The Nineteenth Byte

12:00 AM

>>> g(1)
[3, 5]
>>> g(2)
[12, 25]
>>> g(3)
[144, 325]
>>> g(4)
[3456, 8125]
>>> g(5)
[41472, 99125]

Alex A.

@orlp Are there any performance differences in Python using named lambdas versus def'd functions?

orlp

@Doorknob git clone https://github.com/orlp/dotfiles

python3 -m pip install ptpython

then dotfiles/pcalc/pcalc

Helka Homba

@AlexA. You know, when rendered small it does look like the french flag. But that's fine too :)

Alex A.

@HelkaHomba :)

Doorknob

@orlp ptpython looks nifty

orlp

12:03 AM

@Doorknob also very handy functions I have pre-defined are copy and paste

copy(2**1024) and bam you're done

17976931348623159077293051907890247336179769789423065727343008115773267580550096‌313270847732240753602112011387987139335765878976881441662249284743063947412437776‌789342486548527630221960124609411945308295208500576883815068234246288147391311054‌0827237163350510684586298239947245938479716304835356329624224137216

Alex A.

What is ptpython?

orlp

no need to muk around with selecting the right thing :P

Doorknob

@orlp that's definitely a whole lot easier than <mod+'>irb<cr>2**1024<cr><alt-esc>kbve<cr> :P

orlp

or let's say alex's msg needs to be double spaced

>>> " ".join(paste())
'W h a t   i s   p t p y t h o n ?'
>>> copy(_)

Cᴏɴᴏʀ O'Bʀɪᴇɴ

What's this magic?

orlp

12:06 AM

@CᴏɴᴏʀO'Bʀɪᴇɴ you can use it yourself

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@orlp What is it though?

orlp

@CᴏɴᴏʀO'Bʀɪᴇɴ scroll up to git clone and follow the instructions

@CᴏɴᴏʀO'Bʀɪᴇɴ a python interpreter on crack :P

Doorknob

@orlp for that kind of stuff I use cVim's "pop out vim instance" feature. If a textbox contains asdf, I can hit Ctrl+I to get a gvim instance with the text inside of it, :s/./& /g or similar, then ZZ drops the modified file back into the textbox

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@orlp Uhhh nevermind I don't want to know anymore

orlp

@Doorknob sure, but pasting into a vim I have open somewhere is no problem

lowering the boundary to use a Python interpreter is glorious

I also use it constantly while programming

(1 << 18) // 3

hrm, what if I sum all 5 dividends? sum((1 << 18) // i for i in range(1, 1+5))

Cᴏɴᴏʀ O'Bʀɪᴇɴ

12:12 AM

> 1+5

orlp

@CᴏɴᴏʀO'Bʀɪᴇɴ it's a mental thing I do

Nathan Merrill

hey, I'm struggling with how to design the following code:

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@orlp o okai

orlp

whenever I want an inclusive range starting on 1 in Python I do range(1, 1+n)

Nathan Merrill

In my KoTH controller, a user can create a class with methods that are the "user" methods

Cᴏɴᴏʀ O'Bʀɪᴇɴ

12:14 AM

I guess that makes sense, using less mind to put something down

Nathan Merrill

so, for example, in Good vs Evil, they would make a class with a "boolean vote()" method

how would my controller automatically detect such methods, and autofill them with an implementation that communicates over STDIO

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Shinedown - Simple Man

►

Easterly Irk

@CᴏɴᴏʀO'Bʀɪᴇɴ post yourself to twitch so we can watch you

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@EasterlyIrk XD I'm already dead

Easterly Irk

roflmao

livestream.com/producer/software

Alex A.

12:24 AM

Are there always at least n primes less than n^2?

Sp3000

For n > 1, yes if I remember correctly

Alex A.

Nice, thanks!

Cᴏɴᴏʀ O'Bʀɪᴇɴ

TIL, wow

Sp3000

15

Q: The n-th prime is less than $n^2$?

Let $p_n$ be the n-th prime number, e.g. $p_1=2,p_2=3,p_3=5$. How do I show that for all $n>1$, $p_n<n^2$?

