The Nineteenth Byte

12:55 AM

if i haven't made a mistake, the number of valid brainfuck programs of length n should be given by f(n)=if(n,6*f(n-1)+sum(i=0,n-2,f(i)*f(n-2-i)),1)

TheNumberOne

1:06 AM

I keep figuring out ways and then proving myself wrong :(

Mitch Schwartz

i've conceptualised it as: the first character is either a [ or one of the 6 non-[] characters

if [ then we can iterate over the possible locations of its matching ]

and recurse on the enclosed part and the tail

and if non-[] then we again recurse on the tail

TheNumberOne

I'm currently creating a grammar specification for BF. Each bit will decide where along the grammar to expand.

a ::= c*
c ::= d | '[' e
e ::= ']' | c e
d ::= f | g
f ::= h | '-'
g ::= i | '+'
h ::= '<' | '>'
i ::= ',' | '.'

Alex A.

1:23 AM

@Dennis ^

TheNumberOne

@AlexA. Is it still in the store or did you buy it?

Alex A.

Who knows

Maybe I made it

@Dennis @Dennis ^^

TheNumberOne

1:40 AM

This has to be ILLGOL. — Dennis 17 mins ago

ILLGOL is only a stub :(

Sp3000

@TheNumberOne That could only happen if Dennis, aditsu and I (+ a few more) all took a simultaneous holiday

Mitch Schwartz

2:16 AM

my breakdown should lead naturally to a ranking scheme such as (untested): let f be as above (counting programs of length n) and g be the desired function that accepts a program and returns its rank, and define helper functions a(x) = sum(i=0,len(x)-1,f(i)) and h(x) = g(x) - a(x), i.e. h maps the programs of length n to [1..f(n)].

take h(e) = 1 where e is empty program; arbitrarily map h(c) for c in '+-<>,.' to [1..6]; h(c+x) = (h(c)-1) * f(len(x)) + h(x) where c is non-[] head and x is tail; h('['+x+']'+y) = 6*f(len(x+']'+y)) + sum(i=0,len(x)-1,f(i)*f(len(x+y)-i) + (h(x)-1) * f(len(y)) + h(y)

Dennis

@AlexA. Thanks! Why 80th though? :P

Mitch Schwartz

sorry if i made a mistake but something close to that should work at least, i don't feel like looking at it more atm

TheNumberOne

@Dennis Your problem has fried my brain :(

Dennis

Sorry...

@TheNumberOne bfy.tw/1Urc

:P

TheNumberOne

︻╦╤─<           -     google

Mitch Schwartz

2:22 AM

there's a missing ")" near the end, after f(len(x+y)-i) to close the sum

Dennis

Thank Foo. I'm not sure EoL FIN would have made me suspicious enough, but NB had to be comment syntax.

Oh, man, this would have been quite hard to figure out:

esolangs.org/wiki/Stasis

No interpreter though.

PhiNotPi

3:01 AM

I'm really worried that TPLQ will beat my record for most answered thread.

Dennis

It still needs 85 answers to catch up.

Dennis

3:15 AM

♫ Whenever, Wherever... ♫

Sp3000

3:48 AM

Dennis is catching up pretty quickly :P

Dennis

Knowing you, you'll have cracked 40 when I wake up tomorrow. :P

Alex A.

@Dennis Because that's the cake I found a picture of. So happy 80th birthday, Denis. :P

My goodness, +178 rep for suggesting people use pretty pictures in popcons.

Dennis

My suggestion would be to always do math in base 13.

Alex A.

I think that's more of a general life suggestion.

Dennis

Very helpful for popcons though.

Optimizer

4:01 AM

@Dennis♦ Congrats for the diamond! Finally!

Dennis

You're worse than Geobits and his graduation promises... :P

Alex A.

Haha

@Optimizer There's a space before the diamond, plus the diamonds don't show up in chat pings. :P

Optimizer

@Dennis at least I am not making up your diamond story all together :P

@AlexA. :P

Dennis

They still have to initiate contact, so we'll see what comes earlier.

Alex A.

@Optimizer Oh you.

Optimizer

4:04 AM

Alex A.

@Dennis Close encounters of the third kind?

Something like that.

Optimizer.ping(to="Alex", type="invisible", repeatedly=True)

Optimizer

4:06 AM

cannot call ping on locked Optimizer

Alex A.

Alex.printEyes(type="disapproval")

ಠ_ಠ

ಠ_ಠ

ಠ_ಠ

lol

:P

4:08 AM

fail

Yep

♦_♦

desperate

Dennis has diamond eyes

Dennis

@AlexA. Did you hear? The language I was so desperately searching for was bc.

