@overactor no, I think we should wait until we have a normative spec, and a compliant reference interpreter

@MartinBüttner that is a good point

Are there any differences between the two interpreters at the moment?

I intend to write my own interpreter over the holidays, although I'm not sure any more if it'll provide any new value over es1024's implementation

@overactor I don't know. does the Python one support cylindrical boards?

also the JS one doesn't support includes and a couple of other things

@MartinBüttner Ah no, it doesn't

nor does it support output as decimal values

huh, does the JS one?

JS has standard library includes though

yes

btw, standard library stuff needs to go in the spec as well. and we need to figure out the absolute minimum devices necessary to implement all others

Those are the 4 extra options the JS version offers

@MartinBüttner Is figuring out the minimum strictly necessary?

(lol @ Insomnia - one day the Stack Snippets question will be filled with all the weird esolangs due to people writing interpreters for restricted source questions)

@overactor it seems important for implementations

it would be nice to know, how much of Marbelous can be implemented inside Marbelous

@MartinBüttner Sounds like a valid question for ppcg

implement the marbelous devices using a minimal set of marbelous devices

total code size as tie breaker

yep, that could be fun.

even if we figure out that minimal set ourselves, it might be fun to do a pure code golf to implement the rest in marbelous in terms of these.

synchronizers and portals are obviously impossible

yep. and you need some sort of arithmetic.

you don't need cloners or deflectors or trash bins

but you might need at least 0=, I think

++ can implement all +n and -n devices

maybe that isn't even necessary

++ can be implemented with marble merging

I', thinking too

@MartinBüttner oh right

but you might need a bit accessor or something

but maybe not even that

or bit shifting

:In
01}0
{0{0

that's ++?

=0 or some other comparison will be necessary

@MartinBüttner yup

yes

you could probably even implement the PRNG in marbelous

btw, that needs to go in the spec (what kind of PRNG should back the ?n devices)

@MartinBüttner or it could be left undefined

hm yeah it could... but I don't like bad PRNGs :D

Which doesn't exclude an interpreter using an xkcd approach

hi @PeterTaylor are you about?

@PeterTaylor there is an interesting update :)

@MartinBüttner I agree though

How would you implement a PRNG in marbelous?

once you've got arithmetic, that shouldn't be a problem, right?

@MartinBüttner but all arithmetic is deterministic

and you can't save a state

how do you think other PRNGs are implemented?

hm, right

not being able to save state is a problem

nowhere to persist a seed

@MartinBüttner We could get around it with a two cell wide function

okay, good point, ?n needs to be written outside of marbelous

@overactor yeah, I don't think that's a good idea

@MartinBüttner for users of the functions, the functions should be identical to the devices

You can split marbles, right?

@FryAmTheEggman There's a device for that

can trivially be implemented witha function too

Couldn't you make a game-of-life "gun" style board that spins a seed marble around and spits out a value each time it gets input?

@MartinBüttner ?? would be more practical

@FryAmTheEggman The problem is that boards don't persist in between being called

Ah, right, I just noticed that :(

Obviously, the !! device is also impossible

right?

@overactor I'm pretty sure it is... unless you can silently crash the interpreter :D

@overactor true, then you can get ?n via simple modulo

maybe not quite so simple

but it will be possible

@Marbelous, randomness might be implementable depending on what the spec says about multiple portals with the same index

lol

?? Random Sets the value of passing marbles to be a random integer between 0 and the value of the passed marble, inclusive.

oh okay

even better

?n is not an actual device

?1, ?2, ... ?Y and ?Z are

How we solve the portal thing is still not clear

@MartinBüttner the JS interpreter does it randomly

@overactor I meant those by ?n

:R1 #Returns a random marble between 0 and 1 (inclusive).
}0@0@0
?0++{0
@0{0

@overactor I don't even remember what preference I had when we last talked about it

I know mine, but no one agreed with me

you wanted closest or something?

Note that I used ?0 there, which is trivially implemented as:

:R0
00}0
{0

@MartinBüttner Yeah, not sure if I agree anymore

I was also thinking of making a marbelous android game

yeah, that was also my next plan after writing a JS interpreter (browser-based game though)

whatever happened to asm.js?

@MartinBüttner Were you planning on having functions in the game?

it seemed like a good idea

@overactor maybe you could unlock functions by writing them in earlier levels?

@MartinBüttner I can also imagine having two boards of defined sizes to solve a problem

or three

yeah, possibly

I was also thinking of having two modes

one would be kid friendly

(well, "kid")

and would be a fun way of being introduced to programming

(well, "programming")

wow, you're really building the suspense for what the other mode will be...

@MartinBüttner I am, I thought it was self explanatory?

More extended

I'm not going to bother kids with bit fetchers and bit shifts

Obviously the puzzles would be easier too

And possibly, the marbles only go from 0 to 9

I see

@user2179021 But you're going to make me beg for details? Why so coy?