Just so that Alex can't blame me if maths breaks some day :P

Alex A.

Perfect, thanks for the link.

Now I know who to blame. :P

Easterly Irk

12:45 AM

Doctors remove wrong body part. Sounds worse than it is.

@CoolestVeto I have dubstep on autoplay now. :P

Sp3000

@AlexA. I said n>1 :P

Alex A.

Yes?

Sp3000

Does f(1) work?

Easterly Irk

@AlexA. did you try iTerm2?

It crashes for me. :(

Alex A.

@Sp3000 Ah crap. Easy fix but it costs two bytes: change n^2 to n^2+1

Easterly Irk

12:54 AM

Where is this?

Sp3000

You can do it in 1 with 2n^2 :)

Also, Rational's kinda bulky, gimme a bit

Easterly Irk

true

Alex A.

@Sp3000 Oh but it works so well :P

Sp3000

n->prod(1-2.//-~primes(2n^2)[1:n].^2) maybe?

Alex A.

o_O

I didn't know about .//

Easterly Irk

12:56 AM

What is .//?

Alex A.

Elementwise rational construction

Cᴏɴᴏʀ O'Bʀɪᴇɴ

//?

Easterly Irk

no.

Sp3000

1/2 gives a float, but 1//2 gives a rational. .// is just element-wise over a vector

Easterly Irk

@AlexA. cool

Alex A.

12:57 AM

[1,2,3] .// [5,6,7] is [1//5, 1//3, 3//7]

Yes, very cool. Thanks a bunch, @Sp3. :D

Sp3000

(Yay, I learnt something useful :P)

Alex A.

Oh, overflow

I need a big in there somewhere

Sp3000

... Ah :/ right that.

Alex A.

n->prod(1-big(2).//-~primes(2n^2)[1:n].^2) seems to work :D

oops

Copied the wrong thing

Sp3000

That'll probably work until the n is big enough such that the nth prime overflows

Which would be quite a while, but I dunno if orlp wanted it to work for all inputs :/

Alex A.

1:00 AM

@orlp For your challenge, do submissions have to work theoretically for all inputs? (I'd be surprised if many of the current submissions do.)

orlp

@AlexA. yes

otherwise the submission is obviously incorrect

not sure why people keep asking this on every single question I post

why would your answer suddenly not have to work for n+1 when it does for n? does the question say so? no

Alex A.

Overflow

Limits of the language

orlp

too bad?

work around it

Sp3000

Probably because, for a lot of questions, there's a difference between "working for all practical inputs" and "working for all inputs theoretically"

Alex A.

I usually assume the former applies when unspecified.

orlp

1:03 AM

the only thing that's excusable is not having enough runtime or memory

but given an all-powerful machine that's really really fast with tons and tons of memory your program should just work

Helka Homba

Not all languages have arbitrary precision integers

orlp

@HelkaHomba that doesn't matter

mınxomaτ

@orlp That is hardly the case with any challenge.

orlp

code-golf competes within their own language

mınxomaτ

The usual assumptions is always within the language's limits (with rare exceptions like 8bit langs). Otherwise we'd have to delete hundreds of answers from this site.

3

Helka Homba

1:05 AM

^

Alex A.

To reiterate that not all of the current submissions do that: "Fails for large inputs because only integers up to 2^52 can be accurately represented internally." (from Luis' MATL)

mınxomaτ

@AlexA. Which is reasonable tbh.

orlp

well, I'm in the camp that allows invalid submissions if they're interesting

but they just won't get accepted

Alex A.

@mınxomaτ Yeah, I think so

orlp

if you want my checkmark, your program should work, no exceptions

unless I specifically mention otherwise

Alex A.

1:06 AM

Per our policies, any submission that doesn't fully comply with the spec should be removed.

mınxomaτ

@AlexA. Beyond that, it would just be a requirement for the requirement's sake.

Alex A.

So if the spec is that it must theoretically work for all inputs, then answers that don't work that way should be removed.

