Esoteric Programming Languages - 2015-08-27

« first day (2 days earlier) ← previous day next day → last day (1394 days later) »

Martin Büttner

12:07 PM

@TheNumberOne yeah, it's not too difficult manually

+->< becomes )({} and then you just need to get the layout right to implement the loops. doing that automatically might be a bit trickier, but in essence it will be similar to the BF to MarioLANG conversion.

@Sp3000 I've had an idea for ': pop a value and push the character codes of its string representation. I'm not entirely sure how useful it would be though.

Er... wait how does that work?

Martin Büttner

say you've got -105 on the stack, then ' would push 53 48 49 45 onto the main stack (the 45 on top)

Considering there's output as num, I don't see much of a use case

Martin Büttner

hm, I thought one might want to read input with ? instead of . sometimes, because it gives a different EOF, but it's not too useful. alternatively, it gives you a simple base-10 converter although that shouldn't be too much of a pain to do yourself either

well I'll leave it until I actually come across a case where I want to use it

I think base 10 convert should just be something you do yourself

Martin Büttner

12:22 PM

yeah, fair point

hm, I've been trying to golf tac... I've got three 15-byte solutions now but can't seem to beat that

You could make ' push char code I guess, but with the *10+ thing it's probably not necessary

Martin Büttner

@Sp3000 I'm not sure what you mean

e.g. 'a -> 97, but that's only one shorter than _97 hence the "might not be necessary"

Not sure if it's any more readable though

Martin Büttner

oh right, well that's a problem, because with that I can no longer have unrecognised characters (i.e. letters) as walls

or rather, it depends on the current context if a cell is a wall or not

Does your 15 tac handle NULs?

Martin Büttner

12:30 PM

yes

it only ends at EOF (returned as -1)

I've got 4 variants now...

@MartinBüttner This pastebin has a mapping for each BF command: pastebin.com/q7SZaYFx

Martin Büttner

oh, cool

@TheNumberOne you don't need to make - so complicated. modulo should work correctly for negative numbers

% pops y, pops x, pushes x%y (modulo; the sign of the result is the same as the sign of x).

Martin Büttner

oh right, that's a typo

thanks

I haven't tested any of what I wrote, so use at your own risk :)

Martin Büttner

12:38 PM

I'm not sure it works correctly. wouldn't it enter any loop at least once?

Damn I thought I had 12, but it doesn't handle NULs at the start

The top two rows of lines are for jumps to other parts of the code.

Martin Büttner

@Sp3000 do you have 15?

Not yet, working on it

Martin Büttner

@TheNumberOne the top row is where the instruction pointer is initially though, right?

12:51 PM

:/ I'm having a lot of trouble getting this to work for starting NULs

@MartinBüttner Sorry, I meant bottom two.

Martin Büttner

@TheNumberOne right... but in the left bracked, you just pass straight through when first entering it, right?

Look at the character directly below the third "

Martin Büttner

ohh, that somehow looked like it was a row below

but doesn't that mean it enters the loop when the current cell is 0, and skips it otherwise?

@Sp3000 13!

:/ are you suuuure yours handles NULs right? :P

Martin Büttner

12:59 PM

I don't see why it wouldn't

it's a bit tricky in powershell to test though

might fire up my Ubuntu VM later

wait...

I think you have a point :/

I'm correctly reading NULs, but I stop printing at the first NUL in the reversed string... crap

@MartinBüttner Looks like I got the zero and non-zero mixed up for both brackets.

