The Nineteenth Byte

hyper-neutrino

14:01

hm. i'm considering making my own language (like actually this time - i scrapped yuno because it was too annoying), and I'm probably going to use same/similar tacit structure to jelly. should I add a 2,2,2 rule above the topmost Jelly dyadic rule to do the special case always?

Adám

@cairdcoinheringaahing Essentially, you want APL's trains (=chains)

Anush

hi all

user

@Adám Jelly's chains seem a lot more general than APL's trains (especially since it doesn't seem to have "normal" dyadic function application)

Anush

bpa.st/WT5A is my attempt to make a benchmark for sorting 1 million ints

Wezl

:O I found a 5-byte HQ9+ quine

Anush

14:04

you have to type 1000000 when you run it

can anyone make it faster?

StackMeter

@cairdcoinheringaahing oh

I am somewhat understanding Jelly

hyper-neutrino

IMO Jelly chains are golfier and APL trains are more intuitive. But I'm not a good judge of this because I know pretty little about APL and I am decent at golfing in Jelly but not APL, and I already intutively get (most of) Jelly chains

Adám

@user What do you mean by "normal" dyadic function application?

StackMeter

I understand APL more than Jelly

user

@Adám Like (some expression that isn't a constant and requires parentheses) <dyadic function> (another expression)

You can't directly do that in Jelly, but that opens up the syntax for more chaining rules

Adám

14:07

@user Oh, "it" goes on Jelly.

user

Actually, looks like you can, I'm an idiot

@Adám Pronouns are annoying

Adám

Don't go there!

StackMeter

so you can't do (+)*(+)? AFAICT, that gives (a + b)^2

or a^2 + b^2

hyper-neutrino

or APL

user

@StackMeter Pretty sure you can do that in both Jelly and APL (without parentheses (APL doesn't require parentheses, and I don't think they mean the same thing in Jelly))

StackMeter

14:09

@hyper-neutrino Jelly

hyper-neutrino

first of all * is exponentiation lol

user

^ × for multiplication in APL

hyper-neutrino

at the beginning of a dyadic chain, +×+ will become the special 2,2,2-chain (which is really just a special-case dyadic start and then a regular 2,2-chain), which does λ = (α + ω) × (α + ω)

otherwise, depends on what's before it. assuming the code before it is already completed, this is a 2,2-chain and then a lone dyad, which then depends on what's after it

Adám

So far the same in APL and Jelly.

hyper-neutrino

APL is RTL right?

Adám

14:11

Yes.

hyper-neutrino

cool

assuming the code before it won't chain with this dyad and there is either no code after this or it starts with a monad

Adám

I don't understand why Jelly would take +,_2+ as (+,)(_2+) then.

Why isn't it become (+,_)2+ ?

Though I don't understand what 2 does there.

hyper-neutrino

+×+ is a 2,2-chain into a lone dyad. the first does λ = λ + (α × ω) and the second does λ = λ + ω, so overall the new value is (λ + (α × ω)) + ω

@Adám 2,2,2 is a special case only at the beginning of a dyadic chain

so at the beginning, it is (+,_)(2+)

otherwise, it is (+,)(_2)(+)

Adám

Oh. Yikes.

user

@hyper-neutrino Does that help with golfing?

Adám

14:14

APL's trains are definitely more regular then.

hyper-neutrino

it's... a bit strange. but there is likely a reason behind this, probably just golfiness. i trust dennis made these decisions consciously and had a good reason for each design choice

user

How long did it take him to design Jelly?

Adám

Yeah, because J is identical to APL in the respect.

hyper-neutrino

@user i can't recall the last time I used it, so idk

user

Looks like it was under heavy development for ~2 years

Adám

14:16

@user 4 Dec 2015–17 Jan 2019 according to GH.

hyper-neutrino

the core of it probably took about 2 years, and new built-ins and features were continually added until the last commit, probably

user

@Adám Is there a way to find that without going through all the commits page by page?

