The Nineteenth Byte - 2015-05-01 (page 1 of 2)

« first day (1553 days earlier) ← previous day next day → last day (3284 days later) »

12:22 AM

do/should we have a tag for something like stack-manipulation similar to array-manipulation

12:32 AM

0

Q: Hover text for chat starboard messages

In chat, starred messages appear in the panel to the right but are truncated to a certain length. Rather than have to follow the link to a new page, it would be useful to be able to hover over a starred message and see the full text. Would there be demand for this? Would this be practical to im...

feature-request chat chat-stars

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

StickStack Numbers code-challenge numbers StickStack is a very simple stack-based programming language with only two instructions: | pushes the length of the stack onto the stack - pops the top two elements of the stack and pushes back their difference (second topmost - topmost) Language de...

We seem to have a lot of particularly long starred messages, so I posted a feature request

2 hours later…

New Sandboxed Posts

2:06 AM

0

A: Sandbox for Proposed Challenges

Digits Digging the Deepest Dungon code-golf numbers Given a 9 element list, the counts of the one-digit positive integers, you should determine how deep dungeon can they dig. The input will describe how many 1's, 2's, ... 9's will be available for the work. E.g. [5,2,0,0,0,0,0,0,3] means the a...

2:20 AM

@randomra To make sure I'm reading this right, ||-|- would evaluate to -2, right?

And to decrement any single number, add |- repeatedly.

Similarly, incrementing is ||-- .

(which means I believe you're missing a - at the end of your 2 example)

@Geobits Yes, those are all correct statements.

should I add more examples?

(and added the missing -)

If I was wrong, I'd say to. But I was right, so I'll just go ahead and assume everyone else is on my level of understanding or higher. That's normally a pretty safe assumption ;)

well, I was wrong with my example :P

Yea, I didn't even notice that until I went back to check the increment loop.

I will add an online executer, but ideone isn't as good as the CJam one

maybe I will try my hand in that (or ask Matrin/Optimizer :))

2:33 AM

Anyone but me :P

2:49 AM

@Sp3000 outputting an integer for me means that in a language's standard representation just like lists, so imo trailing L is fine

I'm building the string, so it would actually interleave the output with Ls

oh, I see

Now to wait for xnor to come along and demolish everything with sub 100 :P

3:08 AM

@randomra Can you add 9,6,3 and 10,6,3 as test cases? It's going to mess up a few submissions (including my own)

@Sp3000 damn, I missed that too in my reference code

... yeah :/

Testing each one, isaac uses the same algorithm so that's wrong too

are my examples incorrect too?

luckily not

More like unfortunately not :( Or someone would have noticed

yeah, but given the current state its lucky

3:17 AM

I can't test PARI/GP, but I think the others (except Pyth) pass these two test cases okay

yeah, I dont know what does Mathematica that it's sub 1 sec

3:45 AM

@PhiNotPi You can pull scripts from Github into Screep.

@TheNumberOne I noticed, although I've chosen not to link my Github account just yet.

3 hours later…

6:55 AM

@trichoplax this is my understanding of Pada: http://pastebin.com/SEP6pYZb
I have a low-functioning interpreter currently

as esolangs down, here is the cached page for it: webcache.googleusercontent.com/search?q=cache:http://…

1 hour later…

7:58 AM

one. more. code-golf.

Martin Büttner

8:26 AM

@randomra That seems quite specific. I think array manipulation works well.

Martin Büttner

8:38 AM

@randomra the digging dungeons challenge could be clearer if you mentioned first that digits need to be able to get in and out of the dungeon.

do you have a test case where the greedy approach fails? (use numbers from lowest to highest and always dig in the deepest possible position while keeping the depths sorted?)

@MartinBüttner I think that is optimal. Do you think it's not?

Martin Büttner

I think it is

Not trivial, intuitively seems true, and I thought quite some time about it (and that seemed to support the case) but don't have a proper proof.

In the other challenge is "Your primary score is the total length of the StickStack programs for the below given test cases. (Your submission is only valid if you ran your program on all the test cases and counted your score.)" unambiguous? I (and other) had many problem with these kind of scorings.

