The Nineteenth Byte

2:00 PM

@Rainbolt you could install a browser extension such as Imagus to show images when hovering over. It comes in use in other places too :)

Compass

Well, I'm using Collections.sort so it's not my mergesort, it's the library's ._.

Rainbolt

Oh, I see

Geobits

What comes out of it? Is the "sorted" list in the exact same order as the original?

"Doesn't sort" could mean a few things.

Compass

@Geobits GOOD IDEA

Rainbolt

2:14 PM

Oh, so return thisLong - thatLong doesn't work with longs

Because compareTo returns an int and you might accidentally wrap around when you try to convert

@Compass Ok, so I just implemented the logic from your friend and found that my list of objects did not sort at all.

[-2984990665945965758, 8007777253927337067, 6017403796917550899, -8319492675305266938, -1974614842160663373, -7475941312107594453, 6984756601187698770, -6409319350170508176, 2394843995449090561, -3011597354772439812]
[-2984990665945965758, 8007777253927337067, 6017403796917550899, -8319492675305266938, -1974614842160663373, -7475941312107594453, 6984756601187698770, -6409319350170508176, 2394843995449090561, -3011597354772439812]
BUILD SUCCESSFUL (total time: 0 seconds)

Then, I modified the compareTo() to return -1 when it is supposed to, and it sorted correctly

trichoplax

@Rainbolt Does the non-working version still work when all the items are longs which would fit in an int?

Rainbolt

    public int compareTo(Foo f) {
        if (l > f.l)
            return 1;
//        if (l < f.l)
//            return -1;
        return 0;
    }

That's the non working version

And I don't know what "Does the non working version still work" even means

Geobits

I think it matters more if the result of a-b would fit into an int. Even if both a and b do.

Rainbolt

I'm not using a - b. I'm using the logic that Compass described earlier

Geobits

I know, I thought he was asking about your earlier message where you were talking about overflow.

trichoplax

2:25 PM

@Rainbolt My wording is appalling. I meant does it consistently not work, or are there cases for which it still works

Rainbolt

If there are, my rng isn't doing a good job at finding one.

    class Foo implements Comparable<Foo> {
        public static Random rng = new Random();
        long l;
        public Foo() { l = rng.nextLong(); }
        ...
    }

Geobits

I'm guessing that since the default sort order is usually ascending, the < comparison is used more often than the > internally, so it not working without a reliable one makes sense.

The only question is why would it not work now when it (apparently) worked for years.

Rainbolt

I think Compass has two problems. First he needs to fix his comparator. And then he needs to figure out why each machine behaves differently.

Geobits

Agreed for sure. No excuse to let that comparator stay there unless there's a damn good reason to.

Rainbolt

Some things you just fix because that's how itshould be

(My supervisor on the other hand might disagree with me)

Katenkyo

2:30 PM

@Rainbolt It because he thinks it is a wait of time, as it is actually working. Supervisor are actutally more (often) obsessed with short term results than reducing maintenance time cost

Rainbolt

We had an item in our css that made all iframes look a certain way. Then we started using a reporting tool that made heavy use of iframes, and so I modified the css to only apply to a specific class. He asked me "Did you break any iframes?" and I could only say "I dunno, but if I did then they need to be broken so I can find them and give them a proper class."

Katenkyo

An actually, a lot of IT dev are doing things thiw way...

Geobits

It's not just IT, it's been happening for a long time in a lot of areas.

Katenkyo

(saw in a PL/1 source 4-level nested conditions that were copy paste many times because they didn't want to had the statement in the existing condition block)

Compass

@Rainbolt thanks i think it's a Java issue that recently came out, so I'm going to verify java versions through regression and then document it as such

er the sort breaking is a result* of a Java change

problem on my end, not java's XD

Katenkyo

2:32 PM

@Geobits I prefer speaking of IT only, because that's what I know :)

Rainbolt

@Compass Glad you found it. And also glad the problem isn't within Java's sort function because that would be scary.

Compass

I would never assume sort is broken

especially when its my comparator

if its java's comparator, then #ffffffff

Geobits

@Katenkyo My previous job (military avionics) was much better about thinking long-term, but they've learned they have to be if they want to keep flying the same planes for 50+ years.

