I have Ferry Cross the Mersey stuck in my head again. I should just make this a code golf challenge

Audio kolmorogov complexity?

There should be a tweetable audio challenge.

I love that last photoshop

So much

@PhiNotPi can audio be embedded in an answer?

To play on demand, not automatically on page open, obviously :)

I just saved a copy of the stock exchange revisited question, so it can be used as a case study for FGITW.

Can you imagine the noise with 20 entries on a page?

I can. FERRY, CROSS THE MERSEY. GET OUT OF MY HEAD

I don't think we have done enough audio challenges.

How about tweetable ascii art?

Audio's strictly worse than visuals

I was thinking there should be a kolmorogov complexity challenge to make a Shepard tone

I think the tweetable art question was great for this site, but another visual challenge would be better than an audio one

I definitely agree audio wouldn't be as popular, but I'd still like to see variety

There are quite a few interesting visual questions at the moment

I'm trying to make an answer to the closed curve image question

I tried that

My answer turned out to be horrible

So you won't be posting it?

Could you show us how bad...?

No, I ruined it at some point last night, and now it doesn't even compile

But it was so bad I didn't bother fixing it

Maybe in a day or two I'll fix it up, but it doesn't work well

It was kind of a dumb idea: Imagine a hose spraying ink at one corner, say the upper left

It flows across the image, but slows down corresponding to how dark an area is

when it's done, cheat and make a loop from the end to the source

Ah I see.

The hose waves back and forth, so it approximates the darkness of the grayscale image

I'm trying to get an image the same way the original was made - I recognise it as a travelling salesman approximation

That's probably better. My method was kind of doomed to fail

I've got a very coarse image but my python implementation of a naive method is running very slowly :)

It ignores lines in the original

Coarse

not course

lol I was in the middle of correcting it...

Sorry

No it's good - thank you

I'm bad about that. I can see a page of text, comprehend nothing, and still make out typos

my spelling is ok but my typing isn't

sometimes I start typing the words from the next sentence before I finish this one - trying to type and think at the same time

I hate to ask you this, but could you submit a solution to my search party question? I'll answer any question you have in return

I spent a lot of time on that, and it's dying :(

it'll probably take you 30-60 minutes

I only have one question and a 300 rep bounty didn't draw much attention to it - all part of learning what makes a good question I suppose. I'm interested in yours but I've just been preoccupied with the traveling salesman one - I should probably take a break and work on something different...

That's very optimistic - I don't work at the same speed as most people in here...

I had a bit of a think about what would work after we discussed it

I doubt your self-assessment is accurate. Everyone here works for a long time and then boils things down

I get the impression most people here are programmers.

I'm not sure what you mean by that

I think most people here are pretty aware of algorithmic analysis, etc.

I mean most people do programming for a living so they are pretty adept at it

Yeah, they mostly beat me in that respect

I wouldn't expect to be able to get an answer as quickly as most in here

I might be able to come up with something though

But most people here also understand complexity, and I've learned a fair amount by reading these posts

Mostly stretching a language; the best algorithms are still always in papers

I'm imagining assigning each square to be searched a parent square from which it is visible, to narrow down the squares that need to be visited

if someone here came up with a better algorithm to do something, I would hound him to publish

quadtree

you're describing the quadtree/octtree algorithm

Maybe that too, but here I don't mean to describe a path for moving, just a visibility approach

so each square would only have a parent, no grandparents.

once you set up a parent relationship, grandparenthood is automatic

Separately I'd need to build a tree for branching out to move to those parents, but I wouldn't need to move to their children

Sorry; what exactly are you trying to solve

I mean no grandparents in the sense that it's irrelevant as you'd still need to travel to the parent. Only the children would not need to be traveled to. Although since we're talking about a group of searchers who also have to stay in visible radius, the grandparent would be useful to measure how many people would need to be sent.

you lost me

What is the problem?

OK. If you only have one searcher, you just move your searcher to each of the parent squares and then the searcher has seen all the child squares without having to travel to them

Still back up; what is the problem?

If you have a number of searchers then you can arrange them in a tree to ensure their survival

^

what are you trying to solve?

