CodeWalrus

CodeWalrus Website => Contests => Topic started by: Yuki on February 28, 2017, 05:03:02 AM

Title: CWick Contest #4: The Tau Problem
Post by: Yuki on February 28, 2017, 05:03:02 AM
Wait what? It's back? Well, then. Today I ask you the 1000 first digits of tau. I think @JWinslow23 will love this one. So if you remember how it's done, I ask you a program that have this output exactly (or more):

6.283185307179586476925286766559005768394338798750211641949889184615632812572417997256069650684234135964296173026564613294187689219101164463450718816256962234900568205403877042211119289245897909860763928857621951331866892256951296467573566330542403818291297133846920697220908653296426787214520498282547449174013212631176349763041841925658508183430728735785180720022661061097640933042768293903883023218866114540731519183906184372234763865223586210237096148924759925499134703771505449782455876366023898259667346724881313286172042789892790449474381404359721887405541078434352586353504769349636935338810264001136254290527121655571542685515579218347274357442936881802449906860293099170742101584559378517847084039912224258043921728068836319627259549542619921037414422699999996745956099902119463465632192637190048918910693816605285044616506689370070523862376342020006275677505773175066416762841234355338294607196506980857510937462319125727764707575187503915563715561064342453613226003855753222391818432840398

You're allowed mostly everything, a calc (scrolling or output to a Str variable might be needed), a PC, a Google Home that speaks the output, whatever. This is also a code golf contest: smaller is better, but it will also be judged on how original the method to produce or to show the output is, be it an algorithm, compression or whatever, and also how fast it is. You need not to have that output exactly, variations are allowed, as long as it shows more or less these digits. Standard code golf pitfalls from Stack Overflow applies.

Send the sources of your entries to [email protected]. Contest ends Monday March 13th 23:59:59 EST. (Two weeks this time, because it's a big one.)

Note that the schedule is now whenever I feel like. Feel free to send ideas for future contests like this.
Title: Re: CWick Contest #4: The Tau Problem
Post by: gameblabla on February 28, 2017, 06:53:09 AM
Tau on the Colecovision :
1k of RAM just isn't enough for Tau...

(https://gameblabla.nl/img/coleco_tau.png)

I will send you the binaries and the sauce.
Title: Re: CWick Contest #4: The Tau Problem
Post by: Yuki on February 28, 2017, 07:14:42 AM
Oooh, that nice. On that note, 1000 digits is not a hard limit if the system you're coding for don't allow for it, although there's algorithms that lets you calculate tau digit by digit, you might get around by throwing away the digit once it's displayed on screen.
Title: Re: CWick Contest #4: The Tau Problem
Post by: Unicorn on February 28, 2017, 07:41:43 AM
This reminds me... Did i ever get my Hello World walrii points? :P

Oh yes, and now i have an interesting idea.. not sure if i could pull it off tho.
Title: Re: CWick Contest #4: The Tau Problem
Post by: gameblabla on February 28, 2017, 07:45:36 AM
Ok juju, i found a workaround for the small memory. (just do it in several parts)
Tau fits on two-screen !
I will send you the new version
Title: Re: CWick Contest #4: The Tau Problem
Post by: Yuki on February 28, 2017, 08:18:25 AM
Quote from: Unicorn on February 28, 2017, 07:41:43 AM
This reminds me... Did i ever get my Hello World walrii points? :P
Ah yeah, the WalriiPoints. I'll look at this later.

EDIT: Also, just a tip, you shouldn't merely store the output in the program and print it, some transformation or generation should be done.
EDIT2: Tip #2: If the language you picked only allows for a limited number of decimals, you should find an algorithm that can go over that limit. There's some algos that can print arbitrary digits.
Title: Re: CWick Contest #4: The Tau Problem
Post by: Dream of Omnimaga on February 28, 2017, 05:13:07 PM
Woah nice @gameblabla . Also @Juju if 1000 is not an hard limit, then does it mean that the following code would have been acceptable in TI-83 Plus BASIC (at the cost of being nearly guaranteed to lose)? :trollface:

Title: Re: CWick Contest #4: The Tau Problem
Post by: Yuki on February 28, 2017, 08:08:37 PM
I should put both of those tips I added in the post above in the OP.

EDIT: Here's my entry:

https://www.youtube.com/watch?v=aOsbMWuVA-Y
Title: Re: CWick Contest #4: The Tau Problem
Post by: JWinslow23 on March 01, 2017, 01:33:56 AM
I just now noticed this.

I suppose I have to come up with something. TI-BASIC would be my target platform if I get this done, let's see if I can encode 1000 digits in less than 1KB. :P

EDIT: I already did this in <1KB, actually. For this specific problem, I optimized the previously-done (therefore invalid :P ) solution to 879 bytes, though it may not finish in the lifespan of your battery.

Goto 0
Lbl 1
If min(L₁)>0 and ᴇ10≥max(L₁
End
0→PMT
For([recursiven],1,dim(L₁
PMT+L₁([recursiven]→|N
int(|Nᴇ⁻10→PMT
|N-PMTᴇ10→L₁([recursiven]
End
If not(PMT
End
[recursiven]→dim(L₂
Ans→dim(ʟQ
Ans→dim(ʟR
Ans→dim(ʟT
PMT→L₁(Ans
End
Lbl 2
1+dim(L₁→|N
Repeat Ans or |N=1
|N-1→|N
L₁(|N)-L₂(|N
End
End
Lbl 0
SetUpEditor T,R,Q
{30→T
6Ans→R
{1→Q
ClrHome
"?→Str1
2→I%
Lbl 3
If 1<length(Str1
Then
"             
If not(fPart((length(Str1)-2)/16
Output(6+C,1,sub(Ans+Ans,1,16
2+16int((length(Str1)-114)/16)(130≤length(Str1
Output(1,1,sub(Str1,Ans,1-Ans+length(Str1
End
5ʟT→L₁
For(θ,⁻1,0
If θ
Goto 1
End
L₁→L₂
(9I%-4)ʟQ→L₁
For(θ,⁻1,0
If θ
Goto 1
End
3L₁+5ʟR→L₁
For(θ,⁻1,0
If θ
Goto 1
End
0→PV
For(θ,⁻1,0
If θ
Goto 2
End
While Ans≥0
PV+1→PV
L₁-L₂→L₁
For(θ,⁻2,0
If θ=⁻2
Goto 1
If θ
Goto 2
End
End
sub("0123456789",PV+1,1
If I%=2
Ans+".
Str1+Ans→Str1
(5I%-2)ʟQ+ʟR-PVʟT→L₁
For(θ,⁻1,0
If θ
Goto 1
End
30L₁→L₁
For(θ,⁻1,0
If θ
Goto 1
End
If I%<33
Then
9I%²+9I%+2
Else
(3I%+1)L₁→L₁
For(θ,⁻1,0
If θ
Goto 1
End
3I%+2
End
AnsL₁→L₁
For(θ,⁻1,0
If θ
Goto 1
End
L₁→ʟR
10I%ʟQ→L₁
For(θ,⁻1,0
If θ
Goto 1
End
(2I%-1)L₁→L₁
For(θ,⁻1,0
If θ
Goto 1
End
L₁→ʟQ
3ʟT→L₁
For(θ,⁻1,0
If θ
Goto 1
End
If I%<33
Then
9I%²+9I%+2
Else
(3I%+1)L₁→L₁
For(θ,⁻1,0
If θ
Goto 1
End
3I%+2
End
AnsL₁→L₁
For(θ,⁻1,0
If θ
Goto 1
End
L₁→ʟT
I%+1→I%
If getKey≠45 and 1003>length(Str1
Goto 3
If 1003=length(Str1
Pause
ClrHome
Title: Re: CWick Contest #4: The Tau Problem
Post by: c4ooo on March 01, 2017, 03:21:41 AM
If someone could post an algorithm to find all the terms of tao in succession i would be willing to golf something up in axe.
Title: Re: CWick Contest #4: The Tau Problem
Post by: JWinslow23 on March 01, 2017, 03:55:39 AM
Quote from: c4ooo on March 01, 2017, 03:21:41 AM
If someone could post an algorithm to find all the terms of tao in succession i would be willing to golf something up in axe.
*tau

And I just posted one. :P

Also, I'm working on my TI-BASIC version, and right now, it's 505 digits in 428 bytes (including the 6. at the beginning). That's about 1.18 digits per byte!
Title: Re: CWick Contest #4: The Tau Problem
Post by: c4ooo on March 01, 2017, 09:00:42 PM
Quote from: JWinslow23 on March 01, 2017, 03:55:39 AM
Quote from: c4ooo on March 01, 2017, 03:21:41 AM
If someone could post an algorithm to find all the terms of tao in succession i would be willing to golf something up in axe.
*tau

And I just posted one. :P
One that isn't in TI-BASIC :P
Title: Re: CWick Contest #4: The Tau Problem
Post by: JWinslow23 on March 01, 2017, 09:04:16 PM
This paper has some in Haskell for pi. (http://www.cs.ox.ac.uk/jeremy.gibbons/publications/spigot.pdf) Adapting them to tau shouldn't be hard.
Title: Re: CWick Contest #4: The Tau Problem
Post by: JWinslow23 on March 02, 2017, 06:33:37 PM
Bump.

I've been working on the spigot, and if I let it go infinitely and display one digit at a time, I can get it down to 383 bytes.
{30→T
6Ans→R
{1→Q
ClrHome
For(I,2,ᴇ4
⁻1→V
5ʟT
For(θ,V,0
If θ
Goto 1
End
ʟA→B
Ans+3(9I-4)ʟQ+5ʟR→A
Repeat Ans<0
IS>(V,9
ʟA-ʟB
For(θ,⁻1,0
If θ
Goto 1
End
End
Disp V
(5I-2)ʟQ+ʟR-VʟT
For(W,0,1
For(Z,1,2
If W
(3I+Z)ʟA
For(θ,⁻1,0
If θ
Goto 1
End
30ʟA
End
End
ʟA→R
10IʟQ
For(W,⁻1,0
For(Z,1,2
If not(W
(3I+Z)ʟA
For(θ,⁻1,0
If θ
Goto 1
End
(2I-1)ʟA
End
If W
ʟA→Q
If W
3ʟT→A
End
ʟA→T
End
Lbl 1
DelVar PAns→A
For(M,1,dim(ʟA
P+ʟA(M→N
int(Ansᴇ⁻9→P
N-Ansᴇ9→ʟA(M
End
If P
P→ʟA(M
dim(ʟA→dim(ʟB
Ans→dim(ʟQ
Ans→dim(ʟR
Ans→dim(ʟT
Ans→N
Repeat Ans or not(N
ʟA(N)-ʟB(N
DS<(N,0
End
End
Title: Re: CWick Contest #4: The Tau Problem
Post by: kotu on March 03, 2017, 04:22:34 AM
I'm not joining in because it looks too hard but I found this.....

"Are you trying compute ALL the digits from the first to the Nth? Or are you just trying to get a few digits around the Nth binary digit? There is no reason to use the BBP formula if you want all the digits as there are MUCH faster formulas out there for Pi."
Title: Re: CWick Contest #4: The Tau Problem
Post by: p2 on March 03, 2017, 11:51:22 AM
hehe force the tau beliefers to use a pi formula  :thumbsup:
Title: Re: CWick Contest #4: The Tau Problem
Post by: kotu on March 03, 2017, 12:01:03 PM
:trollface:

const double PI_DIV_2   = 1.5707963267948966192313216916398;
const double PI                = 3.1415926535897932384626433832795;
const double PI_MUL_2 = 6.283185307179586476925286766559;

the ultimate solution!!  ;)
Title: Re: CWick Contest #4: The Tau Problem
Post by: JWinslow23 on March 03, 2017, 12:15:21 PM
Quote from: kotu on March 03, 2017, 12:01:03 PM
:trollface:

const double PI_DIV_2   = 1.5707963267948966192313216916398;
const double PI                = 3.1415926535897932384626433832795;
const double PI_MUL_2 = 6.283185307179586476925286766559;

the ultimate solution!!  ;)
You mean ETA, PI, and TAU. ;)
Title: Re: CWick Contest #4: The Tau Problem
Post by: kotu on March 03, 2017, 12:23:27 PM
er, no   :banghead:   lol
Title: Re: CWick Contest #4: The Tau Problem
Post by: p2 on March 03, 2017, 12:24:59 PM
why isn't there nyone backing ETA as the one true constant...? ???
Title: Re: CWick Contest #4: The Tau Problem
Post by: kotu on March 03, 2017, 12:26:39 PM
Lance Henriksen  8)
Title: Re: CWick Contest #4: The Tau Problem
Post by: JWinslow23 on March 03, 2017, 12:35:15 PM
Quote from: p2 on March 03, 2017, 12:24:59 PM
why isn't there nyone backing ETA as the one true constant...? ???
There is...
https://www.youtube.com/watch?v=1qpVdwizdvI
...but these people are rare to find, and the idea of eta being the "true" circle constant largely hinges on the argument that the right angle should be the fundamental unit, instead of the area of a unit circle with pi, or the circumference of a unit circle with tau. Which, as you know, I think is silly.
Title: Re: CWick Contest #4: The Tau Problem
Post by: Yuki on March 14, 2017, 02:59:38 AM
Reminder that the contest ends tonight. I won't mind if anyone's late due to timezones or anything, but make sure your entry is submitted by email before I process the results. I'll show the results tomorrow on Pi (or Tau/2) Day live on YouTube or Twitch, stay tuned.
Title: Re: CWick Contest #4: The Tau Problem
Post by: Dream of Omnimaga on April 04, 2017, 10:50:09 PM
Were the results ever posted @Juju ?
Title: Re: CWick Contest #4: The Tau Problem
Post by: Yuki on June 05, 2017, 04:31:49 AM
@JWinslow23 wins, stay tuned for contest #5.

JWinslow23

- Code is mostly unreadable and obfuscated (+)
- Uses clever tricks with integers (+)
- Sometimes misses leading 0's (-)
- Program is less than 1K (+)
- 9/10

gameblabla

- Uses an obscure console (+)
- Tau fits on 2 screens (+)
- Tau is stored in variables, then compressed and finally uncompressed for seemingly no reason (-)
- Source code is very readable (-)
- 6/10
Title: Re: CWick Contest #4: The Tau Problem
Post by: JWinslow23 on June 05, 2017, 01:59:09 PM
Actually, the "sometimes missing leading 0's" is not a mistake. My program basically took numbers in base 100 and displayed them, and the numbers in question were digits of tau. Leading 0's on numbers seemed to be impossible to display, so I just trailed the previous numbers with 0 and then continued. All the digits are correct and in the right order.

For example:

8125724179
972560696
5068423413

If I'd have done that "optimally" and made 9725606965 the second line, the third line would have a leading 0 (which I couldn't figure out how to display).

Hope it makes sense now!
Title: Re: CWick Contest #4: The Tau Problem
Post by: Dream of Omnimaga on June 05, 2017, 04:34:31 PM
Congrats!
Title: Re: CWick Contest #4: The Tau Problem
Post by: kotu on June 05, 2017, 04:40:27 PM
teach him how to program in a professional environment why don't ya!!

go @Juju !!! well done!
Title: Re: CWick Contest #4: The Tau Problem
Post by: E37 on June 05, 2017, 08:31:53 PM
If no one looks at the source, you could just display the first ~15 digits and just print random numbers after that :trollface:
Title: Re: CWick Contest #4: The Tau Problem
Post by: gameblabla on June 09, 2017, 10:38:13 PM
I did not know that deliberately obfuscating the code was a feature.
Richard Stallman told me otherwise :(