The Nineteenth Byte

Taken from StackOverflow Question Challenge Given an encoded string in format n[s] where n is the amount of times s will be repeated, your task is to output the corresponding decoded string. n will always be a positive number s will always be lowercase encoded format is number [ string ] stri...

Zacharý

@Adám, ⍺∪⍵ is the same as ⍺,⍵~⍺, right?

Adám

@Zacharý Yes, although ⍵,⍺~⍵ or ⍵,⍨⍺~⍵ might make more sense.

New Main Posts

6:41 PM

3

Q: Number of FIFO cache misses

This challenge is really simple (and a precursor to a more difficult one!). Given an array of resource accesses (simply denoted by nonnegative integers) and a parameter n, return the number of cache misses it would have assuming our cache has capacity n and uses a first-in-first-out (FIFO) eject...

code-golf

Zacharý

6:53 PM

Is ⍎ a schizo as well?

Erik the Outgolfer

yes

e.g. ⍎'+'

the only primitive function that can return a function

Adám

@EriktheOutgolfer Not (semi-)officially. The term schizophrenia is (technically wrongly) applied to those primitives that suffer from dissociative identity disorder, a.k.a. multiple personality disorder. ⍎ is most definitely a function, and user defined (trad)fns can also return functions. It just happens to be the only primitive doing so.

Erik the Outgolfer

@Adám oh, so it applies to e.g. /, \, ., etc.?

Zacharý

@EriktheOutgolfer I don't think .

Adám

@EriktheOutgolfer Only to /, ⌿, \, and ⍀. . is always unambiguously an operator (or syntax).

Zacharý

7:04 PM

Then what's the case with ∘ then?

Erik the Outgolfer

@Adám but it can have ∘ as its left argument, isn't that enough for it to be considered suffering from "multiple personality disorder"?

like, both an operator and an ugh

Zacharý

"And an ugh," is "ugh" a term I'm forgetting about

Adám

@Zacharý ∘. is special anomalous syntax, and doesn't parse as ∘ and .

@Zacharý No. We have "fgh"s, though :-P

Zacharý

Ah, I see.

Erik the Outgolfer

@Zacharý an "ugh" is what usually causes you to say "ugh" ;)

Zacharý

7:07 PM

Ah

Erik the Outgolfer

@Adám but then I can't do e.g. ∘.∘.+ without causing a syntax error, only ∘.(∘.+), so it does try to parse ∘.∘ first, no?

or at least, it parses ∘. ∘

Adám

@EriktheOutgolfer Yes, it parses ∘. ∘ and throws an error before even getting to the rest.

I do wish for ⍺ f\ ⍵ to deprecate ⍺ ∘.f ⍵.

Erik the Outgolfer

@Adám doesn't ⍺ (f⍤0 15) ⍵ already do that job? :P

Adám

@EriktheOutgolfer Sure, but is obscure to many, and not future-proof if we increase max-rank (may actually happen semi-soon).

Erik the Outgolfer

@Adám (what I was about to say)

Adám

7:13 PM

Also, f\is way golfier ;-)

Erik the Outgolfer

@Adám or, simply put, deprecates a >50-year-old syntax oddity that should've never been there

Adám

@EriktheOutgolfer Yes. But then again, all the slashes are anomalous due to history too. ∘.g was originally seen as an edge case of f.g before operators where recognised as a separate class.

The set of anomalous operators ≡ the original set of operators.

Zacharý

@Adám Well, that's one "N/A" down in RAD's docs

Adám

@Zacharý What?

Zacharý

Filling in a use for ⍺ f\ ⍵

Adám

7:19 PM

@Zacharý You had that planned before, no?

Leaky Nun

input length: 100; number of steps used: (mean = 44.4, sd = 4.76)
input length: 200; number of steps used: (mean = 85.9, sd = 6.66)
input length: 300; number of steps used: (mean = 126.0, sd = 7.2)
input length: 400; number of steps used: (mean = 166.0, sd = 8.32)
input length: 500; number of steps used: (mean = 203.0, sd = 7.6)
input length: 600; number of steps used: (mean = 244.0, sd = 8.35)
input length: 700; number of steps used: (mean = 283.0, sd = 8.98)
input length: 800; number of steps used: (mean = 326.0, sd = 9.53)

Zacharý

I had planned that there would be a use, but not any use in specific.

Leaky Nun

> 0.397 n + 5.83

Adám

@Zacharý I forgot, what are you doing about schizos again?