Alex A.

4:10 AM

I didn't hear!

The one that wasn't Haskell or Curry?

Read?

ರಿ.ರಿ

Yep, that's the one.

Did you crack it?

Nope. I had tested it in bc, but bc programs require a trailing newline which got lost in the copy-pasta.

Alex A.

4:11 AM

Haha pasta.

Sorry to hear that you didn't crack it.

But I bet it feels good to finally know what it is!

Dennis

I was mad as hell when I found out, but at least I can stop compiling random Haskell derivates.

Alex A.

ಠೆ.ಠೆ

Haha no more wasted days and sleepless nights

Compiling all the Haskells

ಠಿ.ಠಿ

Two sites I should really bookmark: disapprovallook.com and the Kannada alphabet.

ಝರ

Dennis

@AlexA. ^^

Alex A.

Hahahaha

I think you got your images and text backwards :P

Dennis

Better?

Alex A.

4:25 AM

Hahaha much better

You know what's the worst thing ever?

SAS has three meanings for *.

Dennis

So does Ruby.

Alex A.

Sometimes it's a comment, sometimes it's multiplication, and sometimes it denotes a cross-tabular table.

Is it sometimes and comment and sometimes not in Ruby?!

Dennis

Nope. In Ruby, it's multiplication, string repetition or join.

Alex A.

Oh, okay

That's normal

Something should never sometimes be a comment and sometimes not.

That's terrible!

Dennis

Can't argue with that.

Alex A.

4:34 AM

And the stuff inside of * comments is still evaluated, so you can't have unmatched quotes or brackets in a comment.

Dennis

So don't use comments. :P

Or SAS.

Alex A.

My job title is SAS programmer. u_u

I kind of have to use SAS.

Dennis

Haha. OK, no comments then.

Alex A.

I usually use C-style comments, i.e. /**/, which are also supported, but for whatever reason are less popular among professional SAS users.

I really wish I wasn't a professional SAS user.

Or a SAS user at all.

Dennis

In some operating environments, SAS might interpret a /* in columns 1 and 2 as a request to end the SAS program or session.

Wait, what?

Alex A.

4:38 AM

WHAT.

I didn't know that.

Mitch Schwartz

here's an ungolfed python program implementing my enumerate brainfuck program approach with verifier ideone.com/0wnG0K

Alex A.

I often have /* in the first two columns.

Dennis

That would explain why they're unpopular. :P

Mitch Schwartz

didn't try it out on a large string, not sure offhand if it would need memoisation to be fast enough

Alex A.

@Dennis Where did you read that? Did it say which environments? The company I work for is all Windows and it doesn't do it in Windows.

Dennis

4:41 AM

@AlexA. support.sas.com/documentation/cdl/en/lrdict/64316/HTML/default/…

Mitch Schwartz

yeah print f(100) is very slow, so it would need that

Dennis

Seems a tad too slow for 100, yes.

Mitch Schwartz

actually i don't know if it would be superslow even with that, because of the other recursion

feels like it should be fast enough though

Alex A.

@Dennis CRAP it also has that in the documentation for the version we're using.

Dennis

SAS seems pretty evil.

Alex A.

4:46 AM

The company or the language?

Because both are evil.

Dennis

I don't know the company.

Alex A.

The SAS Institute, Inc. is the company that makes SAS.

Dennis

But quitting a program because of a comment is evil.

Alex A.

They're evil.

SAS started basically as a way to do statistics using PL/I in the 60s. It was written by statisticians, and statisticians tend to be awful programmers.

I have few if any good things to say about SAS, be it the language or the company.

Mitch Schwartz

as far as i tell the answers are produced very fast as long as f is memoised

Alex A.

4:55 AM

@Dennis Best quote from someone at SAS:

> We have customers who build engines for aircraft. I am happy they are not using freeware when I get on a jet.

Dennis

Because that would be a disaster.

Alex A.

It would be a disaster for The SAS Institute, which wouldn't profit!

Dennis

I've used enough free and non-free software to say with authoritah that they can both suck if they want to.

Alex A.

Sorry, I tend to get fired up and ramble a lot when I start talking about SAS. Its market share is dropping, and I'd like to see it go back to hell where it belongs.

Mitch Schwartz

yeah h doesn't recurse in a loop so h can't be called that many times for a particular sting, and each call is pretty cheap

Alex A.

4:59 AM

@Dennis Amen

Dennis

5:34 AM

Second safe submission! \o/

Alex A.

And the crowd goes wild! \o/ \o/ \o/ \o/

Reto Koradi

5:50 AM

@AlexA. It sounds like you need a new job!