Martin Büttner

I think that works.

You could also put a time limit on it as you usually do.

I can't because here people would use it up exactly so I would have to measure all. I use that only to sort out extreme time complexities.

where standard solutions are well below time limit

without the greedy algorithm I can't see an alternative way to solve the dungeons and that's not good

as greedy is hard to prove

Martin Büttner

8:54 AM

well of course there's brute force

linky ?

Martin Büttner

top two sandbox posts

(sorted by active)

or oldest bottom 2, I have 6 on the last page :D

@MartinBüttner Edited, not sure if it's better though.

Martin Büttner

that's not what I meant

my point is that it's not clear why digits need to ascend at all

(until the last example)

yeah, your examples all use a,b,c,d,e which makes it a bit confusing

Martin Büttner

9:00 AM

@Optimizer ummm, he does explain that

Every digit can remove exactly one block of soil from below itself but it has to reach that position from outside the dungeon and after it removed the block has to leave the dungeon. While doing so a digit **can't descend or ascend more than its numerical value** at any horizontal step.

you know what ? It would be really helpful to get a step by step example

your race of the digits is nice

Martin Büttner

@randomra yep

@Optimizer the abcd... provides that imo with a little imagination

assume I have no imagination :)

Martin Büttner

9:09 AM

@randomra it's only 4/6 steps. you can easily fit them next to each other in the code block without a horizontal scrollbar

maybe will do that, but that's so hard to modify :/

@MartinBüttner how would you bruteforce it though?

Martin Büttner

all possible digit orders (although you can probably show that sorted is optimal), all possible digging position orders.

@Optimizer Thanks, I'm in my digits with numeral_value_property phase. :)

@MartinBüttner and positions are 1..input_size?

yeah, its quite visible ;)

Martin Büttner

@randomra I think using the already dug positions ±1 is enough

9:18 AM

@MartinBüttner I don't(? grammar...) see neither that nor the sorted digit order trivial though.

Martin Büttner

why would you dig somewhere that's not next to a hole you already have?

9:33 AM

Aww plannapus took the cake numbers

1,2 comes and need a 2 deep there while the old place needs a 1 only (ofc this never happens but it's not trivial)

for me at least

(Wait no they didn't, all good)

Martin Büttner

just finished itoa in Fission

     /"0"/    \
R'x[@%  //    /
     \$@{   / \
   @    \     <\
   $       ;  ;}+++++++++L
   %@       A {/
   M     \@<+@/
   ~       @
  ;Z       K Y 0'L
         \  !A   /

now I can actually start working on the sequence :P

(which is probably gonna be even worse, because the one I picked has a recursive definition)

10:04 AM

0

A: Sandbox for Proposed Challenges

Finding pairs for letters code-golf simulation? ascii-art?? (do we use ascii-art for input?) You should write a program or function which given a string as input outputs or returns a string or list of the lowercase letters in the order they find their uppercase pair. The input is a string cons...

will post if no comments

no ascii-art

Martin Büttner

we do use ascii art for input, but I'm not sure this qualifies as ascii art.

do we?

Martin Büttner

sure

nether portals?

@randomra maybe you can use grid

10:24 AM

thanks, posted

11:00 AM

@randomra in the first example, d is only 3 distance away from D, right ?

also, what if input is a square and a is not able to find A forever ?

I've got an idea for a challenge...but I just want to get some thoughts on it!

Input: An image which is not centralised (eg, the union jack in the top left of a png). excluding the object (union jack) the whole canvas is blank (white or transparent).

Output: The image, centralised into the canvas

11:17 AM

what is center?

(x/2,y/2)

Image will always be > 100,100

example input

user image

@Optimizer It's american for centre

Example output

user image

@Joshpbarron Is it the bounding box of the non-white pixels that needs to be centralised?

Or the average non-white pixel position?

@trichoplax american is a language ?