Leaky Nun

> input length: 500; number of steps used: (mean = 203.0, sd = 7.6)

I have a solution (the "linear-time linear-space" solution mentioned above, optimized a bit) that scores 1829 on the secret test case. Try to beat it! — user202729 yesterday

Zacharý

7:20 PM

@Adám They don't exist in RAD.

Adám

@Zacharý Right, but do you then have separate symbols for replicate and reduce?

Zacharý

@Adám Yes & No, ϼ is planned as a replicate-like function (when in the dyadic case), and I might have (X/)Y be replicate as well

Leaky Nun

I refactored the whole code

my goal is to eventually reach 0.333 n

since you can move 3 blocks each time

Adám

@Zacharý What about expand then?

Leaky Nun

it's 20:23 and I'm going to eat

Zacharý

7:23 PM

@Adám Same hacky deal with \.

Adám

@Zacharý What, two symbols?

Zacharý

@Adám No, the hack of using (X\)Y

Adám

@Zacharý But then \¨ and \⍨ etc. can't work.

Erik the Outgolfer

@Adám hm, would ⍺ f⊙ ⍵ be any better?

Zacharý

@EriktheOutgolfer That's RAD's plan (for f.g it's that symbol, but for ∘.f it's a circled circle)

Erik the Outgolfer

7:26 PM

like, ∘ + . = ⊙

Adám

@EriktheOutgolfer I thought about that too, but we want to minimise glyph proliferation. Especially when not adding any new functionality.

@Zacharý I personally think f.g should be f˙g which looks operator-y like ¨ and preserves the dot in dot product.

2.5 and namespace.name can use the same symbol, as in a sense . "dots into" the [2,3) domain just like it "dots into" namespace.

Zacharý

@Adám See the edit, f.g is f⊙g and ∘.f is f⌾

Adám

@Zacharý Sure, but somehow they just don't look "right" to me. By all means, go ahead!

Erik the Outgolfer

xahlee.info/comp/unicode_APL_symbols.html

> Here's a complete list of APL symbols in Unicode.

misses ⊙

Zacharý

@EriktheOutgolfer All of these are already in RAD's codepage. I needed to fit a whole byte, so I just used all of the APL symbols from miscellaneous and technical.

Except ≬, and all the greek symbols are actually the greek symbols for RAD

And | and ~ are ASCII

And *...

Adám

7:38 PM

@EriktheOutgolfer I'm not sure how he compiled that list. This one is better.

Erik the Outgolfer

@Adám ah, I was looking for that specific one, thanks...I actually have an account there, so I can edit the page with what I happen to discover that isn't in there

Adám

@EriktheOutgolfer Cool. That's more than I can say. I just email Kai.

@EriktheOutgolfer Ah, yes, I see that at the bottom.

Erik the Outgolfer

@Adám you should make an account there (you must call it AdámBrudzewsky, or, if that doesn't work, AdamBrudzewsky)

Zacharý

⊙ isn't actually an APL symbol, it's "CIRCLED DOT OPERATOR" or something like that

Erik the Outgolfer

then wait for Kai (or some other admin) to approve it, and then you're all set!

by "APL symbol" I don't mean just the symbols that have "APL" in their name ;)

Zacharý

7:42 PM

I don't think it's used in an APL though.

Erik the Outgolfer

otherwise compiling them would've been a piece of cake

that's used in NARS2000

Zacharý

They missed ⋄!

Adám

@Zacharý Who? xahlee?

Zacharý

@Adám Yeah

Erik the Outgolfer

@Zacharý xahlee isn't trustworthy in other ways too :P

Adám

7:44 PM

@Zacharý Again, it isn't Unicoded as APL-specific.

I think APL is the only programming language that has been specially catered to by UCS.

Erik the Outgolfer

@Adám which is part of what makes xahlee crappy ;)

(although not all chars there have "APL" in their names either)

Adám

@EriktheOutgolfer How would he know if he doesn't actually use APL.

Erik the Outgolfer

also, why is ⋄ called DIAMOND OPERATOR >_>

Adám

@EriktheOutgolfer The rest (notably, not ⋄) are visible on the keyboard photo.

Zacharý

@EriktheOutgolfer 'cuz math

Adám

7:48 PM

I would tell xahlee to look at aplwiki, but I'm not willing to sign up to a social medium.

Erik the Outgolfer

and, while I'm there, I'm actually going to make another edit at that table

⍂ was used in APLX

Zacharý

What was it used for?

Adám

@EriktheOutgolfer I don't think so. IBM reserved ⍂ and also ⌻ for some stuff (can't recall what), but it wasn't released, I think.

