The Nineteenth Byte

Not JoeMama, a rickroll link.

@user I know, right!

but in all seriousness, I hate curly braces and indentation because I hate statements. The only time you need (;) or (,) is statements and tuples/lists in SML

What's the alternative to statements?

Expressions

expressions, recursion, and parameter passing

11:02 PM

Ninja'd

1/3

I like how Rust does it

the implicit return?

You can write code as if things are statements, but things like if/else are actually expressions

Implicit return is gross, not that part

Wow, Java lets you embed Rickroll links in code /s

yeah ruby did a surprisingly large amount right, minus newlines and keywords

11:03 PM

@RedwolfPrograms Why is it gross?

What is this, TI-BASIC? Should I use the Ans variable to get the result? :p

@user How is that not one of Java's many SyntaxErrors

@RedwolfPrograms To me, val foo = if (something) { return x } else { return y } just looks awkward

@Ausername It's a label

Try it in JS and C too

@user What on earth is that

having return at all is gross. I disagree with trailing semicolons meaning anything though

11:05 PM

Why is there both return and val foo =

Is this a function or not

@Ausername The stuff after // is a comment, that's ignored. The statement System.out.println is considered has the label https

@RedwolfPrograms That's what I was saying, it looks gross

@user Oh.

@RedwolfPrograms val foo = if (something) x else y makes much more sense

But...yeah. I agree with that.

"Implicit return" or whatever it's called makes sense

11:06 PM

You shouldn't need to return from an if/else statement

val foo = something ? x : y

What were you talking about, then?

Functions

def foo(bar: Int) = {
  //some computations
  baz(bar, food)
}
What about that is gross?

@user line comments are gross :)

11:08 PM

Idk, maybe just lots of time using JS _/(.-.)\_

What is that emoticon?

Upside down shrugging

lol

@user ~~I see you used my indent style~~ :(

@Wzl Lisp has line comments

11:09 PM

so I'm fixing that

lisp made other mistakes like string escapes too

@Wzl I think I need therapy now

@Wzl What's wrong with escapes? They're very useful. Allowiong raw literals in addition to that is important, though

String escapes are bad now?!

(actually I'm considering abandoning all string literals because they cause problems)

@RedwolfPrograms YES

How???

@Wzl "Actually, I'm considering abandoning all food because it causes problems" :P

"Allowing my interpreter to run arbitrary code is causing issues"

11:12 PM

@user say you make a new regex flavor that supports one extra character class like \w, now you have to use double backslashes or, slightly better, use a different escape character too, which you now have to double to include verbatim

@Wzl This is why you also allow raw string literals

whereas with APL/smalltalk/SQL literals you can just have a function that does escaping

So "foo\"\nbar" for normal stuff, and raw"foo\w\n" for regex and stuff

What if you want the string "'

@Wzl I find that extremely annoying

11:13 PM

^

@user yes, except for

Why

@user and I find backslashes annoying

How do those cause problems?

@Wzl Use another character, then

@RedwolfPrograms they remove greppability and the ability to parse (easily) with regex

11:14 PM

~~I find people annoying. That doesn't mean I'm going to eliminate them all~~

@Wzl So you're moving the burden onto the user instead of the interpreter

@Wzl Parsing with regex isn't an important thing

That's not good design

@RedwolfPrograms that's why I'm still thinking about it.

Any moderately complex language probably can't be parsed with regex, that's not what it was made for

11:15 PM

It was designed explicitly for HTML

but with raw literals and a regex function, it's moveing the burden to the standard library, not the user

If you really want to make parsing easier, make regex better. Maybe mix it with BNF

@user I hate moderately complex languages btw

@RedwolfPrograms I can't tell if ~~you're joking~~ you were making a joke

@Wzl I can't tell if you were making a joke either

@user regexes are already not regular expressions anymore, but i do think grammars are important, in a library

11:17 PM

@Wzl The user still has to deal with annoying stdlib functions instead of having native support

I like my regexp like thing where you could use (name:abc.+) to define things for recursion

@user I'm laughing at myself, but I'm not joking

^^^

A language's syntax should be designed to be easy to use and understand. That does mean it should be simple, but not that you can validate/match it with a regex flavor

And there's a balance between simple to interpret and simple to write

11:18 PM

@user you have to use functions to use regex anyway (outside of perl). So replace and match automatically convert to regex. simple.

Brainfuck is pretty simple, but not to use

@user debatable :P

Even if you have a grammar that regex could be used on, why use it when you could parse it properly? (maybe tokenize it first)

@Wzl Simple for the implementor (hau two spel?), not the user

@Wzl Everything's debatable if you're brave enough

JS has regexp literals, maybe we should all just write languages with those :p

@user '.*' match is not any harder than .match(/.*/)

11:20 PM

CMP: Who thinks all syntaxes (syntices??) should be made as complex as possible so that Wzl gets annoyed? :P

3

@Wzl Okay?

It looks simpler, but I don't see how that helps your point

that's the syntax I'm proposing

Oh ok

What about regexes with \b and friends?

Still don't see how this is easier for people, though

^^

Or with ' in them?

