The Nineteenth Byte

Downgoat

12:04 AM

@EriktheOutgolfer hopperapp.com

Destructible Lemon

> Your score reveals fair psychic abilities!

psychic-experiences.com/psychic-tests/zener-cards.php

i have esp 😤

Erik the Outgolfer

@Downgoat reading the FAQ... yuck, looks like a very lazy effort to make money

and, of course, "What customers say about Hopper?"

the only place where there are real reviews

Downgoat

so the return value of this function is a bool but it's has a movsx (sign ext) on the return register, this should be equivalent to a movzx right since it's a bool?

orlp

12:22 AM

@LeakyNun cs.stackexchange.com/questions/93270/…

I'm thinking of making a challenge out of this

Leaky Nun

hmm

Erik the Outgolfer

12:32 AM

@orlp ok that looks a bit impossible

if Alice gets [x, 1] and Bob gets [0, x], they both get x...ya know, you can't have a cake and eat it too

orlp

@EriktheOutgolfer no they don't

[x, 1] has size 1- x

e.g. [0.7, 1] has size 1 - 0.7 = 0.3

Leaky Nun

what is the size of a vitali set?

orlp

@LeakyNun vitali set?

Leaky Nun

nvm

@orlp so here's the story

we try to measure the size of sets

Erik the Outgolfer

@orlp isn't that true for (x, 1] and [x, 1)? I thought the size of [a, b] is b - a + 0.(0)1

Leaky Nun

12:36 AM

let's say we focus on subsets of [0,1]

so we want to build a function μ:P([0,1]) -> R, where P is powerset

orlp

@EriktheOutgolfer 0.(0)1 = 0

Leaky Nun

we list the things our functions should satisfy

1. μ(A) >= 0
2. μ([0,1]) = 1
3. μ is translation-invariant (mod 1)
4. μ is countably additive

it turns out that the 6 conditions are inconsistent (with 2 to be listed)

5. μ is defined everywhere
6. axiom of choice

orlp

what is translation invariant here?

Leaky Nun

I think Vitali pointed this out if I'm correct, and constructed an example using what is now called a Vitali set, which is constructed using 6

orlp

|[a, b]| = |[x + a, x + b]|?

Leaky Nun

12:39 AM

@orlp μ(A) = μ(x+A)

orlp

and countably additive?

Leaky Nun

not just intervals

orlp

μ(A v B) = μ(A) + μ(B) - μ(A intersect B)?

Leaky Nun

@orlp if A_1, A_2, ... is a bunch of disjoint sets, then μ(A_1) + μ(A_2) + ... = μ(A_1 union ...)

@orlp that's finitely additive

orlp

ah so 4 doesn't even apply to non-disjoint sets

Leaky Nun

12:41 AM

well that's a corollary, but there might be some convergence problem with infinite inclusion-exclusion

you can certainly derive finite inclusion-exclusion with 4

orlp

5 seems quite problematic though

Leaky Nun

1-5 are consistent though :P

orlp

being defined everywhere on the powerset of an uncountable collection is quite the task

Leaky Nun

under ZF, that is

assuming ZF itself is consistent, that is

orlp

@LeakyNun don't get me wrong, I haven't looked into this at all

Leaky Nun

12:43 AM

oh

orlp

but my intuition suspects that the axiom of choice is inconsistent with it

Leaky Nun

right, 1-6 are inconsistent

orlp

is because you need the axiom of choice to begin with to see what a mess you've made by defining this function everywhere on the powerset of the reals

Leaky Nun

right

orlp

@LeakyNun it feels a bit to me like giving a telescope to an astronomist and then concluding that "telescope + geocentric = inconsistent"

haha

Leaky Nun

12:46 AM

well

orlp

but that's just my intuition

Leaky Nun

interesting

I see that you believe in the axiom of choice

orlp

not really

I'm not against it either

but I'm just very skeptical about those... very large sets

Leaky Nun

if you don't, I might say that the telescope is flawed

I see

orlp

and when I say skeptical I just mean that my intuition is fairly worthless

I think the axiom of choice is kinda sinister

because it is intuitive when applied to finite sets

or countable sets

Leaky Nun

12:48 AM

finite is a theorem, countable isn't, but ok

of ZF

orlp

sure, but I'm just looking at it from evaluating it intuitively

like someone sees the axiom and will first try it against a finite set

"yeah makes sense"

then a countable set

"sure, I can pick one with some distribution"

Leaky Nun

I think what you mean by countable is actually dependent choice instead of countable choice

orlp

but then it's too easy to just accept it for uncountable and beyond sets too

Leaky Nun

i.e. choose the things one by one inductively

orlp

@LeakyNun can you pick a number in [0, 1]?