VisualMelon

be afraid, be very afraid: envisage-project.eu/…

Compass

If you told me the F-14 was built in the 1970s I would be like NOWAAAY

the difference 30 years can make for aircraft @.@

Rainbolt

2:34 PM

My boss makes me explain the long term benefits to him. It's not that he can't see it so much as he trusts his developers to know what they are talking about

He's a really harsh person when you ask him to do a code review

Katenkyo

@Geobits Haha, it should have been nice to work there !

Geobits

It was, except all the planes were so damn old :D

Compass

Planes built before MS-DOS x_x

Rainbolt

Wow I made Telerik spit out a stack trace :(

Compass

its okay, this navy ship uses floppy disks for firmware upgrades!

Geobits

2:36 PM

Mostly 60s-70s models. Have you ever worked on 40+ year old wiring harnesses? It's crazy.

Rainbolt

It actually posted the entire stack trace inside of a report viewer and totally bypassed our custom error page

I'm just about fed up with this HTML5 buzz word (especially since HTML6 is close) and the stupid reporting tools that look so great on their website but look like shit on ours.

Our CEO sees these awesome demos and thinks we can make websites that looks exactly like the demo.

Geobits

Just Web 3.0 that shit.

Rainbolt

While also adhering to requirements that stretch the boundaries of what these tools can even do

Ok /rant off. This is what we get paid for I guess.

Katenkyo

I personnaly prefer a sober site to a funky animated one.

Compass

wait what

HTML6?!

I AM NOT READY

Geobits

2:40 PM

@Katenkyo Tell that to SE :(

Rainbolt

Funny thing about HTML specs (and also new releases of JavaScript) is that nobody uses them if no browser supports it, and no browser supports it if nobody uses it. So how do these things catch on? I don't know.

Geobits

Nerd-fueled rage at the browsers until they relent?

Compass

=.=

trichoplax

I guess an open source browser implements it because someone decides to, and then the other browsers have to catch up

Lembik

hi all

Rainbolt

2:44 PM

I met a guy at a conference last week (that's why I was missing all week) who was fanatical about JavaScript. And the more he talked about it, the more I couldn't understand why anyone would want to use it.

Katenkyo

@Geobits Sticky following votes aren't so annoying I think, but I wouldn't like it to go further than votes

trichoplax

@Lembik Hi

Lembik

manged to get +1 for my question :)

I am so proud

Compass

javascript is a nice language for when you want to code and don't feel like compiling or actually playing nice with reality

Rainbolt

I learned that JavaScript has === operator because == does things that nobody would expect

Geobits

2:45 PM

I have a feeling I won't like it, based on the discussion there. Unfortunately, I'm not in the activated group, so I don't know for sure.

Katenkyo

@Geobits me neither, but i'd like to see it

Geobits

=== != ==

trichoplax

I like being able to see buttons from the bottom of a post, and not having to scroll back up. It's only the jarring animation that's a (small) problem

Rainbolt

He presented a list of five comparisons and asked us to write down what each would evaluate to. Not a single person in the room got them all right (and admitted it), and there were maybe 200 of us there.

Katenkyo

@Rainbolt === compare by type and == compare by value, right?

(I hate web dev :))

Rainbolt

2:47 PM

=== is "compare by value without help" and == is "compare by value with help"

trichoplax

@Rainbolt Is it the other way round?

Katenkyo

@Rainbolt and what is "help" if I dare ask...?

Geobits

@trichoplax I like having the visual indicator/separator between posts when scrolling through a bunch of them. I think with them moving it'll be harder to scan/skim.

Rainbolt

'' == '0'           // false
0 == ''             // true
0 == '0'            // true

false == undefined  // false
false == null       // false
null == undefined   // true

Compass

yes, the one thing that is creepy about javascript is how == is not friendly

Rainbolt

2:47 PM

That's what I mean

Geobits

That's not very helpful help.

trichoplax

@Geobits Maybe they should introduce a bolder separator if they keep the moving buttons

Compass

but a lot of other things for javascript are kinda nice for when you want to just do something simple without all the boilerplate that... say Java asks you to add

Geobits

That might help. I wish I was part of the testing group so I could give a more informed opinion. Right now it's like someone telling me they're going to move my cheese but not doing it.