Adám

@user I did it manually, but you might be able to find it in api.github.com/repos/DennisMitchell/jellylanguage/commits

caird coinheringaahing

@Adám One example would be +1+,_2+ vs +,_2+. The first is parsed as (+1)(+,)(_2)(+), whereas the second is (+,_)(2+). You can force the first to be (+1)(+,_)(2+) with an extra byte: +1+,_ɗ2+

I believe that this is generally golfier - before the introduction of ɗ it would've had to be +1ð+,_ð2+, and ð...ð is not a common occurrence in Jelly answers

Sandbox Posts

0

A: Sandbox for Proposed Challenges

How much faster than qsort can you achieve? Task Sort 1 million 32 bit integers as quickly as possible. Score I will run your code and compare the timings to my default C code using qsort. That is: #include <stdio.h> #include <stdlib.h> #include <time.h> #include <stdint.h> int cmpfunc (const v...

StackMeter

14:34

asked the necessary question of "Are builtins banned?"

caird coinheringaahing

@cairdcoinheringaahing Wait no, that's not equivalent :/

hyper-neutrino

@cairdcoinheringaahing so would you say it's a good idea to implement 2,2,2 everywhere or is dennis's choice better?

i don't use either often enough to really know what to think. i tend to use +A} as a 2,2 a lot more than a proper 2,2

caird coinheringaahing

@hyper-neutrino Personally, I prefer 2,2,2 everywhere as its more consistent, but I can't speak to the golfiness

@hyper-neutrino The fun one is +A}, being parsed as a 2,2,2 at the start :/

StackMeter

@hyper-neutrino consistency is everything

Anush

I posted my qsort sandbox question. I hope it turns out to be fun!

hyper-neutrino

14:37

@cairdcoinheringaahing oh yeah :/ rip

tbh might make a built-in in my language that just converts a dyad-monad into a x + F(y) combo just to make it a bit more understandable

if i can afford the byte

caird coinheringaahing

Yeah, if there was a builtin like $ which forced that parsing, that'd be cool

Anush

thanks @cairdcoinheringaahing

caird coinheringaahing

@Anush Thank you for giving me the chance to yet again object to builtin bans :P

Anush

@cairdcoinheringaahing and to change my opinion :)

now.. which language will have a much faster builtin sort??

hyper-neutrino

this is just "fastest sort", i don't think the comparison to another algorithm really makes a difference?

Anush

14:40

@hyper-neutrino what do you mean to another algorithm?

hyper-neutrino

does comparing the timing to your qsort actually make a difference to scoring?

i'm not suggesting you edit it out - just wondering

caird coinheringaahing

@hyper-neutrino It provides a "baseline" contender: "beat this or don't post"

Anush

@hyper-neutrino no it just makes it more fun

hyper-neutrino

@cairdcoinheringaahing ah, good point

Anush

@cairdcoinheringaahing and that too

hyper-neutrino

14:41

so no bogosort :P

caird coinheringaahing

Damn, there goes my plans :/

Anush

I think it's called ego stroking :) People like to be faster than builtins I think

hyper-neutrino

i'm surprised we don't have a challenge that's just "fastest-code sort a list" ¯\_(ツ)_/¯

Anush

@hyper-neutrino it really matters what's in the list

amazingly, it's still not obvious what is the fastest sorting algorithm and when

caird coinheringaahing

hyper-neutrino

14:43

TBH i feel like sort in Python is going to be... pretty competitive

unless the list is not very well generated, timsort is pretty good afaict and it'll take a lot of knowledge of theory to be able to make something that will beat it in python

Anush

@hyper-neutrino timsort is very good if the array is almost sorted already. But in my task it isn't

caird coinheringaahing

@Anush Have you tried benchmarks with builtins in languages (e.g. Python's sorted)?

Anush

@cairdcoinheringaahing no just C's qsort but I am fairly sure timsort will be worse

caird coinheringaahing

I think you should probably try sorted and maybe the builtin sorts for Java or other mainstream languages and see how they compare. If it's very close, the challenge may not be that interesting

Anush

maybe I should add some motivation? qsort is slower than it should be because the comparison function can't be inlined

timsort will be slow because the array is not close to being sorted

because it is 32 bit ints it is likely some variant of radix sort will be fast

and quadsort does well in benchmarks

it shouldn't be a competition just between builtins

my guess is that something that takes advantage of the small size of the ints will win

caird coinheringaahing

14:49

@Anush I agree. My concern is that might happen, depending on how good the builtins are. Therefore, I recommend you get some idea of how good they are before posting the challenge

Anush

@cairdcoinheringaahing could you help me do that ?