Leaky Nun

12:50 AM

0.5

orlp

what made you pick that number?

Leaky Nun

how do I know

psychology, I believe

it makes people choose the number in the middle

I don't believe in free will

orlp

(hint: any answer you give is inherently countable due to the language you described your picking algorithm in)

Leaky Nun

hmm?

orlp

there is no finite-time "picking algorithm" that can return an element from an uncountable set

only from a countable subset of that uncountable set

Leaky Nun

12:52 AM

sure, we're just dealing with countably many things anyway

in particular, the computable real numbers

Downgoat

@ASCII-only should VSL have enum Goat { a, b } for enum syntax or like enum Goat { case a; case b }

orlp

@LeakyNun forget computability

Leaky Nun

that isn't what axiom of choice says

you're still choosing from one set

orlp

@LeakyNun the mentionable or describable set of numbers or mathematical objects in general, is countable

Leaky Nun

this is still finite choice

orlp

12:53 AM

because any medium you use to transfer your mention or description of an object

is inherently countable

therefore the objects that you can describe are also countable

Leaky Nun

do you know Konig lemma?

orlp

not by name

Leaky Nun

locally finite tree that is infinite has an infinite path

orlp

can you describe it in a bit more words?

what is a locally finite tree?

Leaky Nun

each node has finitely many children

orlp

12:55 AM

then I don't think it's correct to say it has an infinite path

Leaky Nun

let's just deal with binary trees

orlp

it has an arbitrarily large path

Leaky Nun

you'd be surprised

if a binary tree is infinite, then there is an infinite path

orlp

that is, a path of length ω

Leaky Nun

right

orlp

12:56 AM

but that's distinct from a path of length ∞

@LeakyNun I've seen a similar thing with chess on an infinite plane

Leaky Nun

ok

now this requires dependent choice

orlp

there are configurations in which black is to move

Leaky Nun

you need to choose a child each time

orlp

and then afterwards white will mate in ω steps

that is black can delay its mate for ω steps - an arbitrarily large but finite amount

@LeakyNun ok, left child

:P

Esolanging Fruit

@DestructibleLemon I tried it

> Score: 25 out of 25
> Analysis: Your score reveals excellent psychic abilities!

Leaky Nun

12:59 AM

@orlp oh no, left child has no children

Esolanging Fruit

Just run console.log(shape[rand_number]);

HyperNeutrino

is there a CSS way of changing the style of a button when it is pressed by the user?

orlp

@LeakyNun i.imgur.com/fEHUiH9.png

black to move

by moving up the middle rook N steps he can survive for N + c moves afterwards

Esolanging Fruit

@HyperNeutrino does :active work?

Leaky Nun

@orlp I see

orlp

1:02 AM

@LeakyNun so he can survive ω steps

arbitrarily large, but finite

HyperNeutrino

:D it does. thanks!

Leaky Nun

@orlp ok 1. that isn't what ω means, ω is actually infinite 2. there is actually an infinite path, not just an arbitrarily large path

orlp

@LeakyNun alright, so left child doesn't have children

what about the right child?

Leaky Nun

it must have, or else the tree would be finite

it would only have 3 nodes...

orlp

ok, so if the node we started on has n children

we go to our right child

so we have n-1 children left

Leaky Nun

1:04 AM

our tree is infinite

orlp

it was locally finite

Leaky Nun

no, we don't have n-1 children left

orlp

so I can with certainty state that for some n the node we started on has n children, no?

Leaky Nun

each node has finite children

but the child can have a larger number of children

orlp

ohhhhhhhhhh

you mean direct children?

