The Nineteenth Byte

7:20 AM

Sadly, that is not accepted as the answer to 32 across

For context, the answer that came before that one was MEMEACCOUNT

ATaco

8:14 AM

Oh neat, Synesthesia

I got 200 rep on CGCC and got 100 rep on all other sites

Very surprising!

That's just the assocation bonus

@ATaco isn’t that “seeing sounds”?

ATaco

Not exclusively, iirc

Synthesesia is just confusing different senses

so it could be seeing sounds or smelling touch or any kind of confusion

8:25 AM

@lyxal 2 is red, 3 is yellow, 4 is green, 5 is light blue, 6 is green.

9:42 AM

@pxeger Would it be a good idea to make the 2021 edition of Rust the default on ATO?

Jacob

11:54 AM

is ato broken for everyone or just me?

Seems to be fine for me

works fine for me

status-norepro

Hello, World!

12:16 PM

0

A: "Hello, World!"

Num Stack Arithmac, \$165\times\log_{256}(16)=82.5\$ Bytes ++^++++^++++++++q+++++++++++++++++++++++++++++q+++++++qq+++q-%++^-^---^++++++++q------------q%+++^^++++++q++++++++++++++++++++++++q+++q------q--------q%++^-^---^---q Well, there are actually 13 builtins (Including <> for comments but not...

pxeger

12:30 PM

@mousetail That sounds like a good idea - how can I set that?

You can set --edition 2021 as the default command line argument

pxeger

Is that overrideable again by a later --edition argument?

I don't want to set it if there's no way for the user to unset it

Seems no unfortunatly :/

ato.pxeger.com/…

Would it be possible to just set the default thing visible in the options pane? That way you also avoid breaking existing ATO links that might rely on old behavior

1:16 PM

@hyper-neutrino status-completed

Jacob

2:39 PM

@Jacob ok it works on my iPhone on iOS 16, but is broken on my iPad with iOS 14

Simd

is there an easy way to set up computer chess so that it plays itself on linux?

how does the new posts bot retrieve posts from se

It uses the websocket

wss://qa.sockets.stackexchange.com/

(It also uses the API, but only to check when the question is fully posted, since otherwise it's so fast it posts plain links instead of oneboxes)

@RydwolfPrograms is the api available to the public?

Yes: api.stackexchange.com

2:48 PM

The API is quite solid, rate limits are also generous

hyper-neutrino

3:25 PM

@Ginger what

@hyper-neutrino there is (another) CGCC minecraft server now

hyper-neutrino

(...) well are you going to link it :P (although never mind, I found the chat room for it)

lol

I'm going to be doing another vanilla SMP once everyone's tired of Ginger's

;-;

3:34 PM

Was gonna do it when school ended but I don't want to split the available playerbase in two

I mean there's nothing stopping us from using both at once lol

Well what killed the last one was it was using up too much of our time

Players' time is the most limited resource for running a small server, not ports or CPU cores

3:49 PM

that's why I put "Abandon all work, ye who enter here" in the MOTD q:

4:15 PM

@Ginger cool what's the ip?

@PlaceReporter99 see chat.stackexchange.com/rooms/146156/the-butter-factory

@hyper-neutrino oh wait

nobody can join yet tho

Why not use the CGCC gaming room

That's where past PPCG/CGCC Minecraft servers have talked

4:50 PM

@RydwolfPrograms the chat relay isn't discriminatory; it sends all messages in its room to the server

ah

I put it in its own room instead of the gaming one so that people talking about unrelated things wouldn't be forwarded to the server

ooh nice

7:27 PM

Okay I need help proving something relating to an esolang I'm working on

So instead of brackets, it uses | as block delimiters

To disambiguate left and right, you can put a : on one side, which is the opening side

The rules being that any | with a : on one side but not the other will be forced to be that bracket

Is it possible to unambiguously represent any structure of brackets and meaningful non-bracket characters using this?

are brackets are guaranteed to be balanced?

@RydwolfPrograms I'm not sure why, but I don't get it.

also do you want | and : to be meaningful on their own as non brackets?

: is, but only in situations where it's not adjacent to a bracket

@DarrenSmith Yes

Like 2 * |: 1 + 3 | is 8?

Except this particular case doesn't need disambiguating, so you could simply do:

2 * | 1 + 3 |

?

7:33 PM

Yep

Whereas |a|b|c| could either be [a]b[c] or [a[b]c]

So you'd need to do something like |a|:b|c| to disambiguate

I don't know if it's because I only just had my cup of coffee, but I'm still not processing the question.

@RydwolfPrograms Ah, okay.

so |:| is (), and |:|:|| is ||:|:| is (())

Why are you using brackets instead of parens, though?

@RydwolfPrograms That confused me, too.

@UnrelatedString yep

@RydwolfPrograms Why would it not be possible, if every single opening bracket was a |:, though?