11:23 PM

You're always going to have to use some form of escape characters, unless you decide to do 'foo' + chr(32) + 'bar' to represent the string foo'bar (no idea what the ASCII actually is :) )

@RedwolfPrograms '/\b''' match versus .match(/\/\b'/)

So '' for a '? That's still escaping.

^

yes, but without taking any more characters, or hardcoding any shortcuts

Plus I'd argue \' is more readable

11:24 PM

It's just a very limited, very annoying form of escaping

And using \ makes adding things like \n convenient

I'm getting too involved in this debate, so I'm going to leave now. o/

o\/

4

@RedwolfPrograms newline is more readable :)

You're not getting too involved in this debate, so I'm going to leave now. o/, you're user!

Misspelt Java:

11:29 PM

I probably shouldn't have started this chat/argument/debate in the first place :/, but I like to rant

What is the debate?

see above link

ah

pablik claz man{
    tatik man(stwig[] argz){
      siztem.owt.pritLin('Hello, world!')
    }
}

Misspelt Java

*subscribes*

11:32 PM

Misspelt Python?

forest

*Pithon

Yes

public class Foo {
   public @NotNull String @Nullable [] @NotNull [] foo(String[]... a[]) @Nullable [] { return null; }
}

I'm actually not sure if that's syntactically valid, but I'm like 70% sure it is, and that it fulfills its method contract

You're not actually not sure if that's syntactically valid, but I'm like 70% sure it is, and that it fulfills its method contract, you're user!

Permission to make a bot that does that automatically so I don't have to?

PLEASE DON'T

11:36 PM

Awwww....

@ hyperneutrino please delete this username as a sock of lyxal

:P

forest

Is this the troll that I heard about on MSE?

@Wzl Don't blame A username, it's just that they have lyxalitis

@Wzl Hey!

11:37 PM

@forest No, Wzl was kidding

forest

oh lol

@user Oh no...

The troll doesn't seem to have come back again after that

@forest Lyxal is a user who started saying "No, you're not JoKing, you're <insert name>" whenever someone said "I'm joking, ..."

(context: JoKing is a mod here)

@user Lyxal started that?

Yes

11:38 PM

I think so. That's why lyxalitis was named after them

GRT seems to be working

GRT?

I'll wait for one more suggested edit task, then I'll release it :p

@Ausername Generic Review Tool

An alternative to Review Stalker Reloaded

There are never any reviews :P

And a much better one, IMO :p

11:40 PM

@Ausername it seems to be related to goats

@Ausername That's why sniping them is so important

@Wzl Um...

@RedwolfPrograms Use NMP!

For?

@Wzl or an exclamation used only by T Abraham

CMQ: Feature suggestions for GRT?

11:44 PM

Waits a minute for everyone but you so you always get them?

I don't really care.

@RedwolfPrograms I've always thought goats needed wheels for practicality

and style

I'm thinking an option to ignore certain review types would be nice, that was a request I saw for RSR

For 1.0 I'll keep it very minimal though.

false == undefined and false == null are some of the dumbest parts of JS

Same with false == 0 and undefined == 0 (and null == 0)

Why does this work?!

It turns out you can't use the legacy array notation with varargs, but you can still do some crazy stuff

@Wzl Me too. Mountain goats might need some modifications, but even so, a couple wheels would look very chic

Would the wheels replace the legs, or just be at the bottom of them?

Wheels on legs would look bad.

Nah, they'd be attached on the sides

11:49 PM

Okay good

The things attaching the wheels to the goats would be able to be moved by the goats so they could climb and stuff

Also, there should be functionality to let the goats spin around and make them transcend to Downgoats

@RedwolfPrograms both

*doesn't clarify*

Legs on wheels, duh

@user I'm thiking of naming this "Java by a drunk C programmer who dropped out of college"

@Wzl Hmm, so there'd be wheels on wheels? Sounds really interesting

CMQ: -> or => (or both, with different purposes?)

11:53 PM

-> for value lambdas, => for type lambdas

Type lambdas?

@RedwolfPrograms configurable by the user :P

@RedwolfPrograms Yeah, lambdas, but for types

I guess that's not very helpful, let me find a link

like a list is a function from a type to a type (int to a list of ints)?

Type Lambdas (this page is also not very helpful, but whatever)

@Wzl Pretty much, yeah. Type lambdas would allow you to define new type constructors (and higher-kinded types) inline

11:55 PM

I'll just make them synonyms

I want this to be mostly compatible with JS, aside from adding and removing what's obviously necessary

Scala 3 uses the syntax [K] =>> Map[K, String], but I think =>> looks ugly

@RedwolfPrograms Then just keep a fat arraow, and leave the thin arrow for maps or something

I like thin arrows though :p

Then screw JS and use just a thin arrow. Screw JS no matter what you do :P

@Wzl I seriously think this is the best solution (ask me why and I'll tell you, but it'll take an hour)

@Wzl If different people use different arrows, they'll be confused when they see another person's code base. Configuring arrows should be okay, but I think some amount of standardization is necessary

11:57 PM

Should function x() { ... } also be allowed, or should I just get rid of that altogether?