00:00 - 16:0016:00 - 23:00

4:02 PM
[can't find the question, but it's about code generation]
and the issue I'm facing is basically that. You write (source) code into TIO, choose Code Golf submission (Stack Exchange) part (get generated code), and add (source) explanation to it.

Then when you change the source code, you have to copy the explanation to correct place, which is often quite hard, partially because TIO permalink is long and SE answer box is small.

@user202729 You're aware you can extend that box, right?

That's not the main problem.

Then please type a complete thought instead of messages that are only one sentence so I can read what the actual issue is, bud ;)

4:09 PM
Why is "ais523" hidden from the active user list?

You mean the avatar?

Yes.

Can you explain this to me: https://tio.run/##Sy7WTS5O/v@/tDgzL10huLK4JDXXmosrOSexuFjBkauaSwEICkqTcjKTFYpLEkuAVFl@ZoqCb2JmnoamAli6movTOT@vOD8nVS@8KLMkVcPQQM8SFZgo2NoqGBpqWoM11HLV/v8PAA

Is it because of the binary representation?

@user202729 So your problem is that it's hard to edit the explanation on a corrected post?

4:12 PM
the last line is false, although it has the least amount of difference

It's hard to update the code and use proper TIO link.

Or that it's hard to edit a post when you're just copying from a formulaic post?

@LucaH Well, what's wrong?

@user202729 yeah, that would be valid

@user202729 Right, I run into the same problem

4:12 PM
Great, -1 byte. (MMA is so smart that it treats 0^0 as indeterminate)
@LucaH They are rounded to the nearest (double), so that's expected behavior.

Why don't you just do it right in the first place so you don't have to edit? :P

Generally, I have little problem updating those links manually, but I'm one of those weirdos who are okay with a little data entry

@LucaH yessir that'd be the reason. You can thank IEEE 754

You think it's easy? No. I often have to rewrite the whole explanation.

So do I
It's a pain in the ass
Especially in Actually and Jelly

4:14 PM
@LucaH C# lets you use numeric literals far more precise than it can actually store in memory: tio.run/##Sy7WTS5O/v@/…

But that's often the price of a good but extensive golf

The first two are considered "true" because of floating point error (can't remember the correct term), right? But I mean you would expect .91 and .92 to round to the same number, instead of .74 and .91... It's simply not intuitive

Not in decimal it isn't, but in binary it could be

yeah I thought so... and binary is not intuitive for humans :D

No, but it was a fine enough hack for early computer programmers to assign numbers to bits and leave it there. Now nobody wants to go back and try to implement binary-coded decimal or some other decimal method

4:18 PM
Would you expect 0.4999999999 and 0.500000001 to round to same number (int)? Certainly not.

@user202729 true...
Ok but please, where does this come from: tio.run/##y0osS9TNL0jNy0rJ/v8/…
it's like... wtf?
@user202729 this was a good explanation, but what the hell is going on now?

@Pavel Indeed. From today's internal code golf email thread: The dismembered chipmunk could save us a character, but it's not a very fast creature.

@Adám Why do you guys have an internal code golf email thread in the first place.

@LucaH I fixed it ;)

@LucaH 10.999999999999998 is shorter than 10.9999999999999991. It is probably the nearest available floating point number (for a given definition of nearest)

4:21 PM
@LucaH Just round to different numbers.

@KSmarts ಠ_ಠ

@Pavel Because PPCG has scoring rules which are disadvantageous to APL.

Yes, the 10.9..991 > 10.9..98.

@KSmarts great fix :D
@Sherlock9 yeah but why isn't 10.9999999999999992 also rounded to 10.9999999999999998? It would be more correct, wouldn't it?

I don't know what the definition of "nearest" is in those cases

4:24 PM
@Adám Why? We have really flexible rules for I/o and answer format, and APL can be scored as one byte/character

oh well, still thank you! Programming is weird sometimes ^^

APL doesn't use a custom code page, so a lot of its built-ins count as two bytes
From what I remember, anyway

@LucaH Floats lose precision with longer numbers
Imagine a float as a string of digits and a decimal point, the further rough the decimal point is, the larger the number you can represent, but there's less values after the point so less precision
@Sherlock9 No it does

Which is a rough explanation of "losing precision". The more detailed explanation has to do with the limitations of having only 53 bytes on hand.