7:36 PM

i feel like we can take the "meaningful non-bracket characters" out of the question since when you have other stuff in there it can only ever make it easier to disambiguate the structure

@RydwolfPrograms Well, there could be situations were |:|:||/||:|:| wouldn't be (())

ok so the only ambiguous things would be :|: and |:|
|:| isn't sensical since has | by itself unless another :, but then that becomes normal case
:|: sounds like it also isn't intended to be defined since : means something special if next to a bracket and both are

@AviFS In |:|:, the first one's unambiguously [, but the second has a : on both sides

so really we just have to consider if every dyck word can be represented this way

Ooh risky google that paid off

7:38 PM

@DarrenSmith |:| has the colon affect both bars, :|: is a colon on both sides so it can be either

lmao

@RydwolfPrograms Oh, I didn't understand that :| was also a closing bracket, somehow. I thought this was just about opening brackets.

@AviFS |: is an opening bracket, but :|: can be either

@UnrelatedString But isn't :|: by itself malformed?

ok so then you should be able to define any paren structure (with those requirements on not overloading solo : and | )

@AviFS No, it just doesn't clarify anything

A | doesn't require a :, and a : doesn't need to be used

7:39 PM

@AviFS in isolation, yes, but in |:|: it has to close

@DarrenSmith yeah it seems like it has to work

@UnrelatedString So are we interpreting this as one of [[/]], or are we ignoring the second colon to make it []?

just having every opening bracket followed by a colon seems like the plat

> The rules being that any | with a : on one side but not the other will be forced to be that bracket

you don't actually group the colons in with the pipes, the colons are just context for the pipes

BTW if you are using an infix grammar that doesn't have implicit ops, then you actually would never need : because atom|op would mean atom)op and op|atom would mean op)atom. There will never been two atoms in a row, and two operators in a row are handled by same case

So can you ignore the colons if they're on both sides?

Yes

7:44 PM

Or do you group the first colon with the first |, and then only have one remaining colon to parse with the second |?

@UnrelatedString Oh, I see. Wow, that took me a while to understand.

Seggan

if i need to run async code in sync code in an async context, the right way to do that would be Handle::current().block_on right?

|:| is unambiguous, a concatenation with no exterior colons is unambiguous, and it seems like any nesting of |: ... | should still be unambiguous

@UnrelatedString I was gonna say the same.

time to break out the test harness from that one challenge

@UnrelatedString You could prob prove it recursively like this.

7:47 PM

yeah that's what i was kinda angling for

To extra clarify though, is :|:, in isolation, exactly equivalent to |?

But ::|: is not equivalent to :|, correct?

@AviFS Assuming it's not adjacent to any other brackets, yes

@AviFS Correct, it's identical to :|:

@UnrelatedString Note that there's no whitespace in this lang

Got it. So :: is equivalent to :.

|:|:|| is unambiguous because the first pipe has to open and the third pipe has to close, |:|:|:|||... i think could represent both (())() and ((()))

@RydwolfPrograms yeah i assumed as much, it was just notational clarity :P

ah I see

@UnrelatedString and ()()()

7:50 PM

oh yikes

so it looks like the simple colon-after-every-open strategy won't cut it

@UnrelatedString I'm thinking first replacing every :: with :, and then every :|: with | would be a useful first step in any proof/reduction.

@UnrelatedString So in this case, I believe this would be equivalent to:

[ | | ] | |

Funnily enough it would be perfectly possible if : was directional

Ie |: :|: :|: :| | | => |: | | :| | |

@AviFS exactly

If it only disambiguated ] there'd be no issues

7:55 PM

Should we make both of the possibilities unambiguous as an exercise?

Wait, are you allowed to use spaces...

Cuz what about |: |: :| :| |: :|

6 mins ago, by Rydwolf Programs

@UnrelatedString Note that there's no whitespace in this lang

My bad, sorry.

Any other null-ops that could serve as space? Cuz that would solve your unambiguous representation issue pretty easily. Even if it's not maximally golfy.

I think I'm just gonna make : only apply when it's to the left of a |

Wait, no, I really don't think you can do it. I don't think it's possible to force a [[

You can do |:, but then if you want another, you can only do |:|:, but that doesn't disambiguate it.

Your only options, that I can think of, would be |:|, |:|:, |::|:, |:|::, none of which do the right thing.

And that goes for any n-length string of opening or closing brackets-- only the first opening/last closing can be unambiguous. Right?

Or am I just being slower than everyone-- maybe y'all already got that.

No you're entirely correct

8:02 PM

@RydwolfPrograms You could do the annoying thing of allowing more colons on one side than the other to disambiguate, hehe

So [[ would be |:|::

And if you wanted [[[ you'd have to do |:|::|:::

That may be possible, but I don't know how undesirable/fun that is in your lang

So [[[]]] would be |:|::|:::|::|:|

And [[]][] would be |:|::|:||:|

I think that works!!