Helka Homba

@orlp Then mention that in the challenge, else people will follow the rule minx just mentioned and be confused when you don't accept them

Doorknob

It's a bit ridiculous to expect C answers to import arbitrary precision arithmetic libraries

GamrCorps

^

mınxomaτ

1:08 AM

Also, requiring all inputs to math challenges will just add the overhead of including a BigInt lib to any language that doesn't have it. Which is neither interesting nor does it add anything to the challenge.

Sp3000

@AlexA. Side note: you can probably replace primes(2n^2) with filter(isprime,1:2n^2) and assume n is a BigInt, because otherwise there's no practical way to input anything like f(2^100) anyway :P

orlp

@Doorknob it's a bit ridiculous that C can just hardcode n=1 to 5 and ignore the rest

Doorknob

@orlp except that's a loophole

orlp

@Doorknob well it'd be a fully functional implementation, by your rules

Doorknob

yes, and it would also be deleted as it violates site policy

mınxomaτ

1:09 AM

^

orlp

but why? it has the exact same behavior as the other implementation

Doorknob

because it's a loophole

Alex A.

@Sp3000 isprime is probabilistic for BigInts

orlp

I just find it weird that every languages has to solve it's own challenge, depending on how crippled it is

Doorknob

wget goo.gl/asdf; python asdf is also fully functional

orlp

1:10 AM

all these languages are turing complete

Sp3000

@AlexA. ... D: damnit didn't know that

orlp

they can be implemented

the question is - how many bytes does it take?

if your language does not have built-in multiprecision, well, then it's going to take more bytes

it's that simple

that's the core of code golf: kolmogorov complexity

Alex A.

@Sp3000 Yeah :/ julia.readthedocs.org/en/latest/stdlib/numbers/…

Doorknob

It would really be prohibitively obnoxious to implement arbitrary precision in a language that doesn't have it. It's basically language restrictions for no good reason.

Helka Homba

@orlp But few will want to write a BigInt library for a language that doesn't have one, and won't bother answering at all.

orlp

1:11 AM

there's no point to it if language A gets a free pass because it's a less powerful language

mınxomaτ

@orlp No, because you shift the focus from actually solving the problem to creating a limited BigInt implementation for the challenge, which is likely bigger than the actual code.

There's no benefit.

orlp

and all this is still too vague, do we have an unambiguous policy of how broken a program is allowed to be because it's a less powerful language?

Doorknob

Insisting on arbitrary precision is analogous to insisting on a certain I/O format. Your language can't do file I/O? Too bad, you don't get a free pass because your language isn't powerful enough.

orlp

@Doorknob I/O is well defined though

so far this isn't

mınxomaτ

@orlp You consider it broken. To prevent the ambiguity, many challenges set an upper limit for inputs, which is what you should do instead of just assuming everyone thinks how you think (which, as we already established, is not the case ... in this case) IMO.

Alex A.

1:15 AM

Arbitrary precision isn't well-defined?

orlp

@mınxomaτ but we need a default

I didn't set a limit for my challenge

Doorknob

@orlp "the maximum representable integer range in your language"

orlp

to my knowledge the default was 'no excuses'

@Doorknob so.. does the input have to be representable in that?

the output?

all intermediate results?

Doorknob

yes

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@Doorknob The only problem is boolfuck

orlp

1:16 AM

which is it?

Doorknob

@CᴏɴᴏʀO'Bʀɪᴇɴ what about it?

mınxomaτ

@orlp It wasn't. As is implied by the before mentioned hundreds of answers that would suddenly all be invalid by you "rules".

Doorknob

@orlp all of the above?

Sp3000

5

Q: Integer range defaults for challenges

When writing challenges involving integer I/O, I always find myself having to type the same thing: You may assume that the input and output will not exceed the maximum representable range in your language. I think it would be useful if we had a default for integer ranges in challenges, to avoid n...

discussion defaults

mınxomaτ

^

Cᴏɴᴏʀ O'Bʀɪᴇɴ

1:16 AM

@Doorknob Trivial solutions.