I only know C and python

ngn

@Anush yes, 1000000<256^3, so radixsort should take only 3 passes processing one byte each

caird coinheringaahing

@Anush That's one more (practical) language than I know, unfortunately

I'd suggest messing around on TIO's bash to create a basic script that generates a list of numbers and times each submission. The script here would be a good starting point

Anush

@ngn yes. I think that's a very good idea

@cairdcoinheringaahing I added the C and python timings

@ngn there is inplace radix sort which I guess would be even faster

AncientSwordRage

does anyone here know how TIO works?

caird coinheringaahing

15:05

I have some understanding of it, why?

Adám

@AncientSwordRage In what way?

Anush

@ngn I guess even 32 bit ints only need 4 passes. Maybe I should increase the size to 32 bits

AncientSwordRage

My company is now blocking it from running as it downloads a file....wondering if that's a recent change

caird coinheringaahing

TIO hasn't changed in over 18 months

AncientSwordRage

ah so it's my company

@cairdcoinheringaahing I did hear something about Denis...

caird coinheringaahing

15:06

in talk.tryitonline.net, Feb 3 '20 at 16:39, by Dennis

Hey all. I'm sorry for disappearing on you and for taking bad care of TIO lately. I'll try to explain what's been happening, although it's not easy for me to talk about this...

hyper-neutrino

is there any way for you to see the name of the file being blocked?

caird coinheringaahing

We have had a few minor bugs (e.g. Mathematica's certificate expiring, code refusing to run etc.) that seem to have been fixed after we reported it, and Dennis appears to be "checking" talk.tryitonline (e.g. he was last seen 20 seconds ago), which suggests that he might be handling the absolute essentials to keep TIO going. Beyond that tho, nothing's changed

AncientSwordRage

@hyper-neutrino yes, I think it's a string of numbers

caird coinheringaahing

I've dropped a link to your message in talk.tryitonline, in case its something that does need to be checked out

hyper-neutrino

hm. strange. can you permalink what exactly you're trying to run? i doubt i can actually figure out the issue but i can give it a look

AncientSwordRage

15:13

I've dropped details in the talk.tio chat

hyper-neutrino

thanks!

AncientSwordRage

@cairdcoinheringaahing A loved one has gone through something extremely similar: Fibro -> Depression -> Anxiety -> Spiral

(also whats minigolf in the context of this chat room?)

caird coinheringaahing

@AncientSwordRage TNB or talk.TIO?

hyper-neutrino

@AncientSwordRage Ah, sorry to hear about that :(

caird coinheringaahing

7

Q: Introducing Biweekly Mini Golf!

In order to grow our community and its chat room, The Nineteenth Byte, we decided here to start a new monthly biweekly chat event: Biweekly Mini Golf! During this event, a number of previously drafted "CMCs" (short for Chat Mini Challenge) will be posted, and anyone can reply with their solutions...

hyper-neutrino

15:15

@AncientSwordRage if you're talking about Biweekly Mini Golf, it's an event where we basically just dump a crap ton of CMCs (chat mini challenges). so an organized event where we just solve a bunch of comparatively easy challenges

or read ^^ lol

AncientSwordRage

I'm down for that

user

@Adám Thanks, that's helpful

hyper-neutrino

it's just a casual golfing event where we can post simpler challenges and more notably, unlike on main site, CMCs don't need to be that well specified, as long as there's a pretty reasonable interpretation to agree on

caird coinheringaahing

@AncientSwordRage I can;t speak from experience, but from Dennis said in his messages, it sounds like an extremely tough thing to deal with, and I for one am 100% supportive of Dennis taking as much time as he needs to get better

AncientSwordRage

@cairdcoinheringaahing same here

hyper-neutrino

15:19

things like the mathematica license being fixed and critical issues being fixed indicate that dennis is at least well enough to keep TIO in a functional state which makes me glad to see - of course I hope he eventually feels well enough to return to updating TIO or even return to CGCC itself but as long as he's alright i'm happy

user

pxeger and Bubbler are working on their own TIO clones, which'll take pressure off of Dennis

AncientSwordRage

@hyper-neutrino exactly

Anush

@ngn do you think 3 passes would be slower than 1? 2^20 is not very big

ngn

@Anush i think 3>1, yes :)

Anush

@ngn :) I meant in terms of practical performance because of caches etc

I assume you chose 3 passes for a reason

count sort has poor data locality

ngn

15:32

@Anush i said "3 passes" because you limited your numbers to 10^6, and thats between 256^2 and 256^3