;)

I have 21 for NUL-handling, I believe

Martin Büttner

18, I think (based on my 13)

ah no, doesn't work

hm, I think I need to try a different approach

Still 21 - I can't seem to compress this 3x3 loop or it'll all turn funny :(

Martin Büttner

16

(untested on 0, but I think it should work now)

1:15 PM

Do you do any ><v^? I still don't get how to use those effectively

Martin Büttner

no I don't

I'm not sure they are useful in these small golfs

I just have very tight loops

Maybe with a logical not

Martin Büttner

hm no, not using that either

You don't have one though :P

Martin Büttner

oh, I thought you meant bitwise

I've got 4 lines, 2 spaces, 3 nops btw

1:17 PM

3 lines, 5 spaces, 5 nops

:/

@MartinBüttner Okay, I give up - what do you have? :)

Martin Büttner

1:49 PM

  (}
.)),
"{@
""

@Sp3000 ^

... I forgot I don't have to start on the top left :P

Martin Büttner

ohhh, right

Damn everything's so tight

So where does it start - the top (?

Martin Büttner

yes

first non-wall in reading order

going right

,)"";#@
(" " }
   "{.

Yeah I don't think I can adapt mine like that :P

Martin Büttner

1:59 PM

uh... now that I see yours I have no clue why I'm shifting everything to the auxiliary stack o.O

Er... I thought it was so you can do your fancy tight looping XD

Martin Büttner

haha no. for some crazy reason I initially thought I needed that to reverse the string. which is kinda stupid :D

:P

Does that actually help save anything though?

Martin Büttner

hm, I don't know

2:18 PM

@MartinBüttner Your cat needs more golfing :P

Martin Büttner

lol, I know :D

I thought it looked nice like this though ;)

also, I just discovered the () trick an hour ago :P

,)@
.(

But... 5 6 chars D:

Martin Büttner

"but"?

(this is nice :))

The "but" is in response to "I thought it looked nice like this though ;)"

Martin Büttner

haha, okay

2:23 PM

What are you two working on?

Martin Büttner

golfing in Labyrinth

golfing what?

Martin Büttner

tac and cat

ah, okay :)

Martin Büttner

@Sp3000 I've got 15 for tac when exiting with an error, but if we allow that we can probably golf it a lot further, because we don't need @ at all

2:24 PM

how do you tell when you hit end of input?

Martin Büttner

Ruby returns nil at EOF, which I translate to -1 in Labyrinth

NULL?

:P I've never been a fan of the error thing personally, but I guess

Martin Büttner

(actually only , returns -1. ? returns 0 instead)

@Sp3000 neither have I

btw the 12 byte which assumed no NULs was

,);""
(" ."@

NULs ruins everything though :/

(I think you could probably do better than that as well :P)

2:31 PM

Tac:

,){(.@
}" "{

12 bytes

Is that the no-NULs version or does it work with NULs?

should work with nulls

Hmm doesn't seem to...

Yeah NULs mess up the print loop

Martin Büttner

@TheNumberOne heh, I just started experimenting with a read loop that only increments but doesn't decrement

@Sp3000 Oh, I see :(

2:37 PM

Somehow, I feel like I'm programming in Half Broken Car in Heavy Traffic, except this is a lot easier

Martin Büttner

wow that looks terrifying :D

,)"{"@
}" .(

Infinite loops :(

how ?

Martin Büttner

it prints an infinite number of NULs at the end

once { moves a 0 over the instruction pointer heads back down

2:43 PM

:(

,)1"{@
}" .(

Maybe?

Outputs a single SOH when the input is a single NUL

:(

oh, I see the problem.

,)1{@
}".(

This time, just maybe?

Nevermind.

Fails same test case :/

@MartinBüttner Your Fibonacci takes a while to run :(

Martin Büttner

does it?

it's not a sequence generator

it takes an integer as input

Either that or I'm doing something wrong

2:54 PM

@MartinBüttner You need diodes, cells that can be passed through only one way.

Okay, seems to be like this

1) Works fine piping in a file with a trailing newline

2) Errors when piping in a file without a trailing newline

3) Kinda screwy when I enter a number via shell and end the number by hitting Enter for newline

cat:

,)@
.(

Martin Büttner

hm, I see... I only tested echo 5 | ruby ...