@Pavel Almost all newly added built-ins are not in the standard SBCS, even though the SBCS has plenty of unused characters.

4:28 PM
@LucaH For sure. Definitely try looking up the IEEE floating point standard if it interests you. There are some good resources out there that help explain it

@Pavel do you mean "right" instead of "rough"?

But it's neither here nor there. Point is, you put too many 9s and the floating-point system started shoving those numbers into strange places, which are probably less confusing when shown in binary

If I remember correctly, you can only achieve a given "amount" of precision with a given storage space for mantissa and exponent, right?

Finally complete it... this.

@LucaH Yeah, that's pretty much the reasoning exactly :D

4:35 PM
^ must take 1-based... forgot to add in.

0

Factor Trees code-golf factoring (new tag!) I once saw on the xkcd fora a format for expressing numbers in an odd way. In this "factor tree" format: The empty string is 1. Concatenation represents multiplication. A number n enclosed in parentheses (or any paired characters) represents the n...

@Sherlock9 Yeah, for some APLs, that's true:
39

Prompted by this. The question of APL's encoding often comes up, and many times a helpful soul links to Wikipedia's article on the APL EBCDIC codepage. However, each implementation of APL has its own encoding rules. What are they?

4:51 PM
CMP: The 16 characters ɫ¯¥£¢ý·€⍫´∣¶¿¡ § are in the Dyalog APL character set but are not used in any way (ASCII, built-ins, box drawing, identifiers). The six characters ⊆⍠⍤⌸⌺⍸ are built-ins which are not in the character set (i.e. they force the entire solution into Unicode). Should we count the six characters as one byte each because it is technically possible to swap them for included characters?

@Adám No, but publish this script somewhere and say you're using a custom encoding interpretted by it: tr 'ɫ¯¥£¢ý' '⊆⍠⍤⌸⌺⍸'

@Pavel Ah, that would be enough?

@Adám Yep. Your encoding defines the characters on the right as being represented by the APL codepoints of the ones one the left, and now you have an implementation of Dyalog APL that accepts that encoding, made by tacking the tr script in front of the actual dyalog call.
I think it would work like that.
Might need to also first convert APL bytes to Unicode bytes first.

5:11 PM

@cairdcoinheringaahing Done
@LucaH Yes. Autocorrect.

6:05 PM
0

Find a fixed point META: This is just meant as an introductory challenge for functional input (black box functions) and hopefully establish some more inputs. code-golfmathoptimization Given an integer x(1) and some black box function f: ℤ → ℤ find a fixed point of the sequence x(k+1) := f(x(k)...

6:27 PM
Q: What symbol should be used for "Convert to/from character" in Whispers?

what's Whispers?

My new language

I mean, I expected that answer, but I'd like to see the codepage, theme and philosophy behind it. Got a repo? A wiki? something?

@NieDzejkob Repo, no wiki just yet, no code page

TNB is not going to do your work for you. Make crappy decisions like the rest of us and live with it.

6:33 PM
SURROUND_ATOMS is not very golfy if you'd ask me
anyway, how about '?
or "?
it's like a character literal

@El'endiaStarman Calm down, it's just asking for a suggestion
@NieDzejkob Hmm, that makes sense. Thanks!

@cairdcoinheringaahing I think this was meant as a joke

Half joke. I appreciate the desire to have meaningful characters for functions, but at some point, you just can't really do it the way you want. One could at least come up with a few possibilities as opposed to the whole ASCII range.

6:57 PM
Ugh, why did that "smallest number of participants" challenge get reopened? There are still open questions regarding rounding precision and direction, and it still doesn't make sense how "Rounded to the nearest hundredth" results in 8.97 going to 9.

7:29 PM
3

Today's google doodle has a code-golf slant to it

@El'endiaStarman realistic
the last part ;p

@DigitalTrauma Yup :-)

7:44 PM
@Mr.Xcoder oh, I see. I guess I didn't search back enough pages in the chat transcript

Anyway it's cool

@Mr.Xcoder yep. Such was my excitement at google's embrace of codegolf that I didn't even look at the starred comments!!!

:( I didn't get the shortest solution to 4 (yet)

8:03 PM
arxiv's favicon always confuses me
it's a happy face on a lime bg with some bones?

The favicon looks like it wants to hack my computer, and thinks that a favicon is a good way to do so.