So I should just combine those two things into one and stop thinking of them as distinct

your horror film question

you should answer me

I'm looking at this KoTH: codegolf.stackexchange.com/questions/36515/…

It doesn't look significantly different than mine (in terms of appeal), yet in 10 hours its managed to get quite a few more answers than my current Koth

Is it because it seems more complicated? Because I want the programs to persist?/

Terrible writing style?

Or is that KoTH just plain more interesting?

@NathanMerrill it's probably a combination of things. The phrase "to the Death" may have something to do with it - even though it isn't really relevant to the contest.

odd. I only added that because of the character requirement

@githubphagocyte do you normally participate in KoTHs?

@NathanMerrill I got that backwards - I thought "to the Death" would increase interest.

Oh. That's my koth

@NathanMerrill I'd like to, but not so far

given, his has an interesting title

Stack exchange stock exchange

I tried to write an entry for Hunger Gaming when it was posted back in April but I couldn't get it to work. I want to post a KotH so I need to learn about communicating between running programs. That knowledge might help me with competing in other people's KotHs too.

My KoTH should be easy to use

All you need is to print and read from STDIN and STDOUT

I like doing persistent programs because it allows people to write better bots

they can keep data from execution to execution, they can return certain keycodes to query additional data

And it allows for multiple steps

@NathanMerrill if I host a KotH it will definitely use persistent bots, even if there is no benefit to keeping data, just because it will run faster that way. For contests where considering the history is an advantage even more reason to keep them persistent. It will put some people off though. I don't mind that though, I'd rather have an interesting competition with fewer competitors than trying to please everyone

I just wish I had some competition

I got my first real submission today

but I agree with you

The other KotH is also conceptually simpler. This attracts simple solutions. Although there are 13 answers nearly all of them are variations on very simple strategies. I don't think it has more competition than your KotH, only more answers

In a KotH where it is easier to write throw away answers, there will be more answers. I don't get the impression that is what you want though, so it sounds like your KotH is getting fewer answers because that's precisely what you want - only to get good answers.

If you want competition a bounty might help

so, you would say that that is the reason?

that it is easier to write throwaway answers?

I will be adding a bounty

I always do, but after a week

That's my guess, based on the fact that most of the answers aren't serious

Strategies like "forgets to buy or sell, hopes other traders end at a loss"

My point is that KotHs can take a while to gather interesting entries, and comparing with a KotH which is only old enough to have trivial entries will not give you a meaningful measure

45 consecutive days, woo

And 100 rep (along with bounty and downvote privileges) down the drain: meta.stackexchange.com/questions/139270/…

@MartinBüttner :c

So it just disappears?

@MartinBüttner, according to comments on this meta.CR answer the 200+ questions on each of the top 40 tags restriction for the Generalist badge has been removed.

So it's slightly puzzling why those users you mentioned (Howard and one other?) don't have it.

@PeterTaylor primo and myself

@PhiNotPi Yep.

I'm writing a website which randomly generates and plays music on the fly, viewers can upvote/downvote while it plays :D

@PeterTaylor Either that comment is wrong or the restriction was removed on 5 May this year: codereview.stackexchange.com/help/badges/14/generalist

I thought that that was the day where the 40th tag crossed 200 questions

or maybe it was just lowered to something like 100

I'll ask on meta.SE

@cjfaure How is the tetromino controller coming along?

@MartinBüttner It hurts my brain ;_;

I'm still trying to get board generation to work.

Try random placement until you get the desired density.

I think the target density is low enough for this not to take forever.

@MartinBüttner that has a chance of never finishing

albeit undefined and approaching 0/0

@cjfaure As I said, I believe the target density is low enough for this not to be a problem.

@MartinBüttner I'll try it then. Target density is 1/2, right?

I thought it was even lower?

i mean after placing just the tetrominoes

no you're right

1/2

kk

Run it a few times to get a feel for the necessary iterations, and then add an iteration cap (dependent on number of players). When the cap is hit, discard the board and try another seed.

The problem I have is fitting it with that density into a square - though I can get around that by working from the outside in, right?

not sure what you mean

@MartinBüttner sorry I'm used to programming cellular automata sims, lol

this is screaming "duplicate", but I can't find it:

@PeterTaylor?

Sure you're not just remembering it from the sandbox?

He doesn't have enough rep for meta.