Anush

@ngn yes. But can we instead do a pass using 20 bits instead?

ngn

@Anush you don't know until you test :)

Anush

@ngn all true :)

I will leave the question over night but I think I might increase it to 32 bits. 10^6 is a bit small

or maybe 31 bits because of RAND_MAX

CMC find the minimum of (32/r) *(10^6 + 2^r) for r > 1

caird coinheringaahing

@Anush Approx 2118361.135

hyper-neutrino

ah, i'm too lazy to do calculus right now. i trust caird's right.

pxeger

15:44

I trust caird's ability to use a calculator lmao

caird coinheringaahing

I just stuck log ((32/r) *(10^6 + 2^r)) into desmos, it shows the min point :P

Anush

I was hoping for code :)

user

Desmos is a valid language here :P

hyper-neutrino

@cairdcoinheringaahing oh smart. i put the original in and couldn't find the line

caird coinheringaahing

It's at 10^6 + a bunch :P

Wezl

15:46

Jun 4 at 13:28, by Wezl

17 hours ago, by Wezl

an y feed back?

caird coinheringaahing

@Anush Jelly, 10 bytes

hyper-neutrino

how does this wo- oh wait, you hard-coded the number, didn't you ಠ_ಠ

Anush

oh.. no hardcoding please!

I want the code to solve the problem!

pxeger

@Anush Python: print(2118361.135): Try It Online!

Anush

@pxeger noooo :)

pxeger

15:48

Kappa

user

@Wezl I think you should add DJ's last comment to your question - it makes the task a lot clearer

caird coinheringaahing

Wolfram Alpha is so helpful here :P

@Anush It involves either differentiation, which most languages won't do, or manually finding it, which is kinda boring. But, I'm already bored, so I'll write a Jelly finder :P

Anush

@cairdcoinheringaahing surely a binary search would do it

Wezl

@user I agree, but I'm not sure how to fit it in. I'll find out

pxeger

@Anush assuming it's differentiable and continuous

caird coinheringaahing

15:50

It is and is (aside from x = 0, obviously)

Anush

@cairdcoinheringaahing yes and yes

I said r > 1

pxeger

(well yes, and you'd struggle to find the minimum anyway if it wasn't)

Anush

true

it also has a unique minimum

hyper-neutrino

moreover it has just one critical point so it shouldn't be hard

i'm just too bored to do it right now

Anush

code should not be hard to do this

@hyper-neutrino you could solve my sorting problem instead :)

AncientSwordRage

15:52

My favourite sorting method is sleep sort

6

pxeger

...sleep repeatedly until enough cosmic rays have hit the computer that the array became sorted?

caird coinheringaahing

@Anush Jelly, 15 bytes

Anush

@cairdcoinheringaahing what is it doing?

AncientSwordRage

@pxeger e.g. sort the number 1, 32, 6, 864000, you just sleep for that number of (milli)seconds, and out put the number each time

pxeger

ah I see

AncientSwordRage

15:55

after 1 unit of time, output 1. After 6 milliseconds you output 6 etc...

Wezl

@user thanks, I think I'll post now

caird coinheringaahing

@Anush It generates the range from 1.001 to 20 in intervals of 0.001, runs the helper link over each (which is defined as (32/r) *(10^6 + 2^r)), then calculates the minimum

Redwolf Programs

@hyper-neutrino Can you 11 the pinned message about BMG to say 24 hours away instead?

The edit 11 not the delete one :p

Anush

@cairdcoinheringaahing cool. I still think binary search would be better

Redwolf Programs

Thanks!

caird coinheringaahing

15:56

@Anush Probably, but binary search is Jelly won't be as nice

Anush

:( jelly needs upgrading then

hyper-neutrino

you can keep narrowing down a range so your result gets more and more precise as you keep running it

but like

that sounds annoying

caird coinheringaahing

@Anush If you're using Jelly for production code, you're an idiot :P

4

Yes, that includes me :p

Here's a version which takes an upper bound x on the left and a precision value p on the right (as is, calculates the range to p decimal places with a step of 10^-p )

Anush

@cairdcoinheringaahing :)

Wezl

@Wezl lol, SE says my last "I'm Lazy" question is a "similar question"

