The Nineteenth Byte

6:00 PM

P

:P

We've just been taught what a function is, what a graph of a function means, linear functions and order 1 functions, so not much.

Separately, I have studied derivatives (now studying all their rules), a tiny bit of integration, restrictions and properties of functions

Manish Kundu

The worst part in calculus is when they ask you to find the limits but without using L Hopitals

Mr. Xcoder

The worst part in calculus is when they ask you to approximate definite integrals using Riemann sums :p

ETHproductions

Ah yes, going through that right now myself

Mr. Xcoder

@ETHproductions Oh god good luck!

ETHproductions

6:04 PM

So far it's really just been "and so we approximate like so, and when we take the limit as n->inf we end up with this integral"

I hope there's no quizzes where we actually have to calculate the approximate Riemann sum itself :P

Mr. Xcoder

I find integration quite hard compared to differentiation

DJMcMayhem

Same

Mr. Xcoder

^^^ Good lord, that would be so imprecise :P

DJMcMayhem

Differentiation is waaaay easier

Mr. Xcoder

I really need a professor to get into integration properly, can't do it by myself. I guess I'll wait for 4 more years until the 12th grade :P

ETHproductions

6:06 PM

Yeah, chain rule >>> substitution rule

Mr. Xcoder

Agreed.

ETHproductions

The method my prof used to find the antiderivate of sec(x) was so out-of-the-way I would have never found it

Mr. Xcoder

Although some mathematicians think differentiation is mechanical

Neil

@DJMcMayhem interestingly the differential of tan is slightly less undefined than tan is, since it only diverges to +Infinity whereas tan diverges to either +Infinity or -Infinity depending on which direction you approach the discontinuity

Mr. Xcoder

@ETHproductions Lol, I've just encountered that earlier

ETHproductions

6:10 PM

@Mr.Xcoder I agree, to differentiate you can just apply a certain set of rules. With the substitution rule you have to actively find a way to break up the function into smaller pieces

hmm, are there functions that don't have antiderivatives?

DJMcMayhem

@Neil Huh. That's fascinating

CMC: What PPCG challenge has the most other challenges closed as a duplicate of it?

ETHproductions

The original KC challenge

Mr. Xcoder

Oh misunderstood

Rickroll definitely

@ETHproductions There are some that don't have elementary antiderivatives.

(The only answer is amazing, not that I understand all of it)

ETHproductions

wow, that'll be quite a read

there are 10 dupes of that challenge

Mr. Xcoder

Just 10 o.o

ETHproductions

6:16 PM

non-deleted anyway

J. Sallé

@Mr.Xcoder the top comment on the answer is the best

> Raymond, I don't believe you've given enough detail.

Mr. Xcoder

And it's a community wiki! ಠ_ಠ

Erik the Outgolfer

wow that's like 13.6 pages or something

totallyhuman

hi

Mr. Xcoder

\o

Conor O'Brien

6:29 PM

CMC write a function that returns itself

Erik the Outgolfer

@ConorO'Brien Python 2, 10 bytes

f=lambda:f

DJMcMayhem

Dang

I had def f():global f;return f

Erik the Outgolfer

... what?

DJMcMayhem

I didn't think about lambdas lol

Erik the Outgolfer

dang markdown you have disappointed me

totallyhuman

6:32 PM

@ConorO'Brien Haskell, 3 bytes: f=f

hehe

Anonymous

@totallyhuman Haskell (Lambdabot), 3 bytes: fix

Anonymous

Probably cheating :P

Mr. Xcoder

@ConorO'Brien JS (untested): f=()=>f

NieDzejkob

@Mr.Xcoder -1 byte: f=a=>f

totallyhuman

@Mr.Xcoder f=_=>f

Mr. Xcoder

6:35 PM

@NieDzejkob Is that allowed?

totallyhuman

ninja'd 'cause lag >_>

@Mr.Xcoder javascript doesn't care about what arguments are passed

Mr. Xcoder

Ok then

Anonymous

In JS, arguments are more like suggestions

totallyhuman

also that'd be valid anyways 'cause unused input is allowed

Mr. Xcoder

Solution: Don’t use JS.

Potato44

6:36 PM

@totallyhuman I don't think that counts, I believe that is an infinite loop

totallyhuman

yes but it does return itself :P

it returns itself which returns itself which returns itself...

ETHproductions

@ConorO'Brien Japt, 3 bytes: =@U

J. Sallé

@ConorO'Brien APL (Dyalog Unicode), 1 byte +

totallyhuman

@J.Sallé huh, then Haskell, 2 bytes: id

ETHproductions

I think that would allow the 4-byte JS solution eval, or atob, btoa, etc.

or am I misunderstanding how that works?

Anonymous

6:41 PM

@ETHproductions How does eval return itself?

totallyhuman

no as eval() doesn't return eval

ETHproductions

calling id in Haskell returns id? That's interesting

totallyhuman

uh well

not quite

J. Sallé

@ETHproductions Dyalog APL returns whatever function you type if it doesn't have arguments, as I see it at least. Probably @Adám can explain better.