Kinda cool, actually, it's like binding by precedence

Except precedence is unary

And it almost looks like the colons themselves get nested-- they visually tell you something about the structure of the parens if you compare those two. They look very different, and have a nice symmetry. Again, assuming those samples are correct.

Could be cool for a CGCC challenge, or another #esolang, if not this one

@UnrelatedString you're good at this

@RydwolfPrograms Why not just introduce a bracket character or something, at that point?

Because : in other contexts can serve as a third operator

And you're trying to minimize symbols?

Yep

Dotcomma 2.0

Is this a tarpit, or is this a practical esolang?

@RydwolfPrograms Link to the 1.0?

github.com/Radvylf/dotcomma

Evidently I starred it, but I don't remember it at all, haha

Yup you found it first

8:16 PM

Huh, I see.

And you're trying to get rid of the bracket characters for 2.0?

I need a third operator for storage

To make it TC

And slightly more usable

But you're trying to get rid of the brackets?

Yes

Hmm

How did you decide to keep the colon as closing bracket, rather than an opening one?

Left has an e in it

8:20 PM

Obv most langs have the more verbose header/footer for any structure at the beginning, so it's an interesting choice. Eg procedure/loop/conditionals all have meta-info, and anything that would be important for disambiguation, at the beginning, rather than the end, so it's an interesting design choice.

@RydwolfPrograms So does opening.

Exactly

(Or maybe it was just arbitrary, lol. Thought maybe there was a reason based on the structure/aesthetics of a typical program)

Alright, off! o/

@AviFS Also, what'd you think? not for this lang necessarily

@AviFS No I'm serious

Left has an e and I don't like e

Left brackets get nothing special

@RydwolfPrograms Yeah, but left would make it closing.

And opening also has an e. And left brackets are the ones that are getting special treatment. I'm so confused.

No left as in left bracket

Q: Two java files. Getting IllegalAccessError when running class with main method trying to access a method from the other file

8:24 PM

Oh, left as in opening, not left-facing

\o/

6

Learning Java. I have two files, each containing one java class. When I run the file with the main method, I get the following error: Exception in thread "main" java.lang.IllegalAccessError: failed to access class TapeDeck from class TapeDeckTestDrive (TapeDeck is in unnamed module of loader 'a...

java methods compiler-errors

My answer has twice as many upvotes as the question, haha

nice

11:22 PM

@RydwolfPrograms hate to tell you, but | is sometimes called a "pipe" and that's got an e in it

It's also called "vertical bar" which also has an e in it

You've gotten so caught up in the brackets you forgot to check the replacement :p

11:46 PM

Hey @lyxal I had another question for you about the golf elo of langs. The script you use doesn't count the information contained in flags towards the byte count right (since the answers themselves don't)?

it doesn't

It seems to strip the flags and all other information after the first space

lang = re.match("([^-,\(:]*)", line)

that's the regex it uses

actually how does it strip flags

I see, I get that a flag makes it technically a different language so it seems kinda fair for purposes of answer on stack exchange, but seems kinda unfair to not count that information when grouping flags as all same langauge for elo. One could encode entire programs as flags (I think someone did that)

ohhh because when people put flags in backticks, it gets converted to <code>flags</code> which is removed by the HTML cleaning regexes

so it's not purposefully removing them (I think) it's more just an oversight

I see

@DarrenSmith counter argument: people only submit non-flag answers when there's nothing shorter with flags. By only counting the non-flag submissions, you get an even more unfair representation of the language

and the other flag variants (probably) wont appear because of the 40 answer minimum cutoff

11:52 PM

That would be the downside to counting them as separate languages. But what about adding to the byte count 1 byte per flag char?

then that doesn't account for the reason flags are used in the first place

with that byte count added, people might structure their programs differently, meaning that you've got answers on the site written with one scoring system in mind, while being scored differently elsewhere

That is also true, but the current system is definitely unfair. Suppose I just added option to convert flags base64 to the program, every program could be 0 bytes filesize

@DarrenSmith and then what about languages where it's only 1 flag, but the flag is 8 bytes?

should those languages be penalised for something that others would only be penalised with +1 score?

ELO isn't the best measure anyway - if people selectively use languages in situations where you know it'll be the shortest, it's not a true representation

you'd need a standardised set of challenges to compare against

well on an individual answer, there can be multiple answers and a language is only competing with itself, people typically list versions both with and without flags. The issue I have is more that languages like Vyxal typically use about 1ish flags and there are 50ish flags, so they are getting about 6 bits extra of information compared to languages that don't use flags

yeah I know ELO isn't the best system and the whole thing is a little silly because you can game whatever system is created, I just think it could represent things a little better and more fairly