11:22 AM

@Optimizer Sorry, US English - I was being lazy

@trichoplax I think the bounding box would be centralised

If i understand correctly:

square select so that you enclose all the non white pixels

Will the bounding box and the image have width of the same parity (odd or even)?

selection X co-ord = (canvas X dimension - selection X dimension) / 2

why not center of gravity ?

No, however always round down.

11:24 AM

treating each image pixel as unit weight.

That makes it unambiguous. The other option I was going to ask about was offsetting by half a pixel and anti-aliasing

@Optimizer I think that might be good for round 2 maybe?

@Optimizer That sounds more involved, but also interesting. Would the pixel weight be 0 for white, 1 for non-white, or would the pixel weight increase for gradually darker colours?

@Joshpbarron round 2 will then be closed as a dupe

@trichoplax Anti-aliasing might be good for a bonus?

11:26 AM

@Optimizer I don't know - it sounds pretty different, and you certainly can't use answers from round 1 for round 2

Seeing as it is my first question, i think i'll keep round one simple

@trichoplax I say uniform, as colors don't actually mean anything. 0 for white/transparent/non-image-pixels/whatever-you-call-them

@Joshpbarron I'm not a fan of bonuses, as it means there isn't one clear correct requirement

Ok

Thanks for your help by the way guys.

What do you think of the challenge initially?

@Joshpbarron Thanks for working on a question - it's what we need :)

11:27 AM

I'll chuck it down on paper, and put it into meta

The other thing I would consider is how the image is to be encoded

To be inclusive of as many languages as possible

I don't really understand what you mean encoded, do you mean the output?

If you make it png then the code to process it will be longer than if you allow an array of integers, or pbm

some languages here cannot read or output images

Ah ok

11:29 AM

I mean the way the image is stored (input and output). Sorry encoding was the wrong word

Some people specify a format (eg. pbm). Other people say "You can choose the format that suits your language best"

Posting in the sandbox is the best way to get feedback, as you only get feedback from the people who happen to be in the room in chat, whereas leaving it in the sandbox for a week means more people see it

Thanks for your tips

a week might actually be too much, generally 4-5 days work

Here's a handy place to run things past people, but I expect you'll get things we've missed in the sandbox

FYI, I'm already expecting @MartinBüttner to come along with some fancy retina

:D

retina and image parsing = forget it

11:33 AM

But retina and a string of numbers?

If anyone can force regex to do the wrong job though...

Thanks again...I'll see you on the other side of the sandox!

Sandbox aswell!

:)

(you can edit in here by pressing up or clicking edit on the comment)

Thanks :)

12:05 PM

is it okay to answer this question ?

Why wouldn't it be?

@Optimizer Do you mean a second answer?

no

s/is/was

Ah OK

Looking at your answer @Optimizer

I think its safe to assume that the number is already an int?

to save one byte?

12:12 PM

r reads as string

oh ok

Ignore me :)

@randomra That makes sense now :)

@Optimizer

Would you count my answer as 3?

6

see his example of 85 tokens

His example isn't really tokens, it's more everything :)

12:26 PM

tokens counting is never really clear

so its usually best to do what OP says

Got it

is there no way in Pyth to convert a float to int ?

where is orlp now who is always boasting that Pyth is better than CJam ?

(^ link not visible to low rep users)

@Joshpbarron you should also change your code into a function/program. right now, it assumes input in x but its not much of a difference to convert it to an anonymous function

log2[#+1]@& (or something like that)

I don't know Mathamatica, was just trying to be smart :)

@Optimizer "If a letter newer reaches its pair it shouldn't be outputted at all."

@Optimizer d sees D right away and 2 steps away as noted in the Steps line

12:42 PM

:21407825 It is. input, math, print, math, log, int, input, 2.

Oh, never mind.

@Optimizer s?

really ?

"Reduce + on list, cast to int on other. Real part on complex."

damn.

@Doorknob the solution is incorrect anyways

@Joshpbarron according to the new counting method, your current post is 3 tokens long, but its still not a full program.

Deleted anyhow

Martin Büttner

1:16 PM

atomic code golf is such a broken challenge type.

