The Nineteenth Byte

orlp

17:00

counting from zero

is 0 <= i < n

from 1 is 1 <= i <= n

AdmBorkBork

Yay, fenceposting.

betseg

then why it doesnt fail then there is something in RO data

orlp

@betseg undefined behavior includes 'never fail unless something else happens'

in fact it might only fail once you put it inside a billion dollar space shuttle

one of the reasons C and C++ are so notoriously unsafe

Adám

@quartata Our 2016 user meeting had a half-day workshop on neural networks, and half a day on compiling what they'd made in the first half. Still I find a few promising repos on GitHub…

Cows quack

@Adám I remember watching a video on dyalog.tv about someone transpiling APL to Futhark and then running Michael Nielsen's recognising handwritten digits example in his book

Pavel

17:02

@orlp has that happened

orlp

I actually really like the name differentiable programming instead of neural networks

betseg

@betseg when* why are t and w so close on the keyboard /s

Pavel

Space shuttle death by undefined behavior

betseg

temperature underflow?

Adám

@Cowsquack Yup. One of the repos I found right now.

flawr

17:03

@LuisMendo thanks, I think I passed with @AnderBiguri's triangle stuff :)

orlp

@Pavel death I don't think so

but I'm sure something has gone wrong at some point due to undefined behavior

like a mars lander was lost to a unit conversion error

one team was using metric the other imperial

flawr

@Mr.Xcoder Cogito ergo sum(x,1)

Mr. Xcoder

Lol 10/10

HyperNeutrino

CMP: Does this syntax idea sound like a good idea for Proton? a ~ b ~ c ~ ... ~ d == a(b, c, ..., d)

flawr

@HyperNeutrino what does either of those do?

NieDzejkob

17:13

@HyperNeutrino why?

@flawr the second one is a function call

HyperNeutrino

@flawr well a ~ b ~ c ~ ... ~ d is the proposed syntax and a(b, c, ..., d) is a function call.

Mr. Xcoder

@HyperNeutrino Hmm, separating the function the same way as the elements seems quite odd, but It migh tbe helpful... But isn’t that subtraction?!?

HyperNeutrino

@NieDzejkob not sure, just a random idea

Pavel

@HyperNeutrino I'd rather see a @ b ~ c ~ d

HyperNeutrino

hm

so you propose that b ~ c ~ d is (b, c, d)

Pavel

17:14

@Mr.Xcoder those are tildes, not minus signs

Mr. Xcoder

Oh ~, not -

Then yeah

HyperNeutrino

@Mr.Xcoder yeah. your font might render them similarly :P

I actually would like l ~ f ~ r to be f(l, r) though.

Mr. Xcoder

blame-mobile

HyperNeutrino

lol

Mr. Xcoder

Basically like Mathematica

HyperNeutrino

17:15

that's where I drew the inspiration :P

though that causes infinite recursion which I could fix using a repeat sequence, but I only want 3

quartata

meanwhile just sitting casually over here on pytek's corpse stuffed with emeralds

HyperNeutrino

wait repeat takes args

@quartata uh wat

Pavel

Yeah, I don't think your first proposal is really needed anywhere

HyperNeutrino

pytek as in elendia's language?

@Pavel yeah. The second one can help reduce the number of brackets for chained function calls (as well as @ and |>) but the first proposal is kind of useless and unintuitive :P

quartata

yes, said language where we basically did all of this design

ngn

17:16

isn't that exactly how lisp's s-expressions work?

(the first proposal)

(f x y) is an s-expr
f[x y] is an m-expr

Mr. Xcoder

Who cares about Lisp :P

ngn

x f y is apl :)

quartata

@ngn yes but I think they're more interested in syntax

HyperNeutrino

what's with the emeralds though

quartata

are emeralds not valuable

HyperNeutrino

17:18

how's this: a ~ b ~ c ~ d ~ e == d(b(a, c), e)

and so on

@quartata ok yes but is there a reference i don't get or like why is it a corpse stuffed with emeralds

Mr. Xcoder

Q: Proposals’s for what the syntax character $ should do in Physica.

Dennis

Variable names.

HyperNeutrino

Is Physica meant for golfing or practical

Mr. Xcoder

@HyperNeutrino Practical

HyperNeutrino

ok

$"varname" == varname

and $a = globals()[a] essentially

Mr. Xcoder

17:19

Wait... Lol I completely, completely misread Dennis’ message

HyperNeutrino

or make $ a valid identifier that doesn't form a word with other identifiers

for$in range... :P

flawr

@HyperNeutrino So this means that ~ is right-associative, right?

HyperNeutrino

@flawr no, left associative.