8:32 PM
@orlp if you looked below, you would find that I turned it into a matrix, and found its two complicated eigenvalues

Anonymous
8:54 PM
Any more feedback on this (specifically in reference to additional test cases)?

@LeakyNun I see the matrix
but not the eigenvalues

@Mego I don’t think more test cases are needed. Looks ok to me

but yes, that is one way to solve it

11 hours ago, by Leaky Nun
(x-1)(x-4)-3 = x^2-5x+1 = 0 --> x=(5+-sqrt(21))/2 rip

@LeakyNun well, that's part of the solution :P

9:04 PM
the rest is trivial to me

3

After testing a couple $a_n$ by hand it appears to converge. So let's try to solve: $$x = \frac{x+3}{x+4}$$ $$x(x + 4) = x+3$$ $$x^2 +3x - 3 = 0$$ $$x = \frac{1}{2}(\sqrt{21} - 3)$$ But if we want to be rigorous and prove its convergence: $$a_n = \frac{p_n}{q_n}$$ \frac{p_{n+1}}{q_{n+1}} = ...

@LeakyNun I solved it with generating functions :P

@orlp it's the same

@LeakyNun in a sense
but the interesting thing is that I used two generating functions
and substituted one in the other to solve them

@Adám not sure whether or when I'll get around to this, but can I send you a list of commonly used Unicode operators in Mathematica for your language bar?

9:17 PM
@MartinEnder Absolutely.

9:32 PM
@cairdcoinheringaahing Candidates: ' ⍘ ⍞

9:45 PM
poll: for comments, (specifically python here but applies to other langs) do you leave a space between the comment chars and the text?
i.e.
# test like this
vs
#test like this

absolutely the former

@Riker Yes, always.

@Riker no
The latter

@DJMcMayhem tha'ts nasty

@Riker The first one
@Adám Well, I never said ASCII only :P

9:58 PM
@cairdcoinheringaahing Right. I looked at the source and saw INFIX = '=≠><≥≤+−±×÷*%∆∩∪⊆⊂⊄⊅⊃⊇\∈∉∧∨⊕' PREFIX = '∑∏#√' etc. so I figured.

anyone know what to do if the left side of a keyboard does not work?

@Zacharý One of these?

...

@Zacharý you kontinu on it typing nomilly

i hate using the onscreen one

10:03 PM

@Zacharý I can make you a 12345/QWERTY/ASDFG/ZXCVB language bar

@Riker The former (with a space between # and the rest of the comment). python.org/dev/peps/pep-0008/#inline-comments

my r shift doesn't work either

@Zacharý Good thing you have two Shift keys then.

While I'm typing, I pretty much only use the left shift key. Is that horribly wrong?

10:06 PM
... the left one is on the left
@DJMcMayhem i usually do

@Zacharý Wait, so no Shift key works?

@Adám no (currently using onscreen for left only

@Zacharý Sounds like it is time for a new keyboard.

@DJMcMayhem i'm too pretty to tweet :[

10:10 PM
dyalog ... now this .... ugh

@Zacharý What?

the thing didn't want to install, im concerned with this.

@Zacharý What happened?
@Zacharý Try typing We reserved seats at a secret Starcraft fest.

   .
nope.

@El'endiaStarman did not know that was the standard, makes sense though

10:17 PM
@Riker Yeah, flake8 (Python linter) enforces it by default.

@El'endiaStarman Even on the first line if it begins with #! ?

@Adám Well, that's an exception since that's a shebang, not a Python comment. (Though it's technically also a Python comment.) ... Actually, lemme check.

Obviously not.
ll, i'm oin o y o i hi no, you uy hopully oon

Yeah, flake8 doesn't complain about the shebang.

@El'endiaStarman Good. Would have been awkward.

10:25 PM
0

Birthday Bash, Part 1: Invitations/Index code-golf ascii-art date January 29, 2018 will be my 16th birthday, so in the style of the Advent Challenges set, I'll be posting a series of 8 or so birthday-party-themed challenges. You can find the rest of the challenges in the sidebar. As everyone...

10:46 PM
@AdmBorkBork Can you help me out with a PowerShell question? The prompt function returns the string you want to be your prompt, how can I define it to print a colored prompt?
Like say I want it to be \$> but in Red.

00:00 - 16:0016:00 - 23:00