Leaky Nun

1:05 AM

but that doesn't matter, it suffices to discuss the case for binary trees, i.e. each node has <= 2 direct children

orlp

AKA branching factor

Leaky Nun

sorry

orlp

well duh

:P

the real numbers [0, 1] is such a set

Leaky Nun

that wouldn't be a local condition, you know

@orlp sure

orlp

write them in fractional binary

0 = left child, 1 = right child

then pi is an infinite path in this tree

Leaky Nun

1:06 AM

but that's actually the wrong set to think about it, since 0.0111.... = 0.1000....

if you include everything, you get the Cantor set

Konig lemma and Cantor set and computability are all connected

Fun(Fun(N,2),N) has decidable equality :P

Fun(N,2) is the Cantor set

orlp

what is Fun?

Leaky Nun

functions

orlp

that's still unclear :P

Fun(a, b) = ?

@LeakyNun true, forgot about that

Leaky Nun

eh... I mean the functions from N to 2

orlp

ahh

I don't see how Fun(Fun(N,2),N) has decidable equality

let alone Fun(N,2)

@LeakyNun construct some bijection N <=> polynomials with rational coefficients and arbitrary variables

Leaky Nun

1:16 AM

@orlp Fun(N,2) doesn't

@orlp arbitrary variables?

orlp

then the function f(n) -> 0/1 "does diophantine equation corresponding to n have a solution?" is undecidable

@LeakyNun x, y, z, ... and so on

what's after z?

anything you want

I see

Hilbert's tenth problem

Hilbert's tenth problem is the tenth on the list of mathematical problems that the German mathematician David Hilbert posed in 1900. It is the challenge to provide a general algorithm which, for any given Diophantine equation (a polynomial equation with integer coefficients and a finite number of unknowns) can decide whether the equation has a solution with all unknowns taking integer values. For example, the Diophantine equation 3 x 2 − 2 x y − y ...

it's just an encoding of hilber's tenth problem :P

Leaky Nun

1:17 AM

I thought you told me to do it lol

anyway, that's irrelevant to my claim

orlp

oh no it was just an argument to show Fun(N, 2) has no decidable equality

Leaky Nun

that isn't decidable equality

it's just a noncomputable function

orlp

doesn't that imply that you can't test it for equality?

given the function is a black box

Leaky Nun

indeed, let's say that our functions are computable :P

Apr 4 at 11:41, by H.PWiz

@LeakyNun Can you give the solution to your last CMC please?

cf ^

orlp

I don't really know how to interpret Fun(Fun(N,2),N)

Downgoat

1:21 AM

github.com/jamiebuilds/anti-fascist-mit-license

Leaky Nun

@orlp Fun means computable functions now

orlp

I will look at it more tomorrow

I now need to go to bed

I have an exam :D

@LeakyNun somewhat related I'd say math.stackexchange.com/questions/2469105/…

question by me :P

Leaky Nun

ok

Oliver

Does anyone know how to get the last item from the stack in Actually? tio.run/##S0wuKU3Myan8/7889/9/SwA

Downgoat

1:42 AM

btw @ASCII-only i think i have find alternative to cloud9 call codenvy

@ASCII-only pls ping when u come online

Anonymous

1:59 AM

@Oliver Stack or list? To move the last item on the stack to the top, use (. To get the last element of a list, use N.

Oliver

@Mego Sorry, I did mean list. Thanks!

Anonymous

Welcome :)

Riker

4:10 AM

@quartata I feel like your twitter gave me cancer

and/or I'm now an idiot

quartata

@Riker what tweet did it

I need this for a study I'm running

Anonymous

4:30 AM

@quartata Your flavor of misspelled surrealist humor is a treasure

6

Pavel

@quartata Has your study found anything yet

using the world,s strongest most technologically sophistcated laser pointer to carve 'kilroy was here"" onto mars

Tweeted by quartatertots on June 18, 2018 at 4:27 PM

@quartata ^ Here's what did it for me

quartata

@Pavel i need more fake impressionist paintings

@Pavel this one was a Treat

quite proud of it

@Mego thank you. i hate it

@Riker Just out of curiosity, what were you doing with the Boids?

I never asked

Downgoat

5:08 AM

anyone here good with IDA

I am using MakeUnknown but that doesn't get registered as a patch, so how do I take my modifications w/ MakeUnknown to be able to exported in a diff

Cat Wizard

6:02 AM

@quartata You have a very nice twitter.

Thank you

Stewie Griffin

7:12 AM

I don't see how closing popular questions as dupes of older inactive ones help the site...?

If you notice it early, sure, go ahead and hammer it. If it has twice the number of answers and more upvotes than the old one. Keep it, and if anything, close the other one...

