrino Here, by ZacharyT. When I mistyped "every" in my CMC.

Also, if I get my rep back after content remapping, I wonder if anyone will find it strange that my rep jumped to over 6000 from 500...

@Adám Ah, I see. Hm. Coincedence? I THINK NOT

Conspiracy Theory: Stating someone's name will summon them to TNB if they are on a computer and on the internet.

I don't think I've seriously dropped here since... january, except for one flag, I think (which I shouldn't've followed to here)

that's not much of a conspiracy theory

@Avery No wonder I haven't seen many of your posts around; I thought you were a regular PPCG user originally. :P

anyway gtg now. cy'all later! ಠ/

I really am pretty terrible at CG

also, later to you, later to me.

@HyperNeutrino Coincedence: The act of giving up small change.

@betseg tf2?

@Adám this is funny. please receive your allotted star

it is time for self promotion again

Could I get some looks at this sandbox post? thanks in advance

@DestructibleLemon Please

Actually stop

No one is going to do your KOTH

;_; this is why I am sad

Please be sad somewhere where it doesn't involve spam

;_;

>_>

@Phoenix I noticed earlier that you mentioned Monstercat

It is awesome

What's your favorite genre on Monstercat

brb round just started

r6s

?

So someguy is still loading on apparantly mcdonalds wifi

(rainbow 6 siege)

all right

@Mendeleev Anything by pegboard nerds

@Phoenix doesn't everyone love them

I think they had a tour in Seattle on 3/30/17

What should be returned from a string indexing function if the queried character is not in the string?

@ConorO'Brien make your language 1-indexing and return 0

@ConorO'Brien nil or None

(just like Jelly)

@ConorO'Brien 1+the index of the last char.

@ConorO'Brien append a z to the string and return the length.

@ConorO'Brien undefined, or equivalent.

@Adám Thats length

@Adám that's already there

@Adám that makes it equal to the length

two people got ninja'd

@LeakyNun if the language is 1-based, it makes it the length + 1

Return the string, "Could not be found"

@ConorO'Brien The reason is that then you can append a fall-back element and use the returned index to get the fallback.

@ATaco a fan of stringly typing?

Better yet make index return a container with all the indexes, and return the empty container when it is not found.

@Adám what

suppose the language is strictly typed

@Mendeleev Wrong ping.

ok

@Mego hmm... that's the second person who's suggested this. Now I need to do some research lol

"Hello".index("z") -> ()
"Hello".index("H") -> (0)
"Hello".index("l") -> (2,3)

Write the string to STDERR, then terminate the program.

Crash and delete the offending program

so it can't happen again

pretend I'm making a sane language and not vigil >.>

@Mego So interesting. There has always raged a 0/1 war in the APL community. J decided on 0, but Jelly is 1, and now you say this. I also think 1 is the best, as then you can use the equivalent negative indices to pick from the back. And 0 can be assigned special meaning.

@Adám equivalent negative indices? why not use ~ or the apl equivalent?

@DestructibleLemon That is ~?

62 (b): pop a: push 0 if a==0 else 1; pop "a" or [a]: push 0 if len(a)==0 else 1; pop f: push 0 if len(f)==0 else 1
63 (c): pop a: push character at ordinal a%256; pop [a],b: push [a].count(b); pop "a","b": push "a".count("b"); pop f, [a]: push the number of elements in [a] where f puts a truthy value on top of the stack
75 (u): pop a: push a+1
A1 (í): pop [a],b: push [a].index(b) (0-based, -1 if not found)

ok then

Pass the string to an attached callback function, return the value of that callback function.

return a random integer between 0 and the length of the string.

"String".index("j", (badstring)=>"Ono")

Yield, wait for other threads to process, then try again.

I've always heard 0-based makes arithmetic computations cleaner

Making bad index functions is a lot more fun than making a good one.

Remove the character from your codepage and throw an error.

@WheatWizard stop

Too much fun. Bad suggestions are the only thing I'm good at.

@Adám can't APL specify its indexing?

@ConorO'Brien Most dialects, yes.

@ConorO'Brien For low-level languages, sure. The community yet has to produce a low-level language though.

^ CMC

1-based or 0-based?

@Dennis by community do you mean PPCG ?

@ConorO'Brien Despite a number of very bad suggestions I think returning a container is a good idea, this allows the program writer to specify what they want to happen when the index is not present. If they are sure that there will be z they can just take the first value in the container. You would probably want to implement such a function eventually.

@Dennis brain-flak?