Erik the Outgolfer

@Adám but I have downloaded the APLX interpreter which seems to have it in its docs

that is, the version dyalog.com hosts (5.1.0)

Adám

@EriktheOutgolfer Not confusing it with ⍇ et al.?

Erik the Outgolfer

7:54 PM

@Adám no

the first one is from aplwiki, the second one from APLX's docs

and yes, I checked that the copying worked perfectly too

Zacharý

._. I definitely used these symbols differently than their respective APL's

Adám

@EriktheOutgolfer ⍂ and ⌻ are in APLX's ⎕AV, but they are not used.

Zacharý

@Adám like ⍫

Erik the Outgolfer

@Adám oh huh

yeah, I didn't edit right away for a reason

Adám

@Zacharý In Dyalog APL, yes. But in APL+ it is used instead of ∇ to ⎕LOCK a function.

Wonder why xahlee has ≬ . I don't know of any APL that uses that.

Erik the Outgolfer

7:58 PM

xahlee is...weird

Zacharý

I'll just make ⍺ f\ ⍵ <=> f\¨⍺,/⍵ to keep my sanity with \, ⍀, and ⍂ in RAD.

Zacharý

8:24 PM

CMP is having ⍵ be a niladic atom with the value of a function (the value of ⊢ in specific) too hacky?

Erik the Outgolfer

atom?

Zacharý

Nilad, like 1 or ⍬

Actually, never mind. That wouldn't let & (de/reference a function) work correctly on trains with ⍵

Adám

@Zacharý You'll do an internship at Dyalog eventually, right?

Zacharý

@Adám Hopefully!

Zacharý

8:50 PM

CMP: Which one of ⍰ ⍔ ‽ should be used to dereference a function?

Mr. Xcoder

@Zacharý ‽ imo — It gives me that Uh, what variable are you talking about?!, it's gone feeling

Zacharý

Okay, so ⍺((f&⍬)‽)⍵ will be equivalent to ⍺ f ⍵.

Erik the Outgolfer

um, & is a monadic operator AFAIK, unless there's something else going on in RAD too

Zacharý

& is a monadic operator in RAD (or will be)

Erik the Outgolfer

so, applying a new thread of f to ⍬?

Zacharý

9:02 PM

No. If RAD ever gets multithreading, that'll be ∥

Erik the Outgolfer

so, what is & then?

Zacharý

& "boxes" f into a nilad (that way a function can use it).

And the ⍬ is there because a monadic operator can't return a nilad.

Adám

@Zacharý Why not use ⍍ and ⍔ to array-ify and function-ify as ∆ is a valid name and ∇ symbolises a function?

Erik the Outgolfer

@Adám the latter looks terribly much like APLX's logo

Adám

@EriktheOutgolfer ⊃∨∧⌊○⊂

Zacharý

9:19 PM

@Adám Is ... that supposed to be a face‽

Erik the Outgolfer

Dyalog's logo

Adám

@Zacharý No, compare to:

Zacharý

_¯_

betseg

sleeping bat

Zacharý

@betseg I just went for an upsidedown face

def RTLReduce(f, arg):
	if type(arg) != list and type(arg) != str:
		return arg
	if len(arg) == 0:
		raise SyntaxError("UNKNOWN ERROR")
	if len(arg) == 1:
		return arg[0]
	return f(arg[0], RTLReduce(f, arg[1:]))

^ Will that be horrible performance?

Adám

9:22 PM

@EriktheOutgolfer APL implementers tend to use APL glyphs as logos. APL+ uses a Penrose ∇ but used to use an APL *

Zacharý

@Adám ._. I thought it was called "del"

Adám

Penrose triangle

The Penrose triangle, also known as the Penrose tribar, or the impossible tribar, is a triangular impossible object. It was first created by the Swedish artist Oscar Reutersvärd in 1934. The psychiatrist Lionel Penrose and his mathematician son Roger Penrose independently devised and popularized it in the 1950s, describing it as "impossibility in its purest form". It is featured prominently in the works of artist M. C. Escher, whose earlier depictions of impossible objects partly inspired it. == Impossible object == The tribar appears to be a solid object, made of three straight beams of square...

Zacharý

OH. That makes more sense

Adám

I don't know if it is intentional, but I've always thought Dyalog's logo looked like ⊃∨∧⌊○⊂ which is a valid Dyalog APL expression.