Compass

CxF5

Rainbolt

2:49 PM

Yea who needs boilerplate to establish scope. Just don't ever use var outside of a function and you'll be fine.

Lembik

does it really have === ??

where/how can I look this up?

Compass

yes javascript has ===

Katenkyo

Actually, the only part of web dev I liked, was doing a struts WAS designed project, with the middle tiers in java

Compass

its not a joke

Geobits

@Lembik Some people recommend using it exclusively, and never relying on ==.

Rainbolt

2:49 PM

682

Q: Difference between == and === in JavaScript

Possible Duplicate: Javascript === vs == : Does it matter which “equal” operator I use? What is the difference between == and === in JavaScript? I have also seen != and !== operators. Are there more such operators?

Compass

w3 schools has actually improved its javascript page

Lembik

wow!

Compass

amazing

trichoplax

@Rainbolt I see you're right now. It took me several minutes to work out that I was wrong, and now it's too late to edit my post...

Katenkyo

Incoming soon : ==== and !===

Geobits

2:51 PM

==== should obviously check to ensure value, type, and pointer are all equal ;)

Katenkyo

@Geobits sounds amazing !

Geobits

I suppose ===== should also check variable name and scope. So a thing is only ever ===== itself in the most literal way.

Rainbolt

imgur under heavy load :(

Katenkyo

where ==== would check duplicates

Geobits

Right

And ====== is just never true.

Katenkyo

2:54 PM

@Geobits no, it would check the thread it runs in :)

Rainbolt

Well since imgur won't let me upload, I just have to tell my story :(

Katenkyo

Or the thread who created it

Geobits

Oh, I thought that counted in scope for =====.

trichoplax

Equals Can Mean Anything Script

Rainbolt

I typed in "quadruple equals" and Google suggested that I was searching for "javascript quadruple equals"

Katenkyo

2:55 PM

Oo

Rainbolt

Oh, there we go:

Katenkyo

Ho, and don't forget the closure !

Compass

@Geobits it's not sorting at all XD I just need to figure out why java wasn't sorting it

Geobits

Nice article title: Equals, Equals Equals, or Equals Equals Equals

Compass

time to hand-write merge sort

Katenkyo

2:58 PM

if I define something like f = function(){var a=0;function(){return a++}}

trichoplax

@Compass So that's why they make you do it in interview

Compass

@trichoplax actually all my interviews when i explain a sort method

they look at me and say "You could have also said Collections.sort"

trichoplax

lol

Compass

And I just sit there and go =.= what a cop-out answer

Katenkyo

nevermind, it is covered by the pointer thing

Geobits

2:59 PM

@Compass It's not a cop out answer IMO. Sometimes you want to make sure someone won't go to the trouble of handwriting a sort if they don't need to ;)

Rainbolt

I tend to start with "Do you want the Collections.sort answer or do you want to see me write it myself?"

Katenkyo

@trichoplax it would be a little bit like chicken

Rainbolt

Never had an interviewer get mad for asking

trichoplax

@Katenkyo needs http:// to make the link work

Compass

@Geobits all my interviews were extremely technical

Rainbolt

3:00 PM

I've had two people ask me to do a sort in person, and one via email. On the email one I just wrote two versions and submitted both

Katenkyo

@trichoplax thanks :)

trichoplax

@Katenkyo no problem :)

Geobits

@Compass An even better reason to make sure you find someone who knows not only how to be technical, but also how to just get it done.

trichoplax

@Katenkyo That's terrifying

Agawa001

almost done with my submission

Katenkyo

3:02 PM

@trichoplax It's something !

Geobits

It's not terrifying if you read the whitepaper instead of just the esolangs summary: isotropic.org/papers/chicken.pdf

6

Compass

@Geobits I'm sure my interview woudl have sucked if I had just said "It's on StackOverflow!"

Sparr

@MartinBüttner I'm thinking of a class of marbelous devices where the value of the marble affects what the device does, rather than being useful as a variable coming out of the device, so the output would mostly just be marble-vs-nomarble

trichoplax

@Geobits Are you trying to kill me???