Alex A.

@RetoKoradi Yes.

Yes I do.

grovesNL

Is there specific language known to have the greatest number of operators?

Reto Koradi

Lots of options in Seattle, aren't there?

grovesNL

I know APL has quite a lot but perhaps nowhere near the record

Alex A.

@RetoKoradi Yes, it's just a matter of finding the right one.

@grovesNL As largest built-in library?

Dennis

5:52 AM

@grovesNL I think APL has less than Pyth.

Alex A.

^

Reto Koradi

@AlexA. True. I haven't always had much luck with my jobs.

grovesNL

Ah, really? I thought APL had tons of unicode operators although admittedly I've never used it...

Alex A.

They will all pale in comparison to my new language, DennisLoveMachine.

Dennis

@AlexA. Haha.

Alex A.

5:53 AM

@RetoKoradi Same with me, that's why I'm trying to be very selective this time around.

@Dennis It'll surpass even just kidding.

Dennis

It does. But all letters are variable names.

Reto Koradi

@AlexA. Sounds wise. I partly accepted new jobs because I really needed to get out of somewhere. And then the new job did not work. It can turn into an evil cycle.

Sp3000

@grovesNL One of those Unicode esolangs

grovesNL

Hmm let me rephrase. Which language has the greatest number of built-in functions?

New Main Posts

0

Q: how can i achieve this in java?

A special purpose sorting machine takes text comprising of words and numbers, as input and sorts them both in ascending order and prints them in a manner that a word is followed by a number. Words can be in upper or lower case. The machine converts them into lowercase and then sorts and prints th...

java

Sp3000

5:55 AM

... that's a different question altogether :P

grovesNL

Yeah I just realized what I was trying to ask :P

Reto Koradi

@grovesNL Must be Mathematica.

Sp3000

@Dennis Good news: Short-ish answer to the BF question. Bad news: Convincing myself this is right

grovesNL

I'm too far into the esoteric/Pyth/CJam mindset thinking about operators as functions and vice versa...

Alex A.

Oh yeah, definitely Mathematica. It has the most absurd number of built-ins, for every task possible.

grovesNL

5:57 AM

A quick Google reveals that Mathematica has about 5000 built-ins

Alex A.

:O

Reto Koradi

Python has quite a lot if you count all the included libraries. But must still be much less than Mathematica.

grovesNL

Yeah, I'm counting all included libraries

Alex A.

Holy crap, I found bobince from the famous SO image.

He's REAL.

And he really does have some badges.

Dennis

Yes, he is.

Alex A.

6:01 AM

I wonder if i<3jQuery is also real.

Dennis

@Sp3000 The hard part seems to be going back from natural number to BF code.

I'm pretty sure he made the image himself.

(Which isn't required in the contest.)

Alex A.

No shit?

Hm.

I thought it was Jeff Atwood for whatever reason.

Sp3000

Hmm I wonder if I could reverse this to do natural number to BF

... sounds annoying, though :/

Dennis

I'm thinking about placing a bounty for the shortest submission that comes with a fast inverse.

Alex A.

Ooh bounty.

grovesNL

6:05 AM

Which BF challenge is this?

Alex A.

14

Q: Enumerate valid Brainf**k programs

Golunar/Unary is a way to encode all valid Brainfuck programs, but it is not an enumeration, since most natural numbers do not correspond to a valid program. For the purpose of this challenge, assume a doubly infinite tape and no comments, i.e., a Brainfuck program is valid if and only if it con...

@grovesNL ^

Thanks

Brainfಠಠk

@Sp3000 your solution doesn't give shorter programs lower rank, does it?

Sp3000

It does, how come?

Mitch Schwartz

6:06 AM

oh, interesting

well TheNumberOne's solution is very short and doesn't have that property, my solution is pretty long and has the property

i think i'll still post mine as an answer even though it isn't competitive, here it is in its initial form ideone.com/0wnG0K

Alex A.

ಠ_ಠ_ಠ

Mitch Schwartz

it's just f, h, and g that make up the answer, the rest is just verifying and stuff

Dennis

@AlexA. Brಠಠnfuck

Alex A.

Haha

Mitch Schwartz

i mean, TheNumberOne's solution isn't short yet, but you can tell it will be

Alex A.

6:16 AM

Wow, I just lost a ton of rep, like almost 200, since Digital Trauma's popcon tips was migrated to meta.

Optimizer

why?

its exactly on the same terms as code-golf tips

Alex A.

I guess that's what happens to rep gained from a question on main when it gets migrated to meta.

Optimizer

not that ಠ_ಠ

Alex A.