36 mins ago, by Sp3000

,)@
.(

hifive we both beat Martin

(bet he wasn't trying :P)

Actually, for 3) I need to hit Enter twice for it to run

Martin Büttner

@TheNumberOne hm, would probably be convenient, but I don't really want any control flow that depends on cell types (other than wall/non-wall)

2:58 PM

ok

Martin Büttner

finally...

tac in 15:

   ":
@;`),
 (.

where does it start?

Martin Büttner

at the nop

first non-wall in reading order

ok

Martin Büttner

now let me try using the aux stack

3:02 PM

,);""@
"" .(

nvmind

,);""
"" (.@

Won't work for an empty input.

Martin Büttner

,)"";@
:" .(

1?){{=+
  "   "
  ((}:}
@!;

^^ Fib

Martin Büttner

10:

,)";@
:".(

2

Nice :D

That's... pretty screwy

Now, how can we get it shorter?

Martin Büttner

3:11 PM

it's not far from TNO's second to last attempt

@MartinBüttner It seems as if it would keep going straight after . is executed at end of stack.

Martin Büttner

@TheNumberOne nope, there's a positive copy on the stack as well

:D

Martin Büttner

afk for a bit

Hmm ~ feels like it could be a good alternative to ) if you need left turns instead of right turns

1?){{"
  "  =
  ((}+
@!; :

I shaved a byte by doing something stupid with the colon o_O

Martin Büttner

3:29 PM

@Sp3000 interesting

Martin's Hello World looks funky

So many dead-end corridors

Martin Büttner

heh, I wrote that before starting to implement the interpreter

I was very happy that it worked exactly as intended :D

it's probably shortest without loops

just push the numbers and print them with 13 .s

Doing :P

(but I wrote it all backwards on my first try :/)

Martin Büttner

lol me too

33_100_108:}_114_111:}_87_32_44{{:_101_72.............@

Well... that's a start

Martin Büttner

3:44 PM

you can save 5 bytes on top of that

72.101.108::}..111:}.44.32.87.{.114.{.100.33.@

That was a bit more than 5

Martin Büttner

uhhh... right

and I thought I was really clever -.-

You know, without the _s, stack shifting doesn't save anything

Yeah 72.101.108:..111.44.32.87.111.114.108.100.33.@ same length

Martin Büttner

my idea was this:

33_100_108:}_114_111:}_87_32_44{{:_101_72.......
@

XD

Martin Büttner

3:50 PM

@Sp3000 that's kinda lame :D

Well too bad :P

Martin Büttner

my technique can be used to print palindromes though :D

What, palindromic in terms of code points?

Martin Büttner

ah wait yeah I forgot about that

You can print... drumroll... moo :D

And bee

Martin Büttner

3:58 PM

:D

and acc

Basically 99, 101, 111

Martin Büttner

you can probably do more with upper case letters

If you mix case, yeah maybe

67.79.77.
@

Good enough.

Martin Büttner

:D

66.111.
.
@

Even length words not any better :/

So, unfortunately, not much of a use case for your neat trick :D

Martin Büttner

4:11 PM

T_T

Oh, there's a single 6 letter

67.79.99.
.
@

... yeah that's all I've got :P

Here's yes golfed a bit more:

12
.
01

Martin Büttner

neat :)

I really want to post some of these to the relevant challenges, but I'd prefer to kick it off with an answer to a current challenge where Labyrinth is a valid language :/

:(

Martin Büttner

let me figure out your fibonacci

It's pretty similar to yours

I'm still having trouble figuring out the best way to do loops

The junction thing is funky

Martin Büttner

4:28 PM

I'm glad it turned out to be an interesting mechanic :)

Well, it's more the fact that the first-recognised-char-is-the-starting-point that's screwing things over

e.g. I start off with 1? and I want to halt if zero, so that's going forwards

But for nonzero I turn right, and I need to somehow loop back to that junction again

I can't enter from the left because that's where the zero case (i.e. terminate) will be, so I can try and enter from the top...