Sparr

a "teleport X squares upwards" device would implement arbitrary delays of 0-255 ticks

Compass

3:04 PM

Marbelous?

Katenkyo

@trichoplax check this one, much more readable because it is in chinese !

Compass

Is that a real word?

Sparr

@Compass github.com/marbelous-lang/marbelous.py/blob/master/doc/…

trichoplax

"Would you like the Collections.sort answer or shall I go ahead and implement it in chicken?"

Katenkyo

@Compass Should be Marvelous for a real word :)

Sparr

3:05 PM

@Compass some folks here designed an esoteric language based on marbles in a rube goldberg machine

Compass

That is so sway o.O

Rainbolt

Oh my. Another word I don't know

That's two in two days

Yesterday I learned thirsty

Sparr

@MartinBüttner !! math.stackexchange.com/a/1318135/1995

Rainbolt

My roommate brings home all sorts of words since he teaches high school. It makes me feel old

Compass

@Rainbolt I found the issue

Java 7 decided to redesign sort

Katenkyo

3:08 PM

Well, time for me to leave you there, hafta go. See you tomorrow guys !

trichoplax

Bye :)

Compass

there is a legacy sort

but you have to request it

else it uses comparabletimsort

Thanks, Java 7!

Rainbolt

@Compass Is someone at your work place making you go through the painstaking process of solving this mystery? Or are you just going at it for fun? Congrats on solving it either way.

Compass

Yes.

We need to know why it happens.

So we can release it into PROD

we can't replicate it in QC so we can't ask for a patch fix because we can't guarantee it's a fix

Rainbolt

Oh, tell your boss you want a test environment that is identical in every fashion to actual production.

Compass

3:11 PM

we have one

Rainbolt

Explain how it will save him money down the road because bugs like this won't make it into production

Compass

i cant replicate it on that either

Agawa001

how can i make this with less brackets

0

A: Visual Long Multiplication

C (330 b) int main(char*a){int i,j,r,k,c,h,o,e=15,m=99,A=0,B=0,*C,L[m][m],G[m],*g=G;for(C=&A;(c=*a-48)+48;C=!(c+6)?&B:&(*C+=(*g++=c+1)),a++);for(i=B-1,j=0;j<(r=A+B);i--,j++)for(k=0,o=4*!!((*(g-=!*g))---1);k<*(C=(h=i<0)?&B:&A);k++)L[abs(i)+k][j+k-2*k*h]+=o/(3*h+1)*e;for(i=0;i<r*r;i++)printf("\n%c...

Compass

we "have" I should say

Rainbolt

Does it have the wrong version of Java on it?

We just had the same problem actually. We released a version of our product not knowing that it was dependent on .NET 4.5. Our documentation lists .NET 4 as a prerequisite.

Compass

3:13 PM

I don't know what versions of Java are running on our environment servers. I put in a request to get them.

Sparr

@VisualMelon this is my attempt to get more insight into the probability of different trees in my challenge: math.stackexchange.com/questions/1316839/…

there's a great answer there that might actually be valid

combining that probability with the mutation improbability graph will probably produce a perfect answer to my challenge.

VisualMelon

my thinking is it's just the product of the probabilites of each edge, divided by (n-1)!, times the number of ways to make said tree, which I have a formula for

(can ignore the (n-1)!, same for everything, and can use a spanning tree to get a good upper bound for a brutal search, is my best (working) idea so far)

(I say working, I don't know if it works, but it's the only thing I have any confidence in)

Compass

@Rainbolt yup, I'm compiling in 6, servers compiling in 7. Gah.

Luckily I can fix it pretty easily. Normally would have gone under the radar with using a new optimized sort, but with this lazy sort, things borked.