Sp3000

Related - note that this never got resolved

Doorknob

19

A: Loopholes that are forbidden by default

Abusing native number types to trivialize a problem It is common practice to restrict challenges to cases where input, output and/or intermediate values of the algorithm of choice fit into the language's native number type. At least for input and output, this is generally assumed even if not sta...

orlp

@Doorknob so is my answer automatically OK for any range, as long as my implementation at some point does some massive exponentiation?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@Doorknob I've seen it before though.

Doorknob

@CᴏɴᴏʀO'Bʀɪᴇɴ seen what?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

1:17 AM

Boolfuck trivializing the problem

Doorknob

@orlp if the algorithm wouldn't theoretically work should the integer type be larger, it's a loophole

@CᴏɴᴏʀO'Bʀɪᴇɴ why aren't you using your flags

Cᴏɴᴏʀ O'Bʀɪᴇɴ

I just found out about this loophole today

orlp

@Doorknob but how do you distinguish that from legitimate tricks?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

brb

orlp

like unsigned(-56) to get some specific large number

Doorknob

1:19 AM

unsigned(-56) can be turned into unsigned_type::max_value - 56 without breaking the algorithm

it really should be quite evident if someone tries to abuse what you're describing

orlp

@Doorknob not if there is some 'magic' assumption where integer types get larger

Doorknob

what do you mean?

orlp

w/e I don't care enough to argue about this

I'll just be specific on my challenges

mınxomaτ

Fair enough

Doorknob

the community consensus is clearly in favor of allowing people to not implement obnoxious arbitrary precision. You're allowed to do that, but don't be surprised if it attracts downvotes

Sp3000

1:22 AM

This is a bit of a grey area, unfortunately :/ - I think I remember seeing a Python answer which had exec""*99, with the argument that "that was large enough for practical inputs". I thought about arguing against it, but I don't think I did.

orlp

I just feel that the current rule is vastly more ambiguous than you guys try to make it seem.

Doorknob

@Sp3000 That's a whole different can of worms. That's not a language restriction.

@orlp You do bring up a valid point—the rule could benefit from clarification and formalization

orlp

which has an integer limit of 100

Doorknob

7 mins ago, by Doorknob

19

A: Loopholes that are forbidden by default

Abusing native number types to trivialize a problem It is common practice to restrict challenges to cases where input, output and/or intermediate values of the algorithm of choice fit into the language's native number type. At least for input and output, this is generally assumed even if not sta...

orlp

go above that it and shuts down your PC

@Doorknob no no, it's a new language

and as long as I make that language before a challenge gets posted, it's k

Doorknob

1:24 AM

read the loophole

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@orlp Better x3

Sp3000

btw @orlp if you're clarifying, could you also clarify specifically that the implication is that probabilistic primality methods aren't okay?

orlp

@Doorknob so where does it become deliberate? 100? 300? 2^16?

mınxomaτ

Friggin ping.

Helka Homba

@CᴏɴᴏʀO'Bʀɪᴇɴ Is there yet to be a language that no one has any form of animosity towards?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

1:27 AM

Are the stars being cleared?

Alex A.

Yes

Dennis

@orlp If probabilistic prime check are out, most submissions to your challenge are invalid. Most languages that handle big integers use probabilistic prime checks, since the chance of a false positive is negligible in practice.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

M'kay

@HelkaHomba Hmmm... PHP? :3

Doorknob

@orlp it becomes deliberate when it's deliberate. There's really not much to explain. The answer I linked describes exactly when an answer violates the rule.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@HelkaHomba On a more serious note, maybe Mathematica.

orlp

1:28 AM

@Doorknob how does one objectively score 'deliberate', as it's a guess to the writer's intent

Helka Homba

@CᴏɴᴏʀO'Bʀɪᴇɴ No. People don't like that it's not free

Doorknob

@orlp it's not intent, it's the algorithm

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@HelkaHomba Ah, any form ,right.

orlp

@Doorknob how does one define 'the algorithm' if the code is apparently insufficient to describe it?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@HelkaHomba Rust?

orlp

1:29 AM

is it some 'magical' extended language where we act as if the integers were unbounded?

which brings me back exactly to my point of tricks like unsigned(-56)

Downgoat

@HelkaHomba C?