Not that what?

ಠ?ಠ

?_?

Optimizer

why about merging

Alex A.

6:18 AM

Like why was the question migrated to meta?

I don't know. I think Peter suggested it.

Optimizer

are we going to migrate all tips question then?

Alex A.

Who knows

Probably not

Dennis

@AlexA. After going through the Many Memes of Meta, it seems that the image is in fact Jeff Atwood's work. Sadly, I don't have enough rep on SO to see deleted questions.

Alex A.

@Dennis Oh you have more SO rep than I do so I guess I can't see it either.

Reto Koradi

@Dennis Got a link?

Dennis

6:20 AM

@RetoKoradi stackoverflow.com/questions/2694603/…

Alex A.

Reto, you should be able to see it.

Dennis

@AlexA. Yeah, moderator tools requires 10k on graduated sites.

Alex A.

Didn't realize that was considered a mod tool.

I have about 3.5k rep on SO.

I hang out in low-traffic tags.

Dennis

I have 6.7k. Not sure how.

Reto Koradi

Yes, I can see it. I probably miss the context to know why it's so special. Looks similar to the picture that Alex linked earlier, except with a few more answers.

Alex A.

6:23 AM

@Dennis From being awesome.

Dennis

Could you post a screenshot?

Alex A.

Same way Reto got 26k.

Reto Koradi

Sure, I should be able to do that. Give me a minute.

Alex A.

So the artwork is based on a real question where that happened?

Dennis

I think so.

Alex A.

6:25 AM

:O

Reto Koradi

I probably look computer illiterate right now, but is there a good way of creating an image from a whole web page? If it requires scrolling in the browser window?

Alex A.

Print to PDF?

Idk

Dennis

Print as PDF.

Alex A.

Ninja'd

Dennis

Where did you find that picture? I was young and needed the money...

Alex A.

6:28 AM

@Dennis WAIT. You're Ninja Brian?

aditsu

umm, ninja SEX party, and there are 2 guys on the front page?

Alex A.

@aditsu Idk

Dennis

Well, that's it for today. Good night y'all!

Alex A.

Goodnight!

aditsu

@AlexA. facemaskpalm

Alex A.

6:33 AM

Haha

Sp3000

Having fun caching that BF submission :/

Reto Koradi

3

Is that readable?

Alex A.

Kind of

What I can read makes me so happy

Reto Koradi

Ok, glad to hear you're happy!

Alex A.

:D

I just can't even comprehend that it's a real question and that people actually suggested jQuery.

Reto Koradi

6:44 AM

Most of the answers were probably written as jokes. Not sure about the question. Sometimes you would be amazed (in a bad way) about the level of some questions. In the style of: I'm getting a compiler error saying that a semicolon is missing on line 13 of my code. Why is that? Oh, maybe.... there's a semicolon missing on line 13?

Alex A.

Haha

Reto Koradi

BTW, missed to ping @Dennis. Screen shot of the question/answers you wanted to see is a few messages above.

Alex A.

That's it for me, folks. Goodnight!

Reto Koradi

Yes, time for sleep. goodnight!

Sp3000

7:04 AM

@MitchSchwartz How many length n programs are there for n = 1 .. 7?

Mitch Schwartz

0 1
1 6
2 37
3 234
4 1514
5 9996
6 67181
7 458562

Sp3000

Okay, looking good :)

feersum

There are C(n/2) programs consisting of only brackets, where C is a Catalan number, if that helps

Sp3000

I thought about Catalan but it looked painful :/

jrenk

7:21 AM

hey @all :)

orlp

8:06 AM

@Dennis your brainfuck challenge has a flaw

There is nothing that prevents me from going n -> n*'+' and n*'+' -> n

right now you have these constraints:

"The generated output must be different for all valid Brainfuck programs."

"For every natural number n, there must be a valid Brainfuck program that, when provided as input, generates the output n."

you need one additional one

Sp3000

You give the program as input, not n

orlp

"For every valid Brainfuck program, there must be a valid n that, when provided as input, generates that Brainfuck program."

ohhh

I thought you had to write a two-way function

never mind

Martin Büttner

9:56 AM

ready for review:

4

A: Sandbox for Proposed Challenges

Convert Brainfuck to MarioLANG code-golf code-generation MarioLANG is a two-dimensional programming language where the source code resembles a Super Mario Bros. level. Furthermore, its instruction set is very similar to Brainfuck's. This means that MarioLANG is essentially a 2-D Brainfuck where...

georgeunix

10:10 AM

37

Q: Obfuscated FizzBuzz Golf

Create the shortest possible obfuscated FizzBuzz implementation. To be considered obfuscated, it should satisfy at least one of the following: Does not contain any of the words "Fizz", "Buzz", or "FizzBuzz" Does not contain the numbers 3, 5, or 15. Use any of the above in a misleading way. R...

code-golf obfuscation restricted-source

Aren't the requirements a bit relaxed?