Compass

3:46 PM

So to recap for anyone who cared: JRE6 used Merge Sort. This only needs a comparator to return 0 for <=, and 1 for >

JRE7 uses Timsort, which is a combined merge/insertion sort that creates sorted buckets, from what I've read. That needs -1 for < and a poorly created comparator will break it. QED time to go to lunch :D

Optimizer

4:05 PM

@trichoplax I'd prefer Chicken one

@Compass umm. are you saying timsort is not that great ?

trichoplax

I heard timsort is that great

Unless you think a petrol engine is not that great because it will break if you put coal in it ;)

Optimizer

@trichoplax I must admit, I did not understand his line at all :D

trichoplax

@Optimizer It makes more sense in the context of the problem solving that happened in the previous chat

Optimizer

(at the first go)

Geobits

I read it as "if a comparator was designed to meet the bare minimum necessary for the old method, then it won't work with the new method, which has stricter requirements". Not really anything "bad" about either method.

trichoplax

4:09 PM

How are we supposed to argue with you being sensible??

Geobits

You're not. Optimizer will by default, but llamas are stubborn like that :P

trichoplax

We're not really arguing (apart from Optimizer, who always is for fun)

Optimizer

You don't mess with a llama!

trichoplax

Not twice!

Geobits

Out of curiosity, were you born a llama, or were you Kuzco'd into it?

trichoplax

4:11 PM

@Optimizer I don't really know chicken. I just said something at random while reeling from Geobits' link at the top of the starboard

@Geobits I'm not sure where to put the apostrophe to indicate your possession of something. Geobits'? Geobit's? Geobits's?

Optimizer

@trichoplax same here ("just said something")

trichoplax

:)

Optimizer

@Geobits a secret untold

a mystery unfold

still the secret remains

Geobits

@trichoplax EL&U might have a different opinion, but I don't care at all about which is used. I'm fairly sure Geobits's is "right", though I normally just use Geobits'.

trichoplax

I'll stick with the shorter version then

Geobits

4:15 PM

I'm gonna go with the Kuzco version then.

New Main Posts

4:43 PM

0

Q: Copy array with no variables or extra memory usage

Im curious to see some other ways to make a copy of a fixed array, which potentially use no extra variable or memory and has as small code as possible. Here's my attempt in C: #define ARRAY_COPY(x,y,n) for(*(x)=n;*(x);(*(x))--) (x)[*(x)]=(y)[*(x)]; \ *(x)=*(y); The l...

code-golf array-manipulation c

Martin Büttner

5:14 PM

@Rainbolt comment ninja

Rainbolt

We might want to think a little harder before we blindly send people to SO and CR

Martin Büttner

@AlexA. ^

Rainbolt

Now I will plug my answer that nobody liked, which would clearly have helped us avoid this issue: meta.codegolf.stackexchange.com/a/5041/18487

Agawa001

@Sparr i was surfing thru ur proposed SE section , i discovered out very very wierd orientations

Rainbolt

I wonder if anyone has empirical evidence that most of our off topic questions actually belong on Stack Overflow.

Agawa001

5:20 PM

u said ur program shouldnt print empty chars

Martin Büttner

@Agawa001 empty characters?

Sparr

I'm passing stdin through a Thread and Queue in python to get non-blocking input, and I can't figure out how to detect EOF.

Agawa001

u said ur program shouldnt print empty chars

Sparr

woo! first commits to marbelous in 10 months. github.com/marbelous-lang/marbelous.py/commits/master

Martin Büttner

@Agawa001 do you mean spaces?

Sparr

5:22 PM

@MartinBüttner ^^ not related to turing completeness, I'm just bringing the repo up to date with stuff I never committed. But I still do want to get us at least up closer to TC-ness

Martin Büttner

@Sparr as I said to Nathan, recursion might actually make it TC

because that does allow you an infinite amount of storage

although it might be more like PDA, since recursion only gives you a single stack

Agawa001

yes , spaces after end of diagram

Geobits

That's a fairly standard rule that basically means "I want the answer and just the answer".

Martin Büttner

spaces after the end are fine (provided they don't go beyond the bounding box). but your submission had spaces within the diagram.

Agawa001

why is that forbidden

Geobits

5:25 PM

It doesn't meet the expected output?

Martin Büttner

uhh... I don't know. because it doesn't look as nice? because that's how I set the challenge?

yeah, what Geobits said

Geobits

If I said "output 'Hello world'", I wouldn't expect 'H e l l o w o r l d'.

Agawa001

i cant understand this quite enough

You may print trailing whitespace but it must not exceed the diagram's bounding box.

is my submission violating this rule ?