Doorknob

@orlp have you read the loophole post? It answers all your questions

orlp

@Doorknob also, that post seems to suggest it is not the default

> It is common practice to restrict challenges to cases where input, output and/or intermediate values of the algorithm of choice fit into the language's native number type.

Doorknob

keep going...

> At least for input and output, this is generally assumed even if not stated in the challenge specification.

Helka Homba

in 2016 Programming Puzzles & Code Golf Moderator Election, 8 mins ago, by Helka Homba

To all candidates: Explain why you want to be mod in 12 chars instead of 1200.

orlp

1:31 AM

@Doorknob I'm blind

Easterly Irk

2:10 AM

0

A: Solve the Secretary Problem

I think this is actually how people choose a marriage partner.

@Doorknob ^

pls hammer to death doom and destruction.

thankee

It came up in the first posts review.

El'endia Starman

He's basically wrong too. There is a well-known variant of the Secretary Problem that's exactly the same, except that instead of choosing a secretary, you're choosing a partner.

I'm pretty sure that's not how people pick partners in real life though...

mınxomaτ

> now

El'endia Starman

The difference a single letter makes. >_<

Nathan Merrill

hey, if any of you are interested in helping my KoTH framework along, I'd appreciate an answer here:

0

Q: Code Design: Delegation of arbitrary functions

On PPCG, we frequently have King of the Hill challenges, which pit different code bots against each other. We don't like limiting these challenges to a single language, so we do cross-platform communication over standard I/O. My goal is to write a framework that challenge-writers will be able u...

design delegation

xnor

@Dennis for your base -i+1 answer, it's nice how "carrying" every 2 at once lets you avoid getting digits above 3

i had been trying a similar approach with carrying from right to left, but it eventually messed up the base 10 representation by having digits go above 9

Dennis

2:25 AM

Yeah, that worked out nicely. Doesn't stand a chance against feersum's binary approach though...

Sp3000

Personally I'm pretty impressed with the answers so far, although I do wish the papers I've seen had proofs for the not-so-obvious tidbits :P

xnor

though it works on the test cases, i'm unclear if all cycling cases reduce to 11+111

i mean, all cases where the sum would be 0 and so digits keep getting pushed left

Nathan Merrill

we've had quite a few of these unorthodox base challenges

it'd be cool if a language had builtin support for it

El'endia Starman

Pytek will have built-in base conversion support. Maybe not to the point of having -1+i as a base though...

Dennis

Jelly has support for negative bases, but it stops at integers...

Sp3000

2:28 AM

I'd like to see a language support any arbitrary base, although it'd probably be pretty non-trivial to implement :P

I mean, not even Mathematica has that (I don't think?)

Nathan Merrill

I still don't completely understand how anything besides positive integer bases work

El'endia Starman

They work in exactly the same way.

Nathan Merrill

I mean the base is the number that it "wraps" over, which doesn't make sense for negative numbers

El'endia Starman

Well, if you think of it that way...

a_n*x_n + ... + a_1*x + a_0 is the general form.

Nathan Merrill

right, but a_n has a maximum value of the base

El'endia Starman

2:30 AM

Well, that doesn't have to be the case...

Sp3000

@xnor Not sure about this either, although I have no idea how to prove either way

xnor

@Dennis I think you can save a little by not computing the input length

s=sum(input());L=1
for _ in`s`*8:s+=1098*int(str(s).translate('0011'*64));L*=10
print s%L

Nathan Merrill

so, 3*10^1 + 23*10^0 is still base 10?

El'endia Starman

Technically, yes.

quartata

@Doorknob Have you worked with Shoes before?

El'endia Starman

2:31 AM

It's just convention that we simplify it.

It's like saying that multiplication is repeated addition, which falls apart when you get to imaginary/complex numbers.

Doorknob

@quartata This is the extent of the full experience that I have had with Shoes

(so, no)

xnor

@Sp3000 if checking for (3,7) does work, what's the shortest expression that's 0 only for it?

El'endia Starman

@Doorknob Ahem, you put user:me.