14

Q: Find the Squarish Root

Write code that when given a positive number \$x\$ as input, outputs the largest positive divisor of \$x\$ less than or equal to the square root of \$x\$. In other words find the largest \$n > 0\$ such that \$\exists m\geq n:m\cdot n=x\$ (Exists \$m\$ greater than or equal to \$n\$ such that \...

I'm very tempted to undo Peter's hammer, but I don't want to do it single handedly...

Stewie Griffin

7:48 AM

The old question has not received any answers or votes after the new one was closed as a duplicate. The only result has been to kill all interest in the question.

user202729

8:12 AM

@StewieGriffin Now it does.

I mean, instead of hammer back somebody can just move their answer there, it would be easier right? ...

Without HNQ... probably a good thing.

Why is there so many mathjax-related messages starred :/

I would argue that the old question is no worse than the new one - in fact the highest voted answer there has higher score (and is more interesting, probably faster too)

Q. Did we have HNQ a few years ago?

Adám

@StewieGriffin I find the old question much easier to understand. Maybe a mod can move all the answers to the old one?

user202729

@Adám (also called "merge")

Although there is a problem. Although the core challenge is the same, the output format is different (the old one is {i, n/i}; the new one is i)

So they can't just be moved.

Can anyone help me with this? (need to read the linked question for context)

For me ctrl+shift+up gives A and ctrl+shift+down gives B. — user202729 6 mins ago

Adám

@user202729 I didn't notice. As the new one requires less, I think we should reopen.

Soaku

TypeError: 'HistoryIterator' object is not iterable

user202729

@Adám But the core challenge is the same. Also converting between formats are trivial.

Soaku

8:27 AM

hmm

Adám

@user202729 True. But as you say, it doesn't hurt the site.

user202729

Having multiple similar challenges does. We should only have one.

(in particular: not fully golfed solution without them being aware of a better one, interesting method being not known, etc.)

Solved my problem: unix.stackexchange.com/questions/138501/…

Soaku

9:46 AM

my server hosting doesn't allow access to pip, not even to --user

this is annoying as heck, you have to unpack and install everything manually.

Stewie Griffin

I agree, the best would be if people started answering that one instead. The problem is: they don't (or rather, most people don't).

Erik the Outgolfer

uh, what are you referring to

Stewie Griffin

And the HNQ is not only negative. It attracts new users to the site, which is a good thing.

@user202729 this one.

I'm on the phone, so replying, hyperlinking etc is cumbersome.

user202729

10:01 AM

@StewieGriffin Yes I know...

codegolf.meta.stackexchange.com/questions/1875/… --> "post-by-post".

The 3rd question is "which challenge is better" -- as I said above, IMO they're equally good.

Sherlock9

I was looking through the transcript and saw how Euclid is pronounced something eff-klee-dees in Greek. Now I kind of want to ask Erik about how to pronounce every Greek name I've ever heard in mythology, science, and mathematics

Erik the Outgolfer

...

Sherlock9

But I supposed I'd start with Archimedes, Socrates, Plato, and the person Indonesia calls Aristoteles

user202729

@Sherlock9 Why don't you ask Google? (yes Google can pronounce everything in every language, but a bit inaccurate in my experience)

Sherlock9

Sorry, I won't bombard you

I'm just kind of a linguistics nut about transliteration

ar-ki-mi-dis, so-kra-tis, pla-ton, a-ris-to-te-lis, for those wondering

Stewie Griffin

10:24 AM

@user202729 I too think they are equally good. But I don't see why killing a very active challenge is a good thing. Had it been done when there were 3 answers I would agree, but when there are 30 questions then i disagree.

And I too think it should be a post-by-post thing, although the comments I wrote were general...

New Main Posts

10:36 AM

0

Q: Bake me some moji

Given a byte stream which is valid UTF-8 and Windows-1252 (most languages will simply take this as a UTF-8 string), convert it from (that is, pretend it is) Windows-1252 to UTF-8. Examples: £ → Â£ Â£ → Ã‚Â£ Ã‚Â£ → Ãƒâ€šÃ‚Â£ I ♥ UTF-8 → I â™¥ UTF-8 árvíztűrő tükörfúrógép → Ã¡rvÃztÅ±rÅ‘ tÃ¼kÃ...

code-golf string conversion unicode encoding

Adám

@NewMainPosts Whoa, that was pretty much instant.

user202729

Although obviously /dev/stdin is a byte stream, do we have some "equivalent substitution"?