Pavel

@HyperNeutrino Use $ like Perl or Ruby's $_

label a -- goto a

What's going on with TIO?

Cows quack

17:21

@Mr.Xcoder for the website, it would be nicer if the href attribute was given to the div in the navigation, rather than the text

HyperNeutrino

oh god the mathjax

Mr. Xcoder

@HyperNeutrino I like the idea, but it would need to be ${a}.

HyperNeutrino

hm ok

Pavel

@labela--gotoa appears to be on fire?

label a -- goto a

I can't connect to it .-.

Mr. Xcoder

17:22

@HyperNeutrino Too useless :P I am going to transpile '_' to ' _ '

Pavel

@Dennis panic!

HyperNeutrino

@Mr.Xcoder Oh ok :P

Dennis

I can either panic or fix it

4

HyperNeutrino

rip __magic__ functions

@Dennis well it appears to be fine now :P

Dennis

@labela--gotoa Should be fixed.

Pavel

17:23

It's back!

HyperNeutrino

yay \o/ thanks

Pavel

Ninja'd

label a -- goto a

\\o/

oh, no md escapes needed here .-.

Dennis

Some hiccups with the secondary server. The main one is down for maintenance.

label a -- goto a

ah

wish fast fix

Mr. Xcoder

17:24

@Cowsquack Good idea!

HyperNeutrino

a#b == a[b], should a@b#c == a(b[c]) or a(b)[c]?

Mr. Xcoder

@HyperNeutrino While updating your Pyth answer, I think you overlooked my comment on your Python one

HyperNeutrino

Should @ take precedence over # or should they be on the same level LTR/RTL?

@Mr.Xcoder oh indeed i did .-. thx

Mr. Xcoder

IMO @ should take precedence over #

;-; I have to update 5 pages of HTML

betseg

hastebin.com/mibeyumuco.lisp is this a good idea

flawr

17:27

@HyperNeutrino ah right, that wouldn't make a lot of sense

NieDzejkob

@betseg I don't think it's a good idea to make it a macro

why not a function?

betseg

to avoid the unavoidable *&&**&***&

HyperNeutrino

How should Proton do @ and #?

Pavel

@HyperNeutrino a(b[c])

HyperNeutrino

I personally favor LTR-same-level

Mr. Xcoder

17:29

I chose LTR... ninja'd kinda

DJMcMayhem

@Dennis Will you tag this status-completed?

HyperNeutrino

@betseg "avoid the unavoidable" [...?]

Pavel

@HyperNeutrino Do you support 2d arrays? (a[x, y] rather than a[x][y])

flawr

curry everything

HyperNeutrino

CMP: a[x, y] == [a[x], a[y]] or a[x, y] == a[x][y]?

Dennis

17:30

@DJMcMayhem status-completed

flawr

@HyperNeutrino x]a[y

DJMcMayhem

ty

Pavel

@HyperNeutrino The latter.

betseg

@HyperNeutrino B

HyperNeutrino

@flawr ...uh please...

Dennis

17:31

I wonder why it got closed as off-topic though. Since when are bug reports off topic on meta?

HyperNeutrino

@Dennis There's a close-option for "no longer relevant"

DJMcMayhem

No idea. I'm voting to reopen

@HyperNeutrino I don't see it

Mr. Xcoder

@HyperNeutrino That's an "Other"

HyperNeutrino

was it custom-raised by some random person then?

Mr. Xcoder

The results are... interesting

DJMcMayhem

17:33

No, it's just "off-topic" with no further explanation

Dennis

Maybe.

HyperNeutrino

@DJMcMayhem no the reason is not included in the desc

Dennis

@HyperNeutrino It was. The comment was auto-deleted when the post was closed.

HyperNeutrino

oh ok

well it's reopened now lol

Dennis

.oO(How is that funny?)

Pavel

17:34

@HyperNeutrino You know what I just realized? If you index into a 2d array with a[x,y], then a[x][y] is basically the same thing but curried.

Mr. Xcoder

@HyperNeutrino The latter, and I'll try to implement it in Physica too I think :)

HyperNeutrino

@Pavel lol yes :P

@Dennis sorry, "lol" doesn't mean i'm laughing for me. :P

i guess i use it a bit too often ¯\_(ツ)_/¯

DJMcMayhem

@HyperNeutrino lol

:P

HyperNeutrino

hm how do I respond to that? normally I'd say "lol" but that's too uncreative. I guess this works as a response to that :P

meta-response o0

Pavel

XD?

ngn

17:36

rofl?