10

true

Martin Büttner

@Optimizer doesn't the OPs approach work in CJam for 3 tokens?

ah no

it's the other way round

never mind

I can do 2 tokens - put the code in string, evaluate it. but then using strings is always tricky in atomic

Martin Büttner

he also says you need to count the tokens of any code that you encode in a string

@Optimizer btw, if you use 2b, it's the same number of tokens but one byte less

@MartinBüttner As is exemplified by the OP's self answer.

1:31 PM

People will always find loopholes around atomic code golf. Banning strings? Okay, I'll use comments and read my own file. Banning reading own file? Okay, long variable names. Banning unreasonably long variable names? Well I have this 2000-digit integer which encodes what I need...

@MartinBüttner btw picked a new sequence for Fission yet?

Or still doing the real magic numbers?

Martin Büttner

@Sp3000 chat.stackexchange.com/transcript/message/21406007#21406007 and the following message

I see :)

(btw I gave up on the unique languages thing - it was fun, but I'd like to post interesting sequences too)

Martin Büttner

I'll try to finish it over the weekend, but it's gonna take a while and don't want to spend too much time on this

it's a really fun sequence though, and I'll probably sandbox it as a separate challenge once I'm done implementing it in Fission.

Ahaha okay :)

Martin Büttner

1:46 PM

I wonder if something like a reverse atomic code golf could be fun: you get to define a fixed number of operators (or functions) on numbers - say 2 unary, 4 binary, 1 ternary. and a series of arithmetic tasks (more than than there are operators). and you may only use these self-defined operators to solve them. score is then either plain code golf or number of times you've used the operators across all tasks.

(the definition of the operators would be free in either case)

so it would be something like this except you get to choose what your operators do.

Hmm that does sound fun, but you'd have to come up with quite a lot of tasks

Martin Büttner

most of them could probably be simple things

actually I think this still has the same flaws as normal atomic code golf

I can still define a unary operator which converts its input into some string and evals it.

Oh, yeah that wouldn't work too well would it

Or alternatively a binary operator where one is input and the other determines which problem it solves

Martin Büttner

yeah

I guess this wouldn't be much of a problem if you scored the rest as normal code golf though

because base 10 isn't really all that efficient

I guess you could restrict the operators themselves to being combinations of certain arithmetic operators, if that helps

... actually I don't think that fixes it. Hm.

1:57 PM

I think that sounds like it would be better suited to math.se imo

2:34 PM

0

A: Sandbox for Proposed Challenges

You call it alignment, I call it Code-Golf. Doesn't it annoy you when images aren't correctly aligned? Rules: 1) You are to generate a program. 2) Standard Loopholes are a nono. 3) Input will be a file, however whether it is passed as a file or a list of numbers is up to you. 4) Output will...

Question posted, feel free to edit/comment.

Maybe I'm just not seeing it, but what makes the output aligned and the input not aligned? Aligned to what?

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

You call it alignment, I call it Code-Golf. Doesn't it annoy you when images aren't correctly aligned? Rules: 1) You are to generate a program. 2) Standard Loopholes are a nono. 3) Input will be a file, however whether it is passed as a file or a list of numbers is up to you. 4) Output will...

Doesnt really work on a white background. I'll edit the question to explain better

2:54 PM

@Joshpbarron Your challenge needs some expectations of what constitutes a correct answer. How will you judge whether an image is "aligned"?

Working on it as i type..well...not exactly, but you get my drift.

Also, code-golf is hyphenated in the tag because tags can't have spaces. Doesn't need a hyphen in your title. This is obviously super important.

Super important.

Super-Important?

Super-Important

2:59 PM

<marquee><blink><i><b>SUPER_IMPORTANT</b></i></blink></marquee>

Also, people react more positively to positive statements. You might try:
Before: Doesn't it annoy you when images aren't correctly aligned?
After: Do you enjoy basking in the glory of images that are correctly aligned?

2

On a more serious note: Are your input images guaranteed to be symmetrical?