And I can't find anything like it anywhere on meta either

@MartinBüttner like this? :P

	def gen(self, shapes):
		dim = math.ceil(math.sqrt(len(shapes)*8))
		nsh = shapes
		while nsh:
			nsh = shapes
			map = [" "*dim for i in range(dim)]
			i = 0
			while nsh and i < 1000000:
				j = random.randrange(len(nsh))
				r = random.randrange(4)
				f = random.randrange(1)
				k = nsh[j]
				try:
					p = Piece(shape, random.randrange(dim-len(k)), random.randrange(dim-len(k[0])))
					map = p.place(map)
					del nsh[j]
				except: pass
				i += 1

p.place throws an assertion error if it overlaps another shape.

1000000 is the iteration cap, i didn't test/run the code so it's probably too high

whatever your golfed variable names mean...

@MartinBüttner xD

wait i didn't use r or f facepalm

def gen(self, shapes):
	dim = math.ceil(math.sqrt(len(shapes)*8))
	nsh = shapes
	while nsh:
		nsh = shapes
		map = [" "*dim for i in range(dim)]
		i = 0
		while nsh and i < 1000000:
			j = random.randrange(len(nsh))
			r = random.randrange(4)
			f = random.randrange(1)
			k = nsh[j]
			try:
				p = Piece(shape, random.randrange(dim-len(k)), random.randrange(dim-len(k[0])))
				p.rotate(r)
				if f: p.flip()
				map = p.place(map)
				del nsh[j]
			except: pass
			i += 1

j is the index of the shape I'm trying to place, r is the amount of times to rotate 90 degrees, f is whether to flip it or not and k is the actual shape of it

looks good to me

;D

Ah, there's such a neat solution for the subsequence one in Mathematica, but it's never gonna handle lists of 1k elements.

f=Max[Length/@Cases[Subsets@#,l_/;Less@@l]]&

now, of course, I can't test it until i finish implementing the actual players.

@MartinBüttner It's a repost of a question which hit -7

@PeterTaylor aaaaaahhhh

how did you find that? :D

Google

how the hell do i do this?!

heeeelp

@PeterTaylor ah, clever :)

Btw, did you see my regex submission for the smoothest number challenge?

subprocess sucks xD

I did.

I also upvoted it.

ah k (thanks :)) ... I think based on that it should be possible to make the PCRE version a lot shorter... I wouldn't even use actual recursion, just use the entire snippet that finds the largest prime factor of a number.

but I'm not sure I can be bothered ^^

guys i need some assistance :P

using python, how do you keep a process alive and use it's stdin and stdout as communication?

The docs say, use communicate() on a popen, but the description for communicate is

@PeterTaylor wow, even the plain translation to PCRE is 2 bytes shorter

(using \K to avoid the lookbehind)

^(1+),.*?\K(?=\1)(?=((11+)(?=.*(?=\3$)(?!(11+?)\4+$))(?=\3+$)|(?!(11+)\5+$)1+))(?!.+(?=\1)(?:(?!\2)|(?=((11+)(?=.*(?=\7$)(?!(11+?)\8+$))(?=\7+$)|(?!(11+)\9+$)1+)).*(?=\2$)(?=\6)))1+

I will golf this and beat some Lua answers :D

Interact with process: Send data to stdin. Read data from stdout and stderr, until end-of-file is reached. Wait for process to terminate. The optional input argument should be a string to be sent to the child process, or None, if no data should be sent to the child.

> Wait for process to terminate.

nice.

@cjfaure search PPCG for "hangman solver"

@MartinBüttner every line in that program conflicts with the documentation.

129 bytes :D

Edited.

Now I'm beating both Lua submissions, as well as Cobra, Clojure and C#. :)

If the program hasn't produced any output since the last, does proc.stdout.readline() just return ""?

the only thing C# beats that anybody uses for code-golf is Java ;)

@VisualMelon Too bad nobody golfs in C--.

@cjfaure Or does it block?

is that a thing (C--)?

@VisualMelon Very much so.