Doorknob

... whoops

El'endia Starman

hahaha

Doorknob

2:32 AM

>_<

El'endia Starman

I saw the user id number and I was like "...wait, that's mine." :P

Sp3000

@xnor Sorry, only for it?

quartata

OK. I just figured since you're into Ruby things that you might know it and be able to answer a question I had but anyways

Dennis

@xnor Oh, iterating over the input never occurred to me... Thanks!

xnor

i mean an expression in a and b that equals 0 when a=3,b=7, but is nonzero for any other pair (a,b)of non-negative numbers

except possibly a=7,b=3

Dennis

2:34 AM

@Sp3000 Um, bold question: Can the single non-alphanumeric separator be a plus sign?

Sp3000

Oh, you mean for golfing purposes

xnor

yup

@Dennis i see where this going :-)

Sp3000

@Dennis I was waiting for that question :) The answer is yes, and it was "non-alphanumeric" on the offchance someone needed that + sign :P

Dennis

Great! :)

orlp

gist.github.com/orlp/3de99a1917aba4b7d49c

arithmetic coder!
all I gta do now is load up the Pyth probabilities and I can compress Pyth effectively!

@Dennis let's see if Pyth can compete with Jelly again, if we remove the advantage the custom codepage brings

Dennis

2:38 AM

Well, the custom code page by itself isn't an advantage. Pyth uses ISO 8859-1 by default, so there are 96 more printable characters (and 63 unprintable one) that it could use but doesn't.

orlp

@Dennis Pyth intentionally limits itself to printable ASCII, with that limitation lifted inside strings

Dennis

Yes, I know. That's a design choice though, not a code page thing. Both Jelly's and Pyth's code pages have exactly 256 characters that can be encoded as a single byte.

orlp

@Dennis it is a code page thing, Pyth's builtins are limited to the ASCII codepage

Dennis

Pyth uses ISO 8859-1, not ASCII.

orlp

@Dennis inside strings, not outside of it

Dennis

2:42 AM

You cannot use different code pages inside the same source file.

orlp

@Dennis you can :)

Dennis

How?

orlp

Pyth switches code page when it encounters "

Dennis

o_O

orlp

@Dennis at least in Pyth5 the whole input file is never decoded in one chunk, it stays binary until a character is needed

Sp3000

2:43 AM

@xnor ... this is surprisingly tricky

orlp

@Dennis the distinction is rather pointless though

Dennis

@orlp OK... That's seems unnecessarily complicated, seeing how ISO 8859-1 and ASCII are perfectly compatible. Unless Pyth 5 doesn't use ISO 8859-1 for string anymore.

Sp3000

Can't think of anything apart from a==7>3==b atm :/

orlp

you can say that brainfuck uses ISO 8859-1, UTF-8, ASCII, because anything except it's syntax is ignored

LockOpeners

oh hey this exists

xnor

2:45 AM

@Sp3000 that works, though i dunno if you'd allow the output False for 0

orlp

@LockOpeners your profile picture has a nice top note, pure vanilla, it's a 10

LockOpeners

thanks

i've got that cream

that top milk

Sp3000

Oh wait, we want 0 when it's (7,3) don't we :/ nvm

LockOpeners

:)

xnor

@Sp3000 oh, right, you want the negation

LockOpeners

2:46 AM

I'm the friendly neighbourhood perl6 golfer

Sp3000

a-7|b-3 maybe?

... nope

xnor

what's wrong there?

orlp

@Dennis for now I'll focus my compression only on printable ASCII Pyth (so no funni string business), just because it makes the code eaiser

Insane

thinking of posting a question, it's pretty basic and is related to one that's already been well received. is the sandbox necessary for that?

Easterly Irk

@Insane yes.

Doorknob

2:49 AM

@Insane The sandbox is never necessary.

Easterly Irk

The sandbox will help determine whether it is too similar.

@Insane don't trust @Doorknob.

Dennis

@Insane It's not necessary (it almost never is), but if you're at all uncertain about your challenge idea, it might help.

Insane

since it's pretty simple and you guys could probably tell me right now if it's too similiar

El'endia Starman

The Sandbox is a very, very good idea for anyone that's not ~~Calvin~~ Helka or Doorknob.