AdmBorkBork

@ConorO'Brien PowerShell, 7 bytes -- $a={$a}

ETHproductions

6:43 PM

@J.Sallé oh, I see the difference. Seems kinda cheaty though

totallyhuman

it's the way functional languages work :P

MD XF

@ConorO'Brien C, 11 bytes: x;f(){x=f;} Try it online!

J. Sallé

@ETHproductions I assume it is kinda cheaty, I just answered because it was a CMC :p

Anonymous

@MDXF -1: f(x){x=f;}

MD XF

:o

Conor O'Brien

7:02 PM

@Mego so long as calling fix with no arguments returns fix it should be fine

@MDXF well, if the function really returns itself, then f()() == f, but this is not the case, as your function returns a memory location

Anonymous

Technically calling any function with no args in Haskell returns itself :P

Anonymous

So ap would be shorter in Lambdabot Haskell :P

H.PWiz

Or just id

Conor O'Brien

the thing is, the function has to be called :P so afaict, since APL doesn't evaluate its functions when presented with no arguments, it does not count.

@Mego looking into it, functions in haskell can't be called without arguments, right?

RosLuP

In C would be char*f(){return f;}

In Apl one function that return itself would be {∇} but for me not compile...

Anonymous

7:10 PM

@ConorO'Brien Sure they can. It's entirely possible to define nullary functions: a=1.

RosLuP

In Axiom f(x)==(...;f)

Conor O'Brien

@Mego ok, they can be defined, how might you call one tho?

Anonymous

@ConorO'Brien Just like any other function?

Anonymous

tio.run/##y0gszk7Nyfn/P9HWkCs3MTPPtqAoM69EIfH/fwA

Conor O'Brien

I don't know haskell, but I'm interested in testing the fix solution. how might I verify it?

Mr. Xcoder

7:14 PM

Totally unrelated, but dimensional analysis is a life-saver!

Anonymous

@ConorO'Brien I realized after I posted it that it doesn't work unless you call it with itself as an argument :P

Conor O'Brien

oh lol

Anonymous

Unless you look at it from the perspective of "every function returns itself with 0 arguments"

Conor O'Brien

hm, this calls for another cmc

Anonymous

In which case there are shorter solutions, like id or 1

Conor O'Brien

7:16 PM

CMC: write a function that returns any other function that isn't itself

that looks good I think?

Anonymous

Well it depends on how you define "other function"

Anonymous

With f=id, f and id are functionally identical, but have different names

Conor O'Brien

well, f() != f && f is a function

DJMcMayhem

@ConorO'Brien Python, 15 bytes: lambda:lambda:1

Anonymous

@DJMcMayhem Why not just lambda:id?

DJMcMayhem

7:21 PM

Because I have a nasty golf habit where I implement the first thing I can think of and then post it without thinking of shorter alternatives

Anonymous

me too thanks

Conor O'Brien

nonetheless the lamba lambda is kinda cool :>

Anonymous

Actually: ⌠`1⌡

Anonymous

(which is just lambda:lambda:1)

DJMcMayhem

@Mego Technically you could even do something like str.find

I'm not sure what the shortest class.method combo is though

Anonymous

7:23 PM

I should make Actually's parsing better so that can be written as ``1

AdmBorkBork

lambda lambda red pajambda

MD XF

7:41 PM

@ConorO'Brien wrong

f()() is just not valid syntax

Adám

@ConorO'Brien Non-cheating APL (Dyalog Unicode), 9 bytes Try It Online!:

r←f x
r←f

It takes a dummy argument, actually runs, and actually returns itself (which can then be applied to some other (dummy) argument.

@RosLuP Dfns cannot return functions as a result.

MD XF

@ConorO'Brien or to more clearly disprove that

I had to refer to this that I wrote when I was 13 (and put on GitHub years later) to write that

Adám

@ConorO'Brien Sorry, this is an example run.

Anonymous

@MDXF You didn't free that memory. Shame on you.

Adám

@ConorO'Brien APL (Dyalog Unicode), 9 bytes Try It Online! (non-cheating, takes dummy arg, returns plus)

MD XF

7:47 PM

@Mego You're referring to basics.c? No memory is allocated

...except for foo

Anonymous

Exactly

MD XF

Fixed

Anonymous

@MDXF I think with enough parens in the right places (plus a cast), you can make f()() work, but I don't want to try right now

MD XF

@Mego I'm pretty sure it's completely impossible but I'd love to see it if you do figure it out

Conor O'Brien

@MDXF oh, right, I was thinking you couldn't cast functions for some reason

MD XF

8:03 PM

Anyone know why this happens?

mudkip201

@ConorO'Brien Pyt 5 bytes: 2⁶Đ⁺ṕ

MD XF

@mudkip201 just curious, do you know about Cubically?

Anonymous

@MDXF Change line 47 to char *tmp = malloc(strlen(s) + strlen(lastbrace.since) + 1);

mudkip201

@MDXF haven't really looked into it

MD XF

@mudkip201 just wondering because your first few dozen codepage values are the same as Cubically's

mudkip201

8:16 PM

Huh. Strange

but completely coincidental

MD XF

Interesting

(no accusations here I'm just legitimately interested)

@Mego ლ(ಠ益ಠ)ლ

Transcript for