New Posts

16:00

0

Q: I'm Lazy: Close my Superparens

Picolisp has a feature called "super parentheses": Brackets ('[' and ']') can be used as super parentheses. A closing bracket will match the innermost opening bracket, or all currently open parentheses. src Taking a string containing only the characters []() (and newline if your language requir...

code-golf string parsing balanced-string lisp

caird coinheringaahing

@NewPosts One of those tags is Jelly friendly, the other 4 are the worst tags for Jelly answers :P

Wezl

great, because I always get a jelly answer first :P

hyper-neutrino

lol, syntax highlighting doesn't like elif"..."

it turns the f"..." into an fstring

pxeger

I'd really love to know why it's apparently so hard to get syntax highlighting right

user

Cheapskates using regex

caird coinheringaahing

16:16

@pxeger The most common engines use brainfuck, so its difficult to fix any bugs

Wezl

status-citation-needed

pxeger

@user but so many languages don't have regular grammars for tokenisation. You'd have to be kinda stupid to use regex for that

user

@Wezl Source: It f***s with your brain

@pxeger You're on to something there :)

112

Q: "Hello, World!" in zero lines of code

NPM's sloc is a moderately popular tool for counting source lines of code in a file. The tool will attempt to strip out both single and multiline comments and count the remaining lines in order to get an estimate of the 'true' number of lines of code. However, the parser used is based on regular ...

code-golf restricted-source hello-world

Not quite the same thing, but I'm sure syntax highlighters also use these sorts of techniques

[citation-needed] though

caird coinheringaahing

@pxeger My guess is that there are lots of different languages that people expect syntax highlighting for (e.g. Python, JS, C, Java, C++, C#, Ruby, Haskell, APL and more all have highlighting on SE). You could create standard tokenisation and highlighting for each, but most use rough guesstimates with additional specificities for each language as it's less payload-heavy

user

^

caird coinheringaahing

16:24

For example, IDLE gets Python syntax highlighting correct 100% of the time, because it only has to deal with Python. Things like VS or Notepad++ can have massive libraries of highlighting rules per language installed as part of their software because they won't be that big. But, you try and do that on a webpage, and it could take seconds to load which people don't like

hyper-neutrino

@Wezl do you know how to picolisp?

Wezl

somewhat

hyper-neutrino

do you know why this doesn't terminate?

Wezl

sorry, can't TIO

hyper-neutrino

oh my example is unnecessarily complex anyway

(prin 1)

straight up does not terminate

Wezl

16:35

you're missing (bye). picolisp is pretty weird.

hyper-neutrino

lol wat

ok thanks

Wezl

it's made to be used a lot from the REPL, so standalone programs need (bye) or (i think) (quit) works

hyper-neutrino

ah

caird coinheringaahing

So, most syntax highlighting on SE is tag-based (if a question on SO has the python tag, it uses Python syntax highlighting). We do have language-specific tags which are generally only used on questions that require answers to be in that language. Tag-based syntax highlighting is currently disabled for the site, but would make sense in this case. So:

CMM: Tag based language syntax highlighting, yay or nay?

hyper-neutrino

how often do we get language-specific tags that don't actually restrict submission language, compared to tips or ones that require that language be used

like for example, a challenge about a python feature / detail that can be implemented in any language

pxeger

16:44

and you can always do:

```python
some code
```

hyper-neutrino

yeah

TIO also automatically adds a language header if you are using a highlightable language i believe

pxeger

or

<!-- language-all: python -->
    some code

caird coinheringaahing

bash, c, c#, c++, css, haskell, java, javascript, json, lisp, lua, matlab, perl, php, python, r, ruby, scala, sql, swift are our tags that have syntax highlighting available

Yeah, I'm aware that there are other ways to get syntax highlighting, this is just whether it would be worth enabling it for language tags

Wezl

@hyper-neutrino you can use jdoodle.com/ia/eqH to test out how picolisp parses. Note that it prints () as NIL

caird coinheringaahing

@hyper-neutrino The most recent challenge is tagged lisp, but isn't restricted to lisp answers

hyper-neutrino

16:46

my personal opinion is no, we shouldn't have it

is this already a meta post? i've been through too many lately to remember anymore

caird coinheringaahing

@pxeger If you add that to the question, does it affect the answers?

pxeger

I'd assume not

caird coinheringaahing

@hyper-neutrino No, I'll turn it into a meta discussion depending on this CMM

pxeger

If only we had a sandbox.SE to test it :þ

caird coinheringaahing

@pxeger That's the main difference between tag-based highlighting and HTML-based highlighting. IIRC using specific highlighting in an answer overrides the tag-based one tho