Insane

can i just ask for ya'lls opinion

Doorknob

2:49 AM

sure

Easterly Irk

@Insane just post for like 5 min.

@Insane okay

shoot

Sp3000

@xnor f(0b1, 0b11101) only reaches 3,7, not 7,3

xnor

i might have said it switched, you want to check 3,7

so a-3|b-7

Doorknob

@EasterlyIrk Posting to the sandbox for less than... say, at least a few days is somewhat meaningless

Insane

codegolf.stackexchange.com/questions/75979/full-width-text this reminded me of my favorite ｆｕｌｌｗｉｄｔｈｌａｔｉｎｔｅｘｔ. i was thinking of posting basically saying "inspired by (above), ouput in fullwidth latin".

xnor

2:51 AM

since 7,3 goes to 3,7 but not the other way

Sp3000

Oh, right. Probably my fault - I thought they were interchangeable for some reason

Easterly Irk

@Doorknob @AlexA. once posted for like 15 minutes.

Insane

might add a bit of a twist

Easterly Irk

@Insane Difference between latin and normal?

Doorknob

@Insane It's definitely not a duplicate, but it's kind of boring (not that the original question wasn't either).

Alex A.

2:51 AM

@EasterlyIrk >_>

Easterly Irk

gonna call you in to testify.

Insane

@EasterlyIrk the latin uses the latin fullwidth characters

Easterly Irk

eh

xnor

is there a code page that has fullwidth latin contiguously?

Doorknob

tr a-z ａ-ｚ

Helka Homba

2:52 AM

@Doorknob Not as a staging area

Easterly Irk

I think to similar.

Insane

unicodemap.org/range/87/Halfwidth%20and%20Fullwidth%20Forms

boooooring

^^^^

@xnor UTF-8 does

2:52 AM

and ^^^

sorry

Doorknob

Actually tr !-~ ！-～, but yeah.

Sp3000

Yeah a-3|b-7 seems to work

Insane

@Doorknob i was thinking that too, but it adds a tad bit of complexity i guess

xnor

@Sp3000 that's gonna be hard to beat

Easterly Irk

Too similar, sorry.

Insane

2:54 AM

@EasterlyIrk i already got your opinion

Doorknob

It's not similar at all.

orlp

@Dennis do you have a pyth program I should compress?

Sp3000

Closest other I have in mind is a*b-21 which obviously does not work :P

xnor

@Sp3000 i forget that bitwise operators have precedence below arithmetic ones, that makes the | useful as an arithmetic or for nonzeroness

Sp3000

Sometimes useful, sometimes annoying :P

xnor

2:55 AM

@Sp3000 nor does a/b-3/7 in Python 3

Dennis

@orlp _shM.u,%J/eMeN\12-+PMeNm.B6/J2k,kQ

orlp

@Dennis thanks

Insane

I guess I'll just post it and see how it goes. I'm more inclined to take a mods opinion that it's not too similar to be considered a duplicate, and I'll hope the community as a whole doesn't think it's too boring.

Easterly Irk

@Insane do you even have a PPCG account?

orlp

@Dennis it seems to be, but just to make sure, that's a real program, right?

xnor

2:56 AM

@Insane i'm pretty sure most of us think it's boring

Doorknob

@Insane IMHO it's overly trivial.

El'endia Starman

^ and ^^

Easterly Irk

^

xnor

@Insane if you're going to solicit opinions then ignore them, why ask?

2

Dennis

@orlp Yes, Jakube posted it today.

Insane

2:57 AM

@xnor But I didn't, and I can take the opinions into my own accord in any way I please.

xnor

@Insane of course, you're free to, just don't be surprised if there's downvotes

Sp3000

@orlp Context, although I think compressing a more comparable answer might be more useful, since Jakube's used quite a different algorithm to Dennis here

xnor

for dupe checking there's Small Caps Converter, which is similar in task but different in implementation because there's no contiguous range

Insane

@xnor There'll probably be more downvotes if I post it now than if I had just posted it originally. It's not a duplicate, and it's less boring than the question that inspired it

Transcript for