Dennis

@HyperNeutrino s/a bit too often/1764 times so far/

2

label a -- goto a

haha

DJMcMayhem

@Dennis o_O

Dang

label a -- goto a

"lol" rather than laugh, sounds like "lol? wtf?"

Mr. Xcoder

36 pages of lol...

label a -- goto a

17:38

lol

HyperNeutrino

i need tiebreaker

Adám

Since we're anyway so off-topic. One of the neighbours here has a car with the license plate "TIO RUN". Coincidence?

7

HyperNeutrino

oh wait I can tiebreak because i'm making the language

Mr. Xcoder

That feel when you really want to work on something but TNB swallows you like a black hole.

HyperNeutrino

@Adám what this conversation is 100% on-topic General discussion for codegolf.stackexchange.com

DJMcMayhem

17:40

@Mr.Xcoder s/your programming language/anything

ngn

@HyperNeutrino why don't you make them the same character? :)

DJMcMayhem

TNB is my number one source of being less productive than I should be

HyperNeutrino

@ngn uhhhhhhhhh

label a -- goto a

@Adám what

Mr. Xcoder

@DJMcMayhem status-completed and you need a trailing /

DJMcMayhem

17:41

Oh really?

HyperNeutrino

@Adám ask them if they're on PPCG /s; who knows, maybe they're a user of this site and you'd never know because their account doesn't have their real name! :P

DJMcMayhem

Huh. What if I just claimed it was vim rather than sed?

Mr. Xcoder

@DJMcMayhem Yes

Then ¯\_(ツ)_/¯

Pavel

@DJMcMayhem Then you need a :

DJMcMayhem

Fun fact about vim's :s command: None of the trailing slashes are needed. :s/foo/bar works, :s/foo is equivalent to :s/foo//, and :s is equivalent to :s/<last-search>//

label a -- goto a

17:43

the flag on right on the message looks like a thunderbolt for me. who did make it the logo for flag...?!

well, it looks like a flag in color, but not in black-white

Adám

Btw, our chief architect at Dyalog drives a Mazda Roadster with the license plate "DL06 APL"

Mr. Xcoder

leaves TNB such that the rate of productivity loss is lower

Adám

@Mr.Xcoder Yeah, right.

HyperNeutrino

since @ is right-associative, @# would have to be as well if they were on the same level.

Adám

@Mr.Xcoder I knew you'd come right back.

HyperNeutrino

17:46

his icon isn't in the user list rn :P

Pavel

@HyperNeutrino Consider a#b#c

label a -- goto a

I must have done this... .-.

Adám

@HyperNeutrino He did come right back though. Probably got a notification, and just had to see.

HyperNeutrino

yeah. Since @ and # have different associations, they probably have to be separate. Final decision made for @>#

label a -- goto a

how the heck is it a flag.

.-.

Pavel

17:48

It is clearly a flag, IDK man

label a -- goto a

but the gray icon, not the colored onme

Pavel

Yes

Adám

@labela--gotoa It is exactly the same outline, just a solid grey.

label a -- goto a

i need to get close to screen to say it's a flag

@Adám but it doesn't look like a flag .-.

how

Adám

@labela--gotoa Does the coloured one look like a flag to you?

label a -- goto a

17:51

yes

aaargh

nvm

Adám

@labela--gotoa Which language is your user name in?

@ThomasWard The -- doesn't look BASIC to me.

label a -- goto a

@Adám Actually, a polyglot if you replace -- with a semicolon, because I can't put a ; in my nick

ngn

does basic have "label"?

Adám

@ngn Yeah, at least modern versions do.

label a -- goto a

Maybe I should make an own esoteric language to make my nick work in it

Adám

17:54

@labela--gotoa APL, 4 bytes: a:→a

ngn

@Adám ah, ok. last time I used it was on an 8-bit computer

Pavel

@Adám I did not know APL had goto

Adám

@ngn Last time I did was on 16-bit, and it did have labels.

ngn

@Pavel "branch"

Adám

@Pavel In the old days, that was the only flow control. Today, you can also spell it :GoTo if that makes you happier.

Pavel

17:56

@Adám "modern"

VB has labels, but not that syntax.

Mr. Xcoder

@HyperNeutrino Implemented. integer = 4, inp = Input@, Print[${inp}] now prints 4 when the input is "integer"

HyperNeutrino

Ooh yay :D

Mr. Xcoder

And I only added 2 lines of code :D

Pavel

That's an excellent idea: foo@ == foo() when whatever comes after @ isn't a value.

Adám

@Pavel Right. QBasic used LABEL:GOTO LABEL

Mr. Xcoder

17:58

@Pavel Thanks :)

Note that it only acts niladically for now. (You cannot do function@value yet) :(

Transcript for