Zacharý

Speaking of a bunch of symbols. What was the expression consisting of a LARGE amount of APL symbols that resulted in 42? I remember seeing that somewhere

Adám

9:31 PM

@Zacharý ⍎⊖⍕⊃⊂|⌊-*+○⌈×÷!⌽⍉⌹~⍴⍋⍒,⍟?⍳0

@Zacharý That's the classic version from here, but I like my cut-down version better: ⍎⌽⍕⌈*○~⍴⍬ or maybe even a modern version: ⍎⌽⍕⌈*○≡⍬

Zacharý

⍎⊖⍕⊃⊂|⌊-*+○⌈×÷!⌽⍉⌹~⍴⍋⍒,⍟?⍳0 works.

Adám

@Zacharý Yeah, that's the official one. Comes built-in with APLX, for example.

Zacharý

I wonder if it's possible to cram even more APL symbols in...

⍎⊖⍕⊃⊂|⌊-*+○⌈×÷!⌽⍉⌹~⍴⍋⍒,⍟?⍳≡≢⍬ works as well

Adám

@Zacharý Throw in ⊢⊣⌷

@Zacharý And why not =/≠\?

Erik the Outgolfer

@Adám or even better (⊢⊣⌷)

Adám

9:43 PM

@EriktheOutgolfer Sure, that works.

@Zacharý (⊢⊣⌷)⍨

Zacharý

⍎⊖⍕⊃⊂|⌊-*+○⌈×÷!⌽⍉⌹~⍴⍋⍒,⍟?⍳≡≢=⌿≠⍀⊢⍨⊣⌷(↑↓)⍬ is what I have

Adám

@Zacharý You don't want parens?

@Zacharý Hey: {⍺⊢⍵}⍨

@Zacharý Insert ⊣⌸ before ,

@Zacharý And ⍸ before ⍬.

@Zacharý ⊆ on the very left.

Zacharý

⊆⍎⊖⍕⊃⊂|⌊-*+○⌈×÷!⌽⍉⌹~⍴⍋⍒,⊣⌸⍟?⍳≡≢=⌿≠⍀{⍺⊢⍵}⍨⌷(↑↓)⍸⍬

Adám

@Zacharý And ∪ to the left of ,

Zacharý

Hold on, I have to go eat right now.

Adám

9:52 PM

⊆⍎⊖⍕⊃⊂|⌊-*+○⌈×∘÷!⌽⍉⌹~⍴⍋⍒∊,⊣⌸⍟?@⌷¨⍳≡≢⍣=<⌿≤/≥\>⍀⍪{⍵⋄∇⍺}⍨(↑↓)⍸⍬

Leaky Nun

it looks like people are speaking gibberish

Adám

@LeakyNun We're trying to produce 42 using as many APL primitives (once each) as possible.

Erik the Outgolfer

that's pretty much the point here :P

Adám

@LeakyNun Gibberish computer code is on-topic here, right?

Leaky Nun

sure

Zacharý

10:12 PM

⊆⍎⊖⍕⊃⊂|⌊-*+○⌈×∘÷!⌽⍉⌹~⍴⍋⍒∊,⊣⌸⍟?@⌷¨⍳≡≢⍣=<⌿≤/≥\>⍀⍪{⍵⋄∇⍺}⍨(↑↓)⍸⊢⍬

you left out ⊢

⊆⍎⊖⍕⊃⊂|⌊-*+○⌈×∘÷!⌽⍉⌹~⍴⍋⍒∊,⊣⌸⍟?@⌷¨⍳≡≢⍣=<⌿≤/≥\>⍀⍪{⍵⋄∇⍺}⍨(↑⍷↓)⍸⊢⍬

Adám

@Zacharý Ah, clever, now we need to work in all the other dyadic-only functions.

Zacharý

Stick them to the left of ⍺? That'd be a little cheaty

Adám

@Zacharý #{⍺=⍵} and use ⍨ somewhere else.

@Zacharý No, you need to combine the monadics and the dyadics in the train (mdmdmdm)

Zacharý

Already used = though

Adám

@Zacharý Just as an example.

Anyway, gtg for the night ○/

Zacharý

10:18 PM

Okay, I'll post in APL Orchard if I get all of the primitives in one expression

Zacharý

10:48 PM

⍎⊖⍕⊃⊂|⌊-*+○⌈×∘÷⌹~⍴⍋⍒∊⊣⌸⍟?@⌷¨⍳≡≢⍣=⊤⍨<⌿≤/≥\>⍀⍪{⍵[]⋄∇&⍺}(↑⍷↓≠⊢∧⊆∨∪⍲⌽⍱!⊥⍉∩,)⍸⍬

is as far as I can get.

Transcript for