@PeterTaylor :) See the answer of KFabian mathoverflow.net/questions/168474/…

although I don't get a crucial part of the proof .. it does look elementary

@Rainbolt yeah, I totally agree

I know it does.

I can't believe I missed the chat about Minecraft yesterday >_<

@user2179021 It doesn't necessarily help for partial circulant matrices, though. This looks like a way to find matrices which score at least 1, but there are a lot of Lyndon words which score better than 1 but less well than the optimal score for their length.

@PeterTaylor right... I have no idea yet if it works for partial matrices.. but if it is correct it shows that at least this case is not NP-hard

@PeterTaylor maybe it can be extended to partial matrices?

to be honest.. I am not even sure it is correct yet

I mean as is

I am trying to reprove it

haihai

i was gonna play rollercoaster tycoon (the original one <3) but the browser was still open

I think I might be able to do Gottfried Leibniz in marbelous

@PeterTaylor Do you understand why the sentence that starts "If a solution exists then all [...]" is correct?

that seems unclear to me too

@MartinBüttner How do I count bytes for ideone.com/C3cNpM ? I don't know how to handle non-printable ASCII...

@MartinBüttner oh.. did you get a chance to write the mathematica script we discussed?

for my sand box puzzle?

@Sp3000 filesize of utf-8 file ;3

@user2179021 I did, on Thursday.

Well the thing is if I save it like that it's 1.3k bytes...

unless you're talking about the improved version I mentioned, no I haven't done that yet

@MartinBüttner oh!

@Sp3000 I tend to use mothereff.in/byte-counter

@MartinBüttner where did you put it? I am sorry I seem to have missed it

1376 :/

@user2179021 start reading the transcript here chat.stackexchange.com/transcript/message/18702438#18702438

@Sp3000 but surely, that's with escape codes?

oh thanks!

@user2179021 No.

Yeah, I'm just not sure how to enter the actual codes into the byte counter

@PeterTaylor ok.. maybe it is just wrong

which would be a shame

@Sp3000 print them somewhere and copy them over?

@MartinBüttner that's pretty cool

"EOF while scanning triple-quoted string literal" hmmm

duplicate alert? codegolf.stackexchange.com/q/41872/8478

we've had very similar ones at least

Think so, surely

there's at least codegolf.stackexchange.com/q/38032/8478 and codegolf.stackexchange.com/questions/32085/…

It's closer to codegolf.stackexchange.com/q/1328/194 and codegolf.stackexchange.com/q/10790/194

ah yes, especially the second one

@Sp3000 so did I

I answered, because CJam isn't allowed to compete in the old one

Ahaha hifive

Wow ... Peter can single handedly close code-golf's!!

I just thought the date part was interesting

as can Howard (and marinus I think)

how do you get code-golf gold

:o gold badge

searches

1000 upvotes on 200 answers

100 upvotes, no ?

no

100 upvotes on 20 answers is for bronze

1000 total score means 1000 upvotes?

even though each upvote gives 10 score

it means 1000 upvotes - downvotes

score is answer score, not rep

is there a way to see badge progress ?

check the top users of the tag

and yes, howard and marinus : codegolf.stackexchange.com/help/badges/89/code-golf

u need 26 more 0 vote answers

working on it ;)

@FryAmTheEggman Nice catch :P

i am nowhere near it :D

@Optimizer Only as a duplicate. I keep forgetting it and being surprised.

@PeterTaylor I think that is the best thing!

(forgetting and then getting surprised by your powers)

@Sp3000 Also I think using import time is shorter

Probably is, I forgot how Python's time modules worked

mothereff.in/…

Seems to be 2 chars shorter

these badges are not listed in the badges page ?

yes, there's a section for tag badges

and when you click on those, you get a list of all tags that have badges

It's even shorter if you don't do as d

@MartinBüttner Ah, sneaky . Thanks.

Oops forgot to even check that :P

@Sp3000 Oh, I left in the M==1 :P

...oops

Only 148 bytes to tie with CJam!

which question ?

this

oh, I have <400 answer

i love it when my queues do this

@Sp3000 I think this works: Q=[0,[(D<6)*(D+7),D-24][M>2]][M%11==1]

[(D<6)*(D+7),D-24][M>2]*(M%11==1) golfing it down

:P

Nice :)

Hmm it gives negative numbers for some of them

If I wanted that I guess ((D<6>M<2)*(31+D)+(M==12)*D)-24 would be shorter, but negatives are breaking something

Rats, I thought I tried it with december 20th, etc.

Oh, I found out why :/

Early december days

Yeah

Hmm maybe if there's a better way to fix Q-1 than max(Q-1,0)

Because currently the saving from Q= cancels out with that

~-Q%12?

@Sp3000 seems to work...