A lot of my students could have used C--, but I went with D

Teaching remedial math sucks

@VisualMelon IIRC, my Java entries usually beat C# :p

of course, after checking I see that I only have a single Java golf on a challenge that also has a C# entry. The good news? It beats both of those, though only by 2 characters :D

@EricTressler did you see? I golfed down the regex a bit :D

No; I'll have to look later, because I'm about to go to sleep again

But I'll look then

kk

@Geobits I just found a way to save 17 bytes >:D

:D

@ProgramFOX too lazy to test myself, can you not have the class inherit Form, and then initialize one of those instead? (save two one byte)

@VisualMelon Main has to be a static method, so I cannot.

would that be an issue?

this is where I make myself look an idiot by not knowing my only language

I cannot change these variables in Main, because that's a static method. If I would create another method, I would not save bytes.

Oh wait... my C# solution was more than 17 characters longer than @Geobits' one. So Java still beats C# :(

hangon... there is something very wrong me with

that is longer

Ugh, English terminology for blood relations is a) boring and b) ambiguous.

@MartinBüttner Is it really best to assume both or neither parents are unknown? I guess you've never seen Maury? :p

It gets more fun when you include relations by marriage and cycles.

I seem to recall reading once about a Canadian who married his own great-granddaughter (or maybe there was another great-): apparently that was a sufficiently distant relationship not to be prohibited by law as incest.

So describing her relationship to a lot of people got complicated.

@Geobits I copied that from Calvin's proposal. I haven't yet figured out why that's helpful. (So I might scrap it.)

Greek mythology and the Julio-Claudian family are also good examples of where it gets complicated.

@PeterTaylor Yeah you can easily create rather small incest-free loops.

I think Calvin was right though that non-blood relationships would complicate things too much, because they would probably just not be as well defined.

Tell me that this answer isn't exactly the same as this answer...

What I was referring to was that "M times removed" is ambiguous and that you're lacking the amazing terminology German has for great*X grandparents :D

@Rainbolt the conditional is different isn't it?

But he managed to find a condition that would be false at all of the exact same times that mine will be true, and then flipped the results.

you buy when you have zero shares, he buys depending on relative sizes of money and value?

@Rainbolt k, I can't tell, haven't read the rules.

But he buys as much as he possibly can, meaning that his money will be almost zero when he buys.

It looks like it could diverge. If others sell while he is buying, the overall price could drop, enabling him to buy more on his next turn.

So he could buy even when he has shares, where yours never will.

You're right. He stumbles if the value of stocks somehow drops below the amount of money he has. But that won't ever happen because on every other turn, the price is astronomically high.

Ok, I think they are different answers, and they just behave the same because of a flaw in the game itself.

I haven't seen what patterns actually happen, but yeah, that sounds about right.

@MartinBüttner what is the German terminology for this?

@Geobits The pattern right now is "Everyone buys one round one." "Smart people sell on round two and suddenly have more than ten times their original money." Repeat.

If OP posts his controller, we can see whether the price still behaves as badly with the current 17 answers as it did with the 3 he had last time.

Der glockenspielschadenfreude

@Geobits Basically, there are additional prefixes for every 3 generations, which you use in conjunction with cycling through "_, grand, great grand".

And there are like 8 of these prefixes or so

great-grand-kilofather?

(no one uses or knows them, but apparently they exist in proper genealogic terminology)

gigafather

That sounds tricky

yeah, it's definitely SI prefixes

It's the first column under "Alternative genealogische Bezeichnung": de.wikipedia.org/wiki/Generationsbezeichnungen

So does that mean it goes the other way too? Maybe decifather for descendants?

A clock that derives pleasure from the misfortune of others

I think when you go down the tree, it goes from father to son

decison

My decison has an important decision to make today.

They mean something along the lines of "old", "over", "root", "ancestral", "great-ancestral", "arch", "arch-ancestral"

Swedish uses a simple system: "far", "mor", "farmor", "morfar", "farfar" means "father", "mother", "father's mother" (i.e. grandmother on father's side), "mother's father", "father's father"

I think Icelandic does that too, but they do have short forms at least for grandparents

So in Swedish a great-grand-kilofather could be e.g. a farfarmormorfarfar?

Iceland has an app to avoid incest

wat? :D

Hm

Too many cousin relationships which people don't know about.

The gene pool there is very limited, and people genuinely don't know about their near relations sometimes

So they actually do have an app to help people avoid marrying people they're closely related to

I think that'd be "farfars mormors farfar"

@EricTressler I can't even imagine the disappointment when they check the app. :(

And then, all they have to go back to is lutefisk