Adám

16:49

@cairdcoinheringaahing Nope. Would mess up every challenge that relates to a language.

hyper-neutrino

how does IO work in picolisp

(read) seems to read the next line of source code and inserting input into source code is disallowed

and also really weird

maybe i need to use a function :(

caird coinheringaahing

Are you writing a picolisp answer for the new challenge? :P

hyper-neutrino

... maybe

Wezl

@hyper-neutrino yes, but (read) reads the next form not line, (the code (+ 1 2) can have newlines separating the sub-forms +, 1, and 2)

@hyper-neutrino I'd definitely suggest this. it's much smaller

hyper-neutrino

maybe i should use a functional language functionally. :p

Peilonrayz

16:57

@cairdcoinheringaahing You've missed a couple, like coffee. bsh c clj coffee cpp cs csh css dart erlang golang hs html kotlin java js latex lisp lua markdown matlab ml objectivec pascal perl php proto py r rb regex rust scala swift sh sql vb vhdl xml yaml

caird coinheringaahing

@Peilonrayz Yes, but we don't have tags for some of those (e.g. pascal)

Wezl

@Wezl e.g. '(a(do-stuff-with a] versus (prin(do-stuff-with (read](bye) (not sure if (bye) is necessary)

hyper-neutrino

is there a way to do default arguments?

or should i just make a helper

Peilonrayz

@cairdcoinheringaahing Ah, ok

Wezl

@hyper-neutrino software-lab.de/doc/refD.html#default or do [or var (set'var val]

btw (set'var val) is shorter than (setq var val)

hyper-neutrino

17:02

ah. thanks

why setq lol

Wezl

tradition?

hyper-neutrino

i feel like it should be set! but. ¯\_(ツ)_/¯

Wezl

picolisp is not a very functional language. Scheme is the odd one out there.

hyper-neutrino

ah

the only lisp-like language i've used is racket btw :P

Wezl

e.g. picolisp has dynamic scope, no easy closures, no tail recursion, etc.

user

17:04

Bleh

hyper-neutrino

no tail recursion?

tf

pxeger

no closures? Even C++ has closures!

Wezl

first class functions are literally lists in picolips, so you can make them yourself

user

I can live without closures, but dynamic scope and no tail recursion sound very annoying

pxeger

Depending on what you mean by dynamic scope, I think I could live that that more than I could closures

And I'm interpreting "no tail recursion" as "no tail-call optimisation", not literally "tail recursion is impossible"

Wezl

17:08

the default language is pretty meh, but tbf it has f-exprs and stuff which makes it pretty extensible. For example it has a built-in prolog DSL (pilog)

user

@pxeger If that's the case, it might not be that bad

hyper-neutrino

that's what i figured it meant as well

oh well. python doesn't have it either so i'm used to it

caird coinheringaahing

Does Jelly have TCO? I can never remember

pxeger

I doubt it, being written in Python

Wezl

@cairdcoinheringaahing most likely not
:/

hyper-neutrino

17:14

i think so

ß doesn't terminate

Ȯß successfully outputs all 128 kilobytes of 0s before getting killed

pxeger

oh, huh

caird coinheringaahing

It does use TCO - according to Dennis

hyper-neutrino

1ß€ segfaults immediately, so it's definitely some optimization

Wezl

:O

hyper-neutrino

yeah so it seems that ending a link with just a "call other function" will TCO

Ç}` will segfault but that's to be expected

caird coinheringaahing

17:17

Isn't }` a no-op?

hyper-neutrino

yes

caird coinheringaahing

I know that {` is, as ` reuses the left argument, but I'm not sure about }`

hyper-neutrino

oh wait maybe it isn't?