@LeakyNun not really

not yet

@ConorO'Brien jsoftware.com/papers/indexorigin.htm

@WheatWizard Right now, I'm concerned with writing the framework in C, and am currently occupied with developing a string datatype. In my progress, I have written a string_index_of_char function, and am presently wondering what to return. In this case, I am limited to a number, and am left with the decision to retain a size_t and use the length-based approach, or switch to a ssize_t (signed) and use -1. I think I will implement such a function in the final product, however.

@LeakyNun It's a Turing tarpit, but it doesn't exactly compile to byte code.

@Dennis sesos?

@Mego right, which would seem to make -1 unfeasible. I suppose I could have it return the length internally, but when writing the language itself, make the transition from length to -1

function safeIndexedString(str){
	return {value: str,
	index: function(s, callback){
		var n = this.value.indexOf(s);
		if(n==-1){
			this.watch("value", (name,old,nw)=>{
				var newN = nw.indexOf(s);
				if(newN!=-1){
					callback(newN);
					this.unwatch("value");
				}
			})
		}else{
			callback(n);
		}
	}};
}

var str = safeIndexedString("Hello, world!");
str.index("l", console.log) // "2"
str.index("t", console.log) // nothing...
str.value = "test"			// "0"

@ConorO'Brien Iverson even suggested an operator (adverb) which would select origin per use. ⍳:0 and ⍳:1. Dyalog added ⍠ (as : is already used) as a way to get variant functionality, but we've not added the option to primitives yet.

@LeakyNun Looks like assembly, sure, but it's interpreted by Python, which is already rather high-level.

@Dennis oh...

That's an object that lets you index it if it doesn't contain the subindex, but will wait kindly until it can find it for it to run its callback.

Brain-Flak can already be compiled into byte code either via C or C++

Soon Haskell will be added, if I ever figure out this problem I've been working on, and 1000000000 is working on a compiler that goes straight to byte-code without an intermediate.

@Mego My current implementation lies as this; I somewhat hoped to avoid unnecessary complexity, but your struct idea appeals to me...

@Adám thank you for the resource, it's helped considerably

You can't index strings by strings? only chars?

Fair enough.

.com

I'm just gonna prove Sad-Flak is TC by interpreting or compiling. What language should I use? it would be easiest to use a minsky machine, but I could do something else

Try Brainflak.

@ATaco hmmm. should I interpret or compile that?

@ATaco remember, silence is okay. if you want to talk, feel free to bring up some discussion fodder

also fun fact: because my thing doesn't have a length of stack thing that makes things annoying somewhat

@DestructibleLemon compilation is always easier IMO

yeah it is isn't it?

If you can transpile Sad-flak to Brainflak, you know it's TC.

@ATaco wrong

Is brainflak not TC?

that would prove brain flak is a higher or equal class as Sad-Flak

not that Sad-Flak is TC

Oh, I got it backwards. Transpile Brainflak to Sad-flak.

yeah

Do you have a github?

yes I do

github.com/Destructible-Watermelon/turtl-d

@DestructibleLemon check out something like mini-flak

@ConorO'Brien it's mostly annoying because of arbitrary input

@DestructibleLemon why is your interpreter sorta golfed

no it isn't

wait a minute that is a different repo

@totallyhuman it is badly written, I'll give you that

@Mego stahp bullying me

@DestructibleLemon i was just saying you omitted some spaces in some lines and not in others

wait, I just realised I have a comment that says something is broken in my new langs interpreter

O_O

@Mego fair enough.com

ok, for comparison, see this:

github.com/Destructible-Watermelon

it has comments of why I use seemingly unpythonic things

@EriktheOutgolfer Sorry they're really really recent

it is a much better interpreter

@DestructibleLemon see your GH page?

I am off-point with my links

github.com/Destructible-Watermelon/sad-flak

These are my favourite comments on RProgN's source

I think I will make a BF interpreter for fun even though it is probably going to be incredibly difficult

the main difficulty is accessing commands and numbers at the same time

this will be much hard

actually I'll prove TCness first just in case :P

should I compile minsky machine or brainfuck?

how to format a TC Proof paper?

People were talking about 1- vs. 0-indexing earlier?

Reason to use 1-indexing and return 0 if the substring is not found: that way, the returned index for when there is no match is falsy. (I think this is what Dennis said when I asked him why Jelly uses 1-indexing, if I remember correctly)

@Riker m.imgur.com/gallery/R3eyi

also in emulating bf it could be annoying to emulate input, to impossible for interactive input. so I guess I use a slightly modified version

You don't need io to be TC, just get []+-<> working.

How are BF interpreters hard?

in sad-flak

Ohhhh

currently I'm doing a compiling proof

Ooh, sounds interesting.

(or transpiling but who cares)

¯\_(ツ)_/¯ I wouldn't know the difference :P

they mean the same thing but sometimes people think compiling only means machine code, or they just prefer less ambiguity?