Martin Büttner

I'm not even sure why you are printing those intermediate spaces. it actually costs you a character to add those spaces in. just remove the space from the format string.

@Agawa001 you actually are. there are three columns of trailing spaces that shouldn't be there.

Agawa001

oh , i must narrow the nethermost loop

to tighter bounds then

Martin Büttner

5:30 PM

(and remove that space from the format string)

Agawa001

can u propose any other fitting ascii character instead of space?

Martin Büttner

I don't understand

just remove it

"\n%c " --> "\n%c"

Agawa001

i think u miss how does C works like , C terminal cant just move cursor there and type another \ or / or x , there must be a permeating character between

Geobits

?

Agawa001

oh wait

that space

urrr

and costs me excessive byte :D

fixed

the inversed sight is rooted problem thu , cant be fixed in a minute

Sparr

5:39 PM

@MartinBüttner I had completely forgotten about recursion's storage capabilities...

Martin Büttner

@Agawa001 you're still printing one space past the bounding box. can you somehow just reduces the range of the horizontal loop by 1?

Compass

@Optimizer no i think timsort is great! It just broke my app!

Agawa001

@MartinBüttner that should cost me two bytes

i know there should be some brackets to get ride of somewhere

im certainly sure

Martin Büttner

alright, so what about the reverse order of digits

Agawa001

thats another issue (would take much time)

Chris Jester-Young

5:52 PM

@AlexA. (Speaking for myself, not for the company.) I don't believe that the site icon will ever be PPCG. 1. it's way too long, and 2. nobody in the company calls the site PPCG.

Agawa001

i ll deal with it tomorrow

Martin Büttner

@Agawa001 could you just traverse the input string in reverse?

Agawa001

@MartinBüttner string crossing is done twice , once from right to left , then way round

so i think its nt something to handle in a hour

Chris Jester-Young

@AlexA. This proposal has the most promise, though: meta.codegolf.stackexchange.com/q/788/3

Agawa001

and im tired for now , i leave it tomorrow

thx thu

Sparr

5:55 PM

@MartinBüttner #programming on freenode seems to think recursion makes us TC, even with the tight bounds on storage and code in our language

Chris Jester-Young

@Sparr Of course. Recursive algorithms can be CPS-transformed. ;-)

C0deH4cker

6:10 PM

Are you creating a language?

trichoplax

@C0deH4cker Marbelous has already been created :)

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Captcha cops-and-robbers Cops develop a captcha algorithm in minimum bytes. Robbers tries to implement a recognizer that recognizes 90 of 100 generated captchas.

C0deH4cker

Ah, looks fun!

trichoplax

People have already posted a few answers in Marbelous on PPCG

C0deH4cker

Does it require spaces between the bigrams, or is that just for clarity?

Sparr

6:21 PM

just for clarity

intepreter allows 0,1,2 spaces

Compass

is marbelous designed to be human-readable?

Nathan Merrill

@Compass to an extent.

Compass

>_>

Nathan Merrill

because we have a limit on the size of keywords

its really tough

but we tried to pick smart keywords

and we added comments

Rainbolt

6:35 PM

I feel like ReSharper should be able to pick up on this:

    var s = "";
    for (var i = 0; i < 10; i++)
    {
        s += ".";
    }
    Console.WriteLine(s);

I just wrote that to test if it would detect string concatenation in a loop and suggest that I use Stringbuilder instead

Sparr

6:46 PM

@Compass with whitespace between columns, small marbelous programs should be easy for a human to interpret in their head.

well, easy for humans who can do that with some other language already :)

Martin Büttner

Does anyone know what sort of number system this is, that Randall uses for page numbers in xkcd volume 0?

It's like ternary, but he skips straight from 2(0)^n to 1(0)^(n+1)

(might make a nice challenge)

I guess I should check oeis

apparently there is no nicer definition

Sparr

7:04 PM

he also skips 121 and 122?