F{ is lambda x, y: F(x) so F{` is lambda x: (lambda x, y: F(x))(x, x) which is lambda x: F(x) which is F

F} is lambda x, y: F(y) so F}` is lambda x: (lambda x, y: F(y))(x, x) which is lambda x: F(x) so still F

yeah both are no-ops

caird coinheringaahing

However, `} isn't a no-op, but `{ is

hyper-neutrino

don't think so

caird coinheringaahing

17:22

`} causes the dyad to be reuse the right argument

hyper-neutrino

f` is lambda x: f(x, x) so f`{ is lambda x, y: (lambda x: f(x, x))(x) which is lambda x, y: f(x, x) which isn't equivalent to f anymore

in a monadic chain alone then sure

pxeger

@hyper-neutrino Is this because of a CPython stack overflow? (and also because Jelly does sys.setrecursionlimit(1 << 30)?)

hyper-neutrino

but notice here that +`{ gives left + left and +`} gives right + right

pxeger

Does anyone else remember when the Python designers tried to make Python Turing-incomplete by design?

caird coinheringaahing

@hyper-neutrino Ah yes, `{ is equivalent to `, my mistake

hyper-neutrino

17:24

well, sort of

arity is different :p

caird coinheringaahing

Shame that you can't use ` on monads

Means you have to do <monad>{` (for whatever reason)

hyper-neutrino

yeah

caird coinheringaahing

Guess @lyxal gets to keep their double tag badge

hyper-neutrino

lol

i figured it'd get dismissed; it's really minor :p

caird coinheringaahing

I'm surprised that its full on declined rather than deferred

hyper-neutrino

17:33

now, what if someone gets two gold tag badges? will they get an upgraded hammer? :p

Redwolf Programs

It'd both close and reopen the post with one click :p

caird coinheringaahing

Nah, it closes the target as a dupe of the dupe :P

Redwolf Programs

Thanks, MDN :p

hyper-neutrino

wow i'm dumb

2

i made a picolisp answer to a question about picolisp's multi-bracket closing feature... and didn't use picolisp's multi-bracket closing feature...

fixed within grace period :p

user

18:28

@hyper-neutrino You and ASCII-only have said that many times before

hyper-neutrino

except since i recently started putting the ' in "i'm dumb" / "i'm stupid" / etc these search results are inaccurately distributed

user

@lyxal You'll be happy to learn SE's not going to mess with your double code-golf tag badge

3

@hyper-neutrino Oddly enough, searching for i'm dumb doesn't yield any results

I hate chat search

hyper-neutrino

@user caird already pinged lyxal over that

but i guess fittingly for a double tag badge, lyxal gets a double ping :D

@user s/search/

user

@hyper-neutrino Oh lol

caird coinheringaahing

@user Get ninja'd :P

user

18:37

@hyper-neutrino s/((?<= )h)|search/

hyper-neutrino

I ~~h---ate ---c---hat ---search~~

good job markdown

Redwolf Programs

@user I h ate c h at s ~~hearch~~

Yum

hyper-neutrino

D:

Redwolf Programs

:p

user

@RedwolfPrograms I h ate c h at s ear ch. That's the best part :P

@hyper-neutrino Quite a bold statement there.

~~I h~~ a tech at sea ~~rch~~ (stretching it a bit)

Ooh, Python's pattern matching isn't bad

CMP: What's the best way a language/something else has implemented pattern matching?

Elixir's pattern matching is pretty OP

ngn

18:45

@user by not having it at all :)

user

@ngn Why do you think so? (not challenging you, just curious)

Or were you just ~~joking~~ making a joke?

ngn

generally, the best way to implement something is to avoid the need to implement it

Wezl

@user prolog's pattern matching is very powerful (e.g. A = [1,2|A]), however it does require function scope, which gets in the way of lambdas for example

@ngn this, but you should still implement it anyway

ngn

isn't pattern matching just an if-then in disguise?

Wezl

no, because it can also bind variables

user

18:50

@ngn Assuming you're talking about k, what does it do to remove the need for pattern matching?

Peilonrayz

@ngn Kinda, but patterns like Node(left=None, right=Node(1)) gets a bit unwieldy

user

@ngn I guess you could boil everything down to lower-level constructs, but sugar's always good

Wezl

> always

um, no

ngn

@Wezl ok, then: if-then + assignment + recursion

user

@Wezl Okay, maybe not in a hamburger :P

ngn

18:51

@user nothing to do with k

user

@ngn That can get verbose (not necessarily a bad thing, but it distracts from the task at hand, whatever that may be)

I wish Python had added extractors to its pattern matching PEP :/

ngn

@Wezl does A = [1,2|A] make an infinite list A=[1,2,1,2,1,2..]?

Wezl

I think so

user

@ngn I'm not sure that's pattern matching, though

ngn

@user then, is there a difference between pattern matching and destructuring assignment?

Transcript for