I mean, you could just do 6-1 to get 5

Sp3000

@MartinBüttner Looks okay I think

Martin Büttner

@georgeunix That challenge was posted the day PPCG was created, so don't try holding it up to today's standards. ;)

georgeunix

Haha

OK

Sp3000

Basically anything before 2013-2014 won't have as rigorous specs, I think

georgeunix

OK @Sp3000

Martin Büttner

10:18 AM

posted

now to improve the leaderboard snippet so I can post HW tomorrow

Sp3000

:D

New Main Posts

2

Q: Convert Brainfuck to MarioLANG

MarioLANG is a two-dimensional programming language where the source code resembles a Super Mario Bros. level. Furthermore, its instruction set is very similar to Brainfuck's. This means that MarioLANG is essentially a 2-D Brainfuck where the instruction pointer moves like Mario. So when I wrote ...

code-golf brainfuck code-generation

Martin Büttner

@Sp3000 off the top of your head, can you think of a code golf challenge that mostly follows the "Language, N bytes" format but uses both links on language names and struck-through scores?

georgeunix

The leaderboard work for me with loads of <s></s> in the header

Sp3000

Hmm one of the recent code golfs after your snippet was made?

Oh but link hmm gimme a sec

georgeunix

10:30 AM

Like this? codegolf.stackexchange.com/a/55374/43338

Martin Büttner

@georgeunix yes, the snippet can currently handle that

the problem is it can't handle language names that are links, and I want to fix that... but because the snippet doesn't currently handle it I've been actively editing it out of the answers on many of my own challenges :/

georgeunix

Oh, I see.

Martin Büttner

Sign that word has a single answer with a link

that'll have to do for now

georgeunix

Or someone could edit their answer.

Sp3000

@MartinBüttner Your answer for Retina or Trash? Some incorrect headers though

Martin Büttner

10:38 AM

ah, that's good, thank you

New Sandboxed Posts

10:57 AM

0

A: Sandbox for Proposed Challenges

Get home by ONLY turning right You're finally trying to get a drivers license. About halfway during the lesson, the instructor decides that you've sufficiently mastered telling the gas- and brake pedal apart. It's time to practice the next thing: turning right. The challenge Your program is pr...

Beta Decay

@MartinBüttner I was doing that for this: take the link out of the header and put it in the body to prevent people from getting upset

Martin Büttner

@BetaDecay yes that's what I've been doing as well

but I really want people to be able to keep the links in the header

because that allows me to include those links right in the leaderboard snippet as well

should be working now

Beta Decay

On which question?

Martin Büttner

none yet, but in the snippet on meta

Beta Decay

Oh okay thanks

That's pretty cool :)

Martin Büttner

11:13 AM

So here's a new idea for defining true quines: if you define a new language where you replace all syntax-elements and built-ins of the language in question with different characters (like Alphuck for Brainfuck) but keep their functionality, the output of the translated program in the new language must still be the original source code.

(any arbitrary characters like string contents, custom variable names etc should not be replaced, eval should still evaluate original syntax, and stringification should still yield the original syntax)

@PeterTaylor ^

on a different note, I'm waiting for a sufficiently simple challenge to inaugurate Labyrinth :D

Beta Decay

@MartinBüttner You don't mind my post, do you?

Martin Büttner

no I don't

Beta Decay

Ah thanks

Beta Decay

12:07 PM

If an answer doesn't meet the spec, should I recommend deletion, and if so, for what reason?

Martin Büttner

has the author been informed about the spec violation and has he refused to make an attempt to fix it?

@trichoplax I admire your perseverance on computer graphics :)

trichoplax

@MartinBüttner Thank you :) I'm ready to be put out of my misery now, but it's not over 'til it's over...

Beta Decay

@MartinBüttner He has been informed

codegolf.stackexchange.com/review/low-quality-posts/22553

Martin Büttner

@trichoplax well it's 17 minutes over the extra week now... we should probably hear back some time today

trichoplax

The suspense is torture :)

Martin Büttner

12:15 PM

@BetaDecay I'd give him a day at least to fix it. The comment isn't even entirely correct. FizzBuzz may appear if 3, 5 and 15 didn't appear or if any of those were used in a misleading way. Of course, he doesn't do any of those things.

Beta Decay

@MartinBüttner Oh :P