"Let's face it, aligned images are like little gifts from Heaven. Welcome, but unexpected."

I remember when I first learned HTML and the marquee tag was so cool.

Ah yes... I remember geocities also.

I don't even think we were taught CSS back then

nothing quite like a static web page and copy pasta headers on each page.

3:01 PM

Especially if it had a black background with yellow text.

I made mine in bright teal.

background.

o.o

Sorry, "AQUA"

Coincidentally, W3schools actually has a decent representation of how bad it looks: w3schools.com/tags/ref_color_tryit.asp?color=Aqua

@Joshpbarron Standard loopholes are disallowed by default. Anyone abusing them will be downvoted into oblivion if you point it out in a comment. Sometimes they get straight up deleted as Not an Answer. So you don't have to ban them in your challenge if, like me, you like to keep it short.

It's a miracle that I do UI/UX work now ha ha!

3:03 PM

Don't show any of your work to Optimizer

i was here yesterday

So was I.

that was both your faults for talking about redwood trees.

Seriously. I'm afraid to (jokingly) call anyone rude now.

I am usually super mellow about being offended. Unless it's a betrayal of trust. Then I will never forgive you >:(

3:06 PM

It all begins with redwood trees. The next step is raiding jiggalos on 4chan.

Better watch out, you'll get chat-banned for talking about raiding ;)

Dare I say it, did w3schools actually design their website to look nice?

@Geobits You just said it!

It's a MIRACLE

it's using CSS3, what sorcery is this?!

Weren't you a UI/UX person ?

wow, w3schools redesign

3:08 PM

it looks a lot nicer

i wont say it looks great because that betrays the internet

Wow... it doesnt burn to go there anymore!

now to look at the archive.org stored copy to see how bad it used to look

yup it looked bad

Question Updated

Tear it apart please :)

Ah, that's probably why. Before it operated on the old 1024x768 page width.

Now it fills to however wide you wish it to be.

#progress

Oh, center aligned on the canvas/background. Got it. You might want to use the word "center" at some point to make it more immediately clear.

3:16 PM

"Input will be a file, however whether it is passed as a file or a list of numbers is up to you." Wat

what do you mean, numbers?

anyone posting any code-golf soon ?

I did think that..its more about centralising images really. (But spelt right ofc)

@Compass

If you wanted to, you could have a string of rgba values. But you would have to decode it.

Perhaps best if i just put " input is a file "

"input is an image" ?

@Joshpbarron I think a more common term would just be "centering". Then you also don't have to bastardize the spelling :P

Also, a common term for the "square around the object" is bounding box. Which sounds better to me, because even in your example it's not a square.

Thanks

Anyway, back to tearing it apart :p

3:38 PM

CEO just piloted a toy car with a camera attached to it into the developer Pit...

Everyone immediately starts to act like they are working hard and that they don't care about the car as soon as someone points out the camera

You have a strange workplace. Has anyone ever told you that?

you guys don't have normal cctv cameras in office ?

I probably only talk about the strange bits. Nobody cares about the normal bits

@Optimizer Yes, we have cameras everywhere.

No cameras here :D

But the ceiling is so dark that you can't see them from the floor

3:44 PM

haha

Martin Büttner

@Rainbolt you must have a nice office

@MartinBüttner I think you could find a smaller image of that if you tried hard enough :P

Martin Büttner

is it that big?

It's tiny. It took looking at the filename to figure out what it was.

Martin Büttner

I can't tell which of your two messages is sarcasm

oh

3:50 PM

@Geobits I found one

Martin Büttner

@Optimizer why don't you just 2P*mr? saves two bytes

@MartinBüttner ah.. that would make it a double automatically ...

Martin Büttner

also, I still find it weird that you do 2d and 1d instead of just 2. and 1. :P

1. It seems like this is the first point of a list
2. which is incomplete

@Optimizer Ooooh, that clears it up. Moria indeed.

3:52 PM

One black pixel to rule them all. One black pixel to find them.

@Geobits moria from one piece ?