Hmmm I think I have another way of saving chars though, you gave me an idea

realization:

(oh oops nope nevermind)

the "add without addition" question is just trying to figure out Manhattan distance between (0,0) and (a,b) without knowing that a+b does exactly that

That "realization" was very well timed :P

I've got to go soon, but would a bonus/bounty on my Christmas challenge work for the objective to make the code look like a Christmas tree?

That may not be understandable English...

So turning it into source-layout, right

Er...

Maybe not then...

Yeah I don't think that is the best idea. It seems just tacked on.

@FryAmTheEggman If I do M>11 I save enough chars to make myself not have to do ~-Q%12 :D

Yeah okay.

PPCG would be rather terrible if there was no chat

@Sp3000 I just noticed something

You can move "twel" to the first index and lose the ~-

:o so it'd save... two chars and become 10% more unreadable!

Great success! :D

... oh actually it doesn't :/

Q=max((M>11)*(D-24),(D<6>M<2)*(D+7))
Q=((D<6>M<2)*(31+D)+(M==12)*D)-24

... wait

Nevermind, yes 2 chars

well the %12 and shifting costs nothing, right?

... though tbh if I move twel to the front and change Q-1 to Q...

I can keep the code readable :D

Blasphemer!

IndexError: list index out of range ... nope

Uh, the code you posted doesn't have the %

Is that intentional?

That's not the % one yet

I was trying to golf it down to see if I could avoid it

But -2 chars it is

Oh :P

Now it's got %

@BetaDecay are trailing spaces on the lines allowed?

Man, Martin just golfed out enough chars to beat my last half hour :P

I'm waiting for Optimizer to kill both of us

Maybe if I could figure out compression I'd be below 500

lol, he is pretty good at these

But I don't think it works that way D:

I don't think CJam has any built-in compression

(apart from Unicode packing, but that won't help here)

or maybe it will? wait...

b64encode is so useless :/

use encode("zip") before b64

I don't think it saves enough to warrant the imports/decompression calls...

Yeah I got 680 :P

@Sp3000 hm, I can get it below 500 with byte packing, but I have yet to find a base that doesn't have messed up characters that don't survive copying

Ah :/

Does giving the bits themselves count for questions like this?

Hmm does Python's % operator handle repeats?

e.g. like "{0}{0}".format("ab") gives "abab"

Not as well as that

but you could do "%s%s"%(ab)*2

"%s%s"%(("ab",)*2) hmmm

I think that might save a few chars

meh... copying turns \r into \n and nsbp's into spaces

Yeah I did that a couple times in that detention question

I did think about that earlier but I wasn't sure how much it'd help

Time to find out

@MartinBüttner Yep, I'll add that in

(3 chars, it seems)

@BetaDecay Surely you don't mean to allow leading spaces? :P

@FryAmTheEggman Oh yeah, that would just be terrible :/

@Sp3000 You could do something like removing " the " from some part of the string after/before each of the other replacements to save some more, I think

@Sp3000 Well 3 characters is better than none

Yeah that saves 3 more

Dunno if there is a better string

I think " the " is a bit too short to save chars

I tried it and it seemed to work..?

Oh, maybe you're doing it differently?

"Well%sWhen%skids start singing and the band begins to play\nOh%sSo let%sbells ring out for~"%((" I wish it could be~ every day\n"," the ")*2)

I did that and got 668 total

Yeah

Oh, I was looking at unupdated count :/

Ah, right :P

I wish Python had a better way of doing .replace multiple times...

Whee, sed golfing is fun

D: 12 days of Christmas in sed?

@cjfaure now I wonder if that would even technically be possible, within the space limits of a Minecraft world...

@Sp3000 hm, makes me wonder if we have such a kolmogorov-complexity task

@FireFly if you modified gcc's optimizations to turn some of the jvm data into physical wiring instead of using RAM (which is space-costly) and used a full 256 block tall world, you should be able to fit it

the hard part is circuit sync

Electronic Super Joy has a really weird distribution of achievements.

on the first playthrough you get 4 achievements for bosses and game completion. then there are 3 special achievements for individual levels (like survive 40 seconds, or clear the level without killing anything), which are fairly tough. then you've got the usual "collect all stars", "beat all time runs", "clear each level without dying". and then you can unlock 5 bonus level,s and you get one achievement each for just completing every one of them. o.O ...

Ahaha, on to a new game, are we

like... just unlocking those 5 levels is pretty tough, but then it's basically 5 free achievements ^^

@Sp3000 well...

I might revisit VVVVVV afterwards...

but I think I've come as far as I ever will...

Not that, was just wondering how many games you were planning to achievement rush

I might get the flip mode achievement at some point, just for the fun of it

@Sp3000 my steam library is fairly vast... :D

and there are a lot of games in there I haven't played yet

anyway, ESJ is really fun. some parts are pretty tough, but it's not as hard as Super Meat Boy.

Ahaha k :P

(wow, I've played SMB so much, and I've still got so many achievements left... I still haven't beaten the first Cotton Alley level)

Well I have too many from Humble

So I don't think I'll be playing those any time soon