CMC: given a positive integer, output that many dashes. e.g. 4 -> ----

Python: lambda x:'-'*x

@LeakyNun Àé-

Jelly, 3 bytes: ”-x

@HyperNeutrino '-'.__mul__

oh right

I could do this in sad-flak with a semi-large bytecount

I forgot that operators are functions are objects

or woefully with a mega byte count

nice

@DestructibleLemon that is very sad :( [no pun intended]

should I do it?

@DJMcMayhem What language is that?

@DestructibleLemon Depends. Do you have a bunch of time to devote to doing it? :P

it isn't as much of a tarpit as woefully and this is a simple task

Ah, okay. idk, depends on whether or not you want to :P

@LeakyNun ”-ṁ works too :P

and the same thing as my first solution but with a dot on top of the x

@HyperNeutrino V of course

@DestructibleLemon what is sad-flak?

Oh right, it's DJMcMayhem, of course it's V. :P

this is what I have so far:

<>({((()()()())<>())})

oh something was borked

UserScript idea: make it constantly send updates of the contents of the message box to a server so even if you don't send a message, it will be seen.

@LeakyNun do we need to handle 0?

I just found yet another way in which review is completely nonsensical on PPCG, although at least this time it seems to be intentional?

@HyperNeutrino Hehehe

:P

I only noticed that looking through the review histories (which is something i do from time to time), and saw something that didn't make sense

@DestructibleLemon oh, that would add a byte to mine if we do

@DestructibleLemon refer to the third word of my CMC (cc @DJMcMayhem)

I get the feeling that Stack Exchange is optimized for it being 100% obvious and objective whether posts should be closed, and the only reason they ask for multiple votes is because of people not even putting in the minimal effort needed to review them correctly; whereas the actual situation is that many posts are borderline or reasonably arguable and if we had a working voting system on them, it'd help solve the disputes

@DestructibleLemon unless you French.

What does that have to do with it?

In France, "positif" means >=0

Interesting

Iirc, fatalize and tux are French

@Mego I'm not at all sure any of our review queues work at all, except for suggested edits and possibly first posts

@DJMcMayhem indeed.

even suggested edits doesn't really work, but it works better on PPCG than it does on, say, meta SE

actually I think I'll go back to my original assumption, which is that nothing on SE actually fills any approximation to its intended purpose

I also feel like it would be more affective to make reputation and review not exactly correlated; my idea was perhaps have a system where people with more "helpful" reviews (that is, reviews that turn out the way they reviewed it; for example, I voted to Leave Open and it never got closed) get higher priority/weighting. Kind of like how some people have like 10 flags per day and some people have 100.

Also LQA is bad for PPCG because a lot of answers get kicked into LQA because it's too short, whereas that's HQA for us :P

well, the review queues mostly only give exposure to actions you could do anyway

e.g. I can vote to close either inside or outside the post

although if I see a post getting close votes and I agree, I'll typically try to find it in the review queue anyway just for the badge progress

> just for the badge progress

that's the reason behind me doing half the things I could do one out of two ways :P

it comes to the same thing either way, but one way you get rewarded

if you vote to close a post directly on the post, and it's in review

Rewards are nice :P

you're just denying people the chance to get badge progress for no benefit to anyone

so it's a fairly antisocial thing to do

That's an interesting way to think about review on a "forum" site. :P

hey guys, first sad-flak answer to anything!

<>({((()()()()())<>[()])})≤()≥
(({}))(({()}[()]))<>≤()[[]]≥
<>≤≥

64 bytes in my "codepage"

would you guys like an explanation?

I would like a GitHub link; an explanation would be nice too.

@HyperNeutrino ok, let me write it up then

unfortunately I can't use newlines for fancy formatting

okay

hm, this looks slightly confusing

but interesting

also plus 3 for ascii output

so 67 bytes

@HyperNeutrino that was the aim

it is like a more confusing version of brain flak, but golfier woefully

Is there a specific reason it's called "Sad" Flak?

IIRC it was based off the idea of writing another language like woefully called "Sadly"

<move input to the off stack>(this evaluates to 45, by multiplying and stuff{((()()()()())<>[()])})≤(jump to next line)≥
(({pop and push to stack twice}))(({(pull the input from the off stack)}[(subtract 1)])push twice)<push one of them to the off stack>≤jump by 1-(is the input number still not 0)[[]]≥
<push one dash to the off stack, removing it from output>≤halt≥

@HyperNeutrino because woeful is sad

so ATaco is kind of right

Ah. I see.

CMC: Given n, produce the identity matrix of n ** n.

wait let me test that still works

0/10 dupe of main

@ATaco vtc as dupe

@LeakyNun vtc your vtc as a dupe as a dupe of my vtc as a dupe

yeah it does

hey no fair editing it!

@HyperNeutrino vtc your vtc as a dupe as offtopic

Still a dupe

@DestructibleLemon I was going to reply with another vtc but that would get vtc'd as spam. Or I'd get kicked, either way I don't like the outcome :P

:(

CMC: Take input.

what do you think of sad-flak? I think it was pretty good at tying some of the graph together

@ATaco vtc unclear/too broad

do we need to halt once it is finished

wait, are people implementing PPCG-style esolangs simply to make the graph of influences more connected?

@ATaco Realize that under standard I/O formats, I'm allowed to make a function that accepts input as a parameter. Therefore, I think I can do Python, 0 bytes.

@ais523 no, I made it, not just implementing it

@HyperNeutrino no you can't

specifying and implementing, then

There are stack-based, queue-based, tacit, and memory-taped based languages. Is there a tree-based language yet? (If not, dibs)

... I might have been...

@DestructibleLemon aw :(

grr, I hate it when I say something which isn't precisely what I mean

but the idea of a concise woefully was on my mind for a while

@HyperNeutrino Lisp

and this was a pretty nice crossover I think

@ais523 do you mean "something that isn't"? :P (you probably meant to say what you said but I just felt like saying this <----- anyway)

@HyperNeutrino hev

@HyperNeutrino treehugger

@ais523 oh :( well rip not original idea then

@HyperNeutrino BRB, writing a binary tree language.

ಠ_ಠ

How about a heap based language?

> binary

wait no

i'm being stupid

heaps are trees

just special trees

Are there Christmas Tree based languages?

you could use the heap's operations as bases for the language, rather than its internal implementation

@ATaco it was binary

treehugger uses a binary tree

I'd imagine you'd end up with a language which used a (precise version of) sleep as both control flow and data storage

Hm. Interesting.

I've been meaning to implement that for a while just so that nontrivial programs become impossible to run without timing out

it seems like a fun way to exploit corner cases in PPCG rules

@DestructibleLemon That shall not stop me.

ok

> exploit corner cases

lol ಠ_ಠ :P

How about a map-based language?

Oh wait that's every language with variables

if you can access the variables by name, yes

I remember being very impressed at some anagolf solutions in Perl

which used the entire space of variable names as a hash table

to save a few bytes

like, if they wanted to store something opposite the key foo, they'd assign to the variable $foo

Interesting.

The way my tree based language will work will be using tree based code, not just memory.

almost all languages have tree based code

with the exception of a few esolangs and a few concatenative languages

root {
    node {
        child {
            etc...
        }
    }
}

the initialism "AST" stands for "abstract syntax tree", and it's one of the most fundamental notions in compilers

Also things ultimately become trees so like 2 + 3 * 4 becomes + (2, * (3, 4) ) (I think)

yes, that's exactly how 99% of compilers work internally

Binary tree syntax.

@ATaco ah, okay. I remember this from one of my classmate's presentations in grade 8 :P

| ?- A = +(2,*(3,4)).
A = 2+3*4

What would be a non-gimmicky way to have an esolang in a binary search tree structure

^ an actual interaction I just had with GNU Prolog; Prolog stores unevaluated arithmetic expressions as parse trees

@ais523 :o

I mean, it would be gimmicky

Make a language where the only memory is the source code file

Because it's an esolang

but still

@LeakyNun it's actually stored as +(2,*(3,4)) internally, it just gets printed as 2+3*4 because that's the way most people would write it

@HyperNeutrino that's about half of esolangs.org/wiki/Category:Self-modifying

Essentially prefix notation

@ais523 oh right. but they have other memory too, not only the source file

there are several languages which use prefix notation

Pyth is the best-known one here

@HyperNeutrino not all of them (although many of them do)

ah ok

wait you said half

ಠ_ಠ

the language I invented recently for this answer is a good example of a language where the only data storage is the source file

and I did that because Jelly's bad at storing two mutable values at once

so I didn't want a separate program (implying a mutable instruction pointer) and data storage

Ooh interesting

how do I make sure my python code doesn't error out with many recursions? aka how to make recursion deep

import sys #(if you haven't already)
sys.setrecursionlimit()

but what do I put in parens?

how big should it be? a million?

a really large number

IIRC it's safe to make it way too large, it won't hurt anything

making it larger than the amount of addressable memory your computer has is clearly going to be high enough

9 ^ 9 ^ 9 ^ 9 should do you.

you don't want it to be so large it takes time to calculate :-)

wtf I broke it somehow just by making some numbers bigger

as in I made the input to the program 1000 instead of 100

like wtf