@Optimizer Even if I'd watched One Piece, the answer would be no. That pixel looks nothing like him ;)

That guy looks like the Earl from D Gray Man

Moria ?

yea

3:56 PM

yeah, just a bit more curvy..

moria was a nice arc

one. more. byte.

if only } worked like ]

code-gold!

I figured it out

we can use puzzle pieces as our badge icons

puzzling.SE

what about it?

it's still in beta, it doesn't have anything on us <_<

puzzle pieces belong there

do they?

4:11 PM

like we are graduating tomorrow

puzzle pieces belong on board games!

Can we just get keyboard keys then?

but puzzling is about puzzles

PPCG is not about puzzles

[?] <- the question mark key

no

why

4:11 PM

because no

Riot.

Is this a duplicate?
Reorder tuples with minimal number of deletions so there would be an increasing sequence in every element position e.g. (1,2,3),(2,3,2),(5,4,3) => (2,3,2),(5,4,3) with 1 deletion

in some time complexity/limit ?

sure, vastly underspecified, just interested should I write out details at all

(esolang is back!)

The Pada language has a neat mechanic but it seems to be vastly unusable for any programming. :/

Martin Büttner

4:43 PM

@randomra I thought that is why esolangs are fun?

how do I use dupe-hammer ?

Martin Büttner

@randomra Sounds like a challenge Lembik would write. :)

@Optimizer you don't, because you won't have gold until tomorrow morning :P

@MartinBüttner I already have it. su**er! :P

Martin Büttner

wut?

go see

Martin Büttner

4:44 PM

that's odd

that's just me

Martin Büttner

I thought tag badges were always given out at 3am UTC

Optimizing on time

now tell!

@MartinBüttner i got within 30 minutes

Martin Büttner

it's no different

just cast your close vote

then how do I normal close vote ?

(damn there is no question to test to)

Martin Büttner

4:46 PM

you can't normal close vote any more

(well not for duplicates)

ohhhhh

@MartinBüttner yeah, you can see it as matrix row permutation, but there are some good algorithm to solve it

with great power, comes great irresponsibilities!

Martin Büttner

@Optimizer I'll hold you to that :P

@randomra yeah it sounds interesting. I'm not aware of it being a duplicate.

hi all

4:48 PM

@MartinBüttner well, I might choose BF quicksort over Pada GCD (to implement it)

@MartinBüttner cool

Martin Büttner

talk of the devil...

@MartinBüttner its all your fault

which devil?

oh.. me :)

Martin Büttner

@randomra "These are useful because both "?" and "*" don't touch the 3 commands already in the bit tree, so if you need nothing to happen then using those no-ops are good." I've been re-reading that sentence so often and trying to make sense of it... could it be that commands only fall one level per tick and are pipelined like that? (as opposed to falling all the way through before reading the next command)

@MartinBüttner yes, that is the case though only that sentence implies it

Martin Büttner

4:55 PM

okay, that makes it really messed up :D

@MartinBüttner even the top users code-golf page got updated. usually it does not update at this hour

that is the smallest difficulty imo

just some extra padding

Martin Büttner

@Optimizer the definitely only updates daily

@MartinBüttner yeah, but not at this hour of day for me

Martin Büttner

oh

"got updated"

I read "not updated"

maybe they changed their cronjob for tag scores.

4:58 PM

@MartinBüttner that does not fit grammatically anywhere in that sentence :P

Martin Büttner

I know, but your messages are regularly missing words, so I've become used to interpreting your grammar a bit more loosely :P

@MartinBüttner yeah, the new user profile page might have brought upon that change. It was congratulating me on code-gold too

every jump needs a ton of plus work before and after and all is relative so subject to change
and jump is the only way to do anything as you have only the (0 1 not) functions and you work one bit at a time

@MartinBüttner I don't miss words in between. I only trim them.

« first day (1553 days earlier) ← previous day next day → last day (3284 days later) »

Transcript for

May '151

The Nineteenth Byte

The Nineteenth Byte: General discussion for codegolf.stackexc...

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