For example is an array of ints in 0..255 ok?

Adám

@user202729 Asking me, right?

user202729

Yes, but I think this can apply to other challenges too.

Those which requires "byte stream".

Adám

@user202729 Well, how about a string or character list?

@user202729 Good meta Q.

user202729

10:44 AM

Because a non lazy list is not exactly equivalent to a byte stream.

But streams can be buffered too.

In this particular challenge it does not require interactive so laziness doesn't matter.

Adám

@user202729 Would it be better if it said "sequence" instead of "stream"?

user202729

For example: For Jelly taking an array of bytes (an Python3 list of Python3 int) gives a huge advantages over taking a Python3 string.

Adám

@user202729 But how about a Python3 list of length-1 Python3 strings?

user202729

Equally hard.

In Jelly the only way to get bytes from string (without Python eval) is to get its Unicode codepoint then encode it into UTF8 manually.

Oops...

Jelly string is Python list of length 1 Python string by default.

Adám

@user202729 :-)

user202729

10:50 AM

Except in input. Confusing.

Adám

@user202729 That's because Jelly tries to mimic J, where there are no strings, only character lists.

user202729

That's because Jelly tries to be golfy. Having functions to manipulate strings and lists separately is not golfy.

Anyway... is an array of integers in 0..255 acceptable?

Erik the Outgolfer

@Adám nah, Jelly doesn't try to mimic anything here :P

Adám

@user202729 Consensus does not mention that possibility. The opinion that allows it is at +6/-4 so I guess not.

user202729

A valid UTF8 byte stream is not a string...

Adám

11:00 AM

@user202729 Any ideas how to phrase it?

user202729

You may specify that the input is an UTF8 string?

CMP: How do you write string literals that is too long in Python? (1): Use backslash; (2) concatenate many strings together.

Erik the Outgolfer

what is "too long"

user202729

Something that doesn't fit in the currently working window.

(I asked that because of an annoying bug in sublime text fold)

Adám

@user202729 It already says most languages will simply take this as a UTF-8 string.

user202729

sublimetext.userecho.com/communities/1/topics/…

@Adám So "if your language can take an UTF8 string then take it. Otherwise do whatever you want"?

Sounds a bit weird. And have some language bias.

Adám

11:06 AM

@user202729 Something like that.

Erik the Outgolfer

@user202729 actually, I'd recommend a combination of (1) and (2)

s = 'This ' + \
    'strin' + \
    'g is ' + \
    'long'

user202729

"Don't allow / ask for different things in languages with different capabilities"

@EriktheOutgolfer Why do you need \ then... ah... you do need that when there are no parentheses.

Adám

@user202729 I'm not. All languages may use all methods. It is probably just easier and shorter to take a UTF-8 string.

user202729

That's about my suggestion‌.

Erik the Outgolfer

@user202729 if you do something like this, the result will be ruined

s = 'This \
     strin\
     g is \
     long'

user202729

11:11 AM

@EriktheOutgolfer Yes I know, leading spaces.

Adám

@user202729 OK, but what I mean is that you may take a UTF-8 string, if you can, but feel free to use any method.

@EriktheOutgolfer We have discussed this problem regarding allowing APL strings to span multiple lines.

user202729

I mean in cases like

Erik the Outgolfer

@Adám the problem is different here; the leading whitespace on every line except the first will be included in the string

user202729

s = ('This' +
       'string' +
       'is')

I can't just do

Adám

@EriktheOutgolfer No, that is exactly the problem we encountered. I proposed:

user202729

11:13 AM

  s = 'This\
string\
is'

because sublime text folding bug.

Adám

s←'This
   string
   is long'

Erik the Outgolfer

yeah ^^^ = yuck

Soaku

12:34 PM

Python: is there await for in arrays? eg. [x await for x in r.users()]

user202729

So r.users() is a coroutine?

What about list = await r.users(); [x for x in list]?

Soaku

that throws an exception, wait

 TypeError: object ReactionIterator can't be used in 'await' expression

And yes, r.users() is a coroutine.

never mind, it had to be async, not await. problem solved

Zacharý

1:08 PM

I broke my Linux :(

Transcript for