Martin Büttner

well that's what I mean

he skips from 20 to 100

so he also kips from 120 to 200

Sparr

ahh

Compass

i assume it's related to after 111, a 2 clears all rows o_o

er all columns

so it'd go 1111, 1112, 1120, 1200, 2000, 10000

Martin Büttner

indeed

sandboxed it

Sparr

lies! the bot would obviously have let us know by now if you had. :p

New Sandboxed Posts

7:21 PM

0

A: Sandbox for Proposed Challenges

xkcd-Style Page Numbering code-golfnumbersequencebase-conversion Randall Munroe's book "xkcd, volume 0" uses a rather odd number system for the page numbers. The first few page numbers are 1, 2, 10, 11, 12, 20, 100, 101, 102, 110, 111, 112, 120, 200, 1000, 1001, ... This looks a bit like ter...

Martin Büttner

@Sparr there you go

Geobits

7:41 PM

@Martin Reference solution? forums.xkcd.com/viewtopic.php?f=11&t=46142

Martin Büttner

@Geobits I'm wondering if I should require fast processing of sufficiently large input to avoid just counting up ternary naively.

(I've got a reference implementation like that in CJam)

Geobits

K, was just curious if you had something handy, it was just the first result for searching "xkcd volume 0 page numbers".

I'm not sure how feasible it is to calculate without counting up, honestly, but I haven't looked at it much.

Sparr

7:58 PM

my idea would be to just count in ternary and filter out strings that match /2.*1/

/2.*[12]/ that is

trichoplax

Is there a way to convert it to a linear recurrence relation? (like the Fibonacci sequence)

Compass

i think so ._.

trichoplax

In that case it has a closed form, although that might not help in golfing since the coefficients may require a lot of digits...

Sparr

sure. f(x) is f(x-1) incremented until it doesn't contain 2.*[12]

trichoplax

8:13 PM

I'm not sure that's linear recurrence...

It looks like the number of terms in the recurrence would keep increasing so there wouldn't be a fixed way to write it

Sparr

ahh

trichoplax

I was so amazed when I first heard about the closed form formula for the Fibonacci numbers and remembering that made me want to find a similar approach here but I'm pretty sure it can't work like that

def fib(n):
    return ((((5**0.5)+1)/2)**n - (-(((5**0.5)+1)/2))**(-n)) / 5**0.5

Geobits

8:30 PM

Needs more parens :P

trichoplax

Some are redundant but I left them in for symmetry...

Sparr

that form parallels the fact that fibonacci numbers follow the golden ratio.

trichoplax

yes I just replaced phi with the root 5 formula rather than define it separately

(phi**n - (-phi)**(-n)) / 5**0.5

It looks much nicer on wikipedia

Mitch Schwartz

@Lembik your probability problem is solvable with dp

ideone.com/SAViLx

Rainbolt

Any problem is solvable with Dr. Pepper

Mitch Schwartz

8:36 PM

idk how interested i am in optimising it though

i mean, python is not the ideal language since your scoring method is based purely on speed

maybe a golfing score is better?

but with the requirement that it has to have a similar speed to what i posted

Sparr

there are too many ascii art challenges lately :(

something like 1/4 of new challenges are ascii-art

trichoplax

Does that mean we have too many AA challenges or too few other challenges?

I have a non-AA challenge in the sandbox that I can't decide whether to time limit or not. Other than that it's ready to post (I think)

4

A: Sandbox for Proposed Challenges

Worst Case Manhattan Exclusion code-golf geometry grid packing Imagine a W by H grid of squares that wraps toroidally. Items are placed onto the grid as follows. The first item can be placed on any square, but subsequent items must not be within a Manhattan distance R of any previous item (als...

Compass

I think I figured out the formula for the xkcd page thing

A bit bloated

trichoplax

8:51 PM

And I had gone and convinced myself it was impossible...

Compass

function f(n) {
o = 0;
for(i = 0; i < n; i++) {
s = o.toString(3);
k = s.length;
b = 0;
for(j = 0; j < k; j++) {
if(s[j]==2)
o = o + Math.pow(3,k-j-1),b=1;
}
if(!b) o++;
}
return o.toString(3);
}

ew

well it's javascript

its not pretty

trichoplax

if it was pretty it wouldn't be golfed enough...

Compass

basically, it's a lot of base conversion, and if any digit is 2

it propagates upwards

Rainbolt

We have a fire drill coming up any minute now and it's pouring down rain :(

Lembik

@MitchSchwartz Well... that's very impressive!