... except that'll make the program start from a different place

Loop back here how
  |
  v     " <-- not going to work
1?";!@  "
  "     "
  """""""

Martin Büttner

yeah that's true

1?_";";!@
   1 "
   """

I guess that might work in principle, but it's expensive

1
?
1;";!@
" "
"""

Yeah, that's why I have a ) ( in there

Martin Büttner

hm, one longer than yours:

)?
_
1;(;!@
_ "
{ }
{ :
=+}

ah wait

Hmm I'm having trouble inputting two ints

Martin Büttner

4:41 PM

you need to consume the delimiter with ,

Oh... I was doing newlines :/

Martin Büttner

)?
"
"(";!@
{ }
{ :
=+}

Hmm this still doesn't seem to be working, weird...

Martin Büttner

1
?
)((;!@
{ }
{ :
=+}

So.. no need to cater for 0 input for Fib? :P

... oh you shifted it up one

Martin Büttner

4:44 PM

huh?

Before you had F(10) = 89, now you have F(10) = 55

(not that it matters :P)

Martin Büttner

oh, right

so now it's 1-based again

:P anyway I think I might need to catch some sleep, getting late

Martin Büttner

no problem, goodnight!

GL with the snippet, looking forward to HW :)

Martin Büttner

4:48 PM

the snippet is done, I think

I'll either post it around midnight UTC (in about 7 hours) or about 10 hours later

Martin Büttner

4:59 PM

@Sp3000 20 bytes, 0-based:

})?
}
{{(;!@
= }
+}:

and the sequence in 13 bytes:

1
:!\
{ :
=+}

Martin Büttner

5:35 PM

I have implemented (but not golfed) FizzBuzz:

 )::}_101-@
 "       {
 "       :
 "       }_5%122:_117_66
 "          "          "
 "          """"";{"""""
 "                :
 "                }_3%122:_105_70
 "                   "          "
 "                   """"""1_""""
 "                        ;
""""""""""""""""""""""\!{;``{
;                           ;
\""""""""""""""""""""""""""""
 ."

there's now an annotated version in the repo

2 hours later…

Martin Büttner

7:18 PM

Labyrinth now has a page on esolangs.

Martin Büttner

7:45 PM

@Sp3000 I guess it could be useful for quines

then again I'm not a massive fan of adding a feature that is mainly used to simply quines

Martin Büttner

8:33 PM

so, here's a quine... github.com/mbuettner/labyrinth/blob/master/examples/quine.lab

8:58 PM

@MartinBüttner Have you used the code modification features yet?

Martin Büttner

only for random numbers

oh I have an idea how I could use them to shorten the quine

I've golfed the quine by about 50% without that now (haven't pushed to github yet, though)

yeah, using code modification makes it quite easy to put the data on a single line

That's what I'm doing right now.

Martin Büttner

here is the new base:

 101_100_99_98_97_>
_
3
2
.
_
`"
`}";{();@
  1 : _.
  " _
  " (}
5."  ":_10
9    "   :
_~{;"/"=%}
 ~  "
 !"_(

which can still be golfed a lot

372 bytes

I think it gets even shorter if I put a < at the front

Martin Büttner

9:38 PM

287, still very golfable

this could actually be a pretty solid technique for golfing in Labyrinth in general

Martin Büttner

9:54 PM

228

1 hour later…

11:22 PM

@MartinBüttner Nice! Did you take your Prelude idea again or something? :P

Martin Büttner

@Sp3000 it's fairly similar, yes. push character codes of decoder; then the decoder first reconstructs the string that pushes the data and then prints itself using the data

:)

Martin Büttner

and using grid rotation significantly shortened the code :)

and it might generally be a way to work around the "start at first character in reading order" issue

« first day (2 days earlier) ← previous day next day → last day (1394 days later) »

Transcript for

Aug '1527

Esoteric Programming Languages

A room for discussing, creating, using, golfing and discoverin...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile