« first day (2686 days earlier)      last day (2147 days later) » 

6:24 PM
<LBPHacker> oh yeah it's log2(3)*29-bit
I swear, how is a log2(3)*(2**29) word size possible
<LBPHacker> 29.584962500721-bit
 
hi @H.PWiz, I saw your improvement on ascii's answer for the 10-adic cbrt(3)
I wanted to ask how you got that, simply rearranging it gives me d×10^k≡(3-x^3)/7, but I don't know how you remove the 7 from the bottom
from what I understand, the 7 is removed from ascii's answer because he switches to mod 10, where 7 is the inverse of 3
wait what tio.run/…
 
6:54 PM
anyone here built tensorflow w/ cuda before
I have GCC-7, 5, and 4.9 yet for some reason bazel keeps complaining that my build tool isn't GCC.
 
oh wait what was I thinking, you just simplified ascii's expression by getting rid of the division by p, why did I not see that
 
I've set it to /usr/local/bin/gcc-7 which is definetly gcc
> external/local_config_cuda/crosstool/clang/bin/crosstool_wrapper_driver_is_not_g‌​cc
T_T bazel y you use clang when gcc explicitly specify
 
@Cowsquack Your byte save doesn't work for n=2
@Cowsquack Have you understood what I did now?
 
7:36 PM
CMC: Create a non-empty cat program, and, in the same language, a (non-empty) reverse cat program (i.e. reverse the input). Your score is the Levenshtein distance between the two programs. Lowest wins, and the two programs may not (somehow) be identical.
 
@cairdcoinheringaahing Will the input be on one line?
 
@cairdcoinheringaahing As in and ?
 
@DJMcMayhem I'll say no, assuming it isn't immensely difficult to take multiline input in your language
 
wow, apparently it is a bug in bazel and not my fault: gist.githubusercontent.com/pavelmalik/…
 
@Adám Yep, that works
 
7:38 PM
@cairdcoinheringaahing So that's Dyalog APL with a score of 1.
 
@cairdcoinheringaahing So each line and the order of the lines should be reversed?
 
@DJMcMayhem Yes. Example:
hello,
world!
=>
!dlrow
,olleh
 
blech
 
I'm guessing you're thinking about doing it in BrainFlak?
 
No, V
 
7:40 PM
0
Q: Equal Pair sum of 10?

Muhammad SalmanChallenge : Check : if there are exactly n exclamation marks ! between every pair and whether their sum equals to 10. Input : A string of arbitary length. The number of exclamation marks to check for Output : if there are exactly n exclamation marks between each pair, return truth...

 
0
A: Sandbox for Proposed Challenges

sundarDecoding the Kaadi system code-golf string parsing alphabet You have come across an old Indian manuscript, one that describes mounds of buried treasure. The manuscript also tells you the location of the treasure, except that some crucial numbers have been encoded indirectly into the text. You f...

 
@cairdcoinheringaahing Dyalog APL, score 1: and
 
@cairdcoinheringaahing JS, f=c=>"!"? c : [...c].reverse().join('') & f=c=>""? c : [...c].reverse().join(''). You should've added bytecount as secondary scorer.
 
@NewMainPosts sighs I honestly don't have the effort to recommend the Sandbox
@dzaima ¯\_(ツ)_/¯ It's just a CMC
 
7:43 PM
Ugh, can't retract delete votes
 
@cairdcoinheringaahing V, reversed: Try it online!, and regular: Try it online!
That was much harder than it should be
 
@DJMcMayhem Nice!
 
I have a 2-point brain-flak answer.
Anyone wanna try to figure it out, or should I just post it?
 
Add++ scores 2: Reversed and Normal
@DJMcMayhem Spoiler link?
 
You should try to figure it out!
 
7:46 PM
I can't brainflak to save my life :P
 
Although, I suspect the change might be [...] to (...) (reversed to normal) Nevermind :P
 
That's got to be optimal
@cairdcoinheringaahing Keep in mind that [...] changes the value but has no side-effect
One nice way to think of brain-flak is that every atom has a value and a side-effect. So (...) has the side-effect of pushing n, but the value is just n (no change). Whereas [...] has no side-effect but the value is -n and <...> has no side-effect and the value is 0 (which is surprisingly useful)
 
Yeah, I've never understood the point of <...>
 
@DJMcMayhem Just add a comment to reverse cat for regular cat. Score: 1
 
7:51 PM
Oh yeah, doi
 
@H.PWiz let me get this straight, dn**3/p*7%10*p(n**3-3)/p*7%10*p (backing up)≡7*(n**3-3)%p (you remove the division, now you take mod p because that is equivalent to taking mod 10 on d after dividing by p) tada
 
@cairdcoinheringaahing It lets you get at values under the TOS. So without it, if we wanted to duplicate 'b', then when we push back 'a', the value of duplicating 'b' is 'b', so it pushes back 'a+b' instead of 'a'.
For example, ({}(({}))) turns [a, b] into [a+b, b, b]. But ({}<(({}))>) turns [a, b] into [a, b, b]
Does that make sense?
 
@DJMcMayhem Just working through it at the moment
So the first {} pops a, then the <(({}))> is executed, which pops b, then pushes it twice, then a+0 is pushed back to the stack with the outer (...). So shouldn't that make [a, b] => [b, b, a]?
 
@Cowsquack Pretty much, except you're really calculating 10**k*d. I didn't get it by directly modifying ASCII-only's answer, but just thought of it because n**3 is 10**k*(10*a+b)+3 where b is what you want to multiply by 7
 
No because <(({}))> finishes executing before ({}...) does
But your first sentence is correct
 
7:57 PM
@DJMcMayhem Isn't that what I said? b is pushed twice before a is repushed?
 
@cairdcoinheringaahing Oh wait, this might be a terminology thing
In [a, b] is a or b the TOS?
 
I assumed a is the TOS as you said "It lets you get at values under the TOS"
 
@H.PWiz oh yeah, I forgot about the 10**k*d, nice finding btw
 
So ({} --> [b]; <(({} --> []; ) --> [b]; ) --> [b, b]; >) --> [a, b, b]
 
and there's xnor's 33 byter, that seems very interesting
 
8:01 PM
@DJMcMayhem Ok, I'm following you now, I had the stack the other way to you
 
in xnor's answer, 10**k*2/3+1 calculates 1/3 (mod 10**k). I'm not sure exaclty why it is valid to bring the mod up into the exponent though
 
ah so that just becomes 3^1/3 (mod 10**k)
@H.PWiz because you only need the last k digits?
so that's why the mod inverse is 10**k...
 
@cairdcoinheringaahing I think this might make for a nice challenge if the code length is also incorporated into the score, because it's probably rather easy to do something like if" ">""then id else reverse in most languages.
 
@cairdcoinheringaahing BTW, without <...>, it's still possible to do the same thing. <...> just makes it more convenient. Without it, we could do this: Try it online!
 
8:16 PM
@MuhammadSalman (don't want to continue in the comments) apparently, my comment was 12 seconds before your edit, according to the timestamps
@EriktheOutgolfer : Done 2 secs before your comment :) — Muhammad Salman 1 min ago
anyway, it looks like a few comments including mine are obsolete now...
 
+1 because every Stack Exchange site needs a "Why am I not dead?" question. — Adam Haun 5 hours ago
O.o
 
8:32 PM
@EriktheOutgolfer : Cheers.
 
> I golf code for fun. How am I not dead from my brain exploding at all this weird code?
:P
 
8:52 PM
CMC Last n digits of 0! + 1! + 2! + ...
 
I don't think that's really possible, unless you've forgotten an "n!" over there
or maybe it's just another case of 10-adic numbers :P
 
yes, 10-adic numbers
 
btw, since I'm Greek, I always read "10-adic numbers" as "decadic numbers" and not as "ten-adic numbers" lol
(δέκα = dhéka = ten)
 
@EriktheOutgolfer since the last n digits of (5n)! are all zero...
 
n ≤ 5n
 
9:06 PM
@EriktheOutgolfer I think it is possible, as the trailing digits of +/!⍳∞ stabilise.
 
@Adám well, of course the trivial answer would be, e.g. in Jelly ”0ẋ, i.e. a string with n zeroes
 
@LeakyNun Can you give us a few first?
 
@LeakyNun SOGL, 12 bytes
 
9:44 PM
@EriktheOutgolfer That transliteration scheme tho'
 
well, IPA makes stuff harder to understand :P
 
The <h>?!
I've seen it as déka
 
no, it's not ντέκα, it's δέκα!
dh ≠ d
dh == th(is)
 
I know
 
d == d(ata)
 
9:49 PM
I've just mostly seen grapheme-grapheme transliteration schemes before, where ντ is transcribed as nt, and maybe all the i sounds being i
 
that's pretty much not how we pronounce that in modern Greece :P
 
I know, I've looked at modern greek phonology.
 
I've seen such nonsense a lot of times
 
@EriktheOutgolfer Why? The last few digits of +/!⍳∞ are …0940314.
 
@Adám it depends on whether you follow common sense or Leaky Nun sense :P
 
9:54 PM
the number ends ...20420940314
@dzaima so this is invalid
 
@LeakyNun ಠ_ಠ off-by-one error. fixed
 
@LeakyNun Haskell: 36 bytes
 
The all-new version of (Q)Nial has been released: github.com/danlm/Nial-dev. Last change was a while ago, but this version wasn't previously public.
 
@mınxomaτ Wow, I thought it was dead. Maybe time for Dennis to try adding YAA to TIO.
 
Just because dev was slow, doesn't mean nothing happened. Takes some time to adjust when every contributor has a different goal for the language ;)
Sadly, most of the dev history is sort of lost, because before moving to GitHub, development meant sending the latest tarballs to a mailing list :D
 
10:25 PM
doesn't that still exist? if so, then duh :/
 
10:39 PM
Woah, never knew Jelly added so many two-byte nilads.
 
the δέκadic power series Σ n! x^n should converge for every x
 
*δεκádic
 
how did you put that accent there, it isn't even greek
oh, that isn't that accent
so let f(x) = Σ n! x^n, then f'(x) = Σ (n+1).(n+1)! x^n = Σ [(n+2)! - (n+1)!] x^n = [f(x) - 1 - x]/x^2 - [f(x) - 1]/x
x^2 f'(x) = (1-x) f(x) - 1
 
@H.PWiz 35: Try it online!
 
Oh, thanks.sometimes I forget that backticks aren't always short
 
10:54 PM
I also forget it from time to time. There's actually a tip pointing it out.
 
11:39 PM
@Cowsquack i was hoping it would be faster for large numbers. not sure if it actually is (plus exec removes the performance difference)
 
11:55 PM
:D @H.PWiz I think the new method will be pretty useful for quite a few of the submissions (well, at least the java one I guess? nvm. xnor's answer blew mine out of the water >_>
 

« first day (2686 days earlier)      last day (2147 days later) »