Please no

Geez people, I leave for half an hour and you nuke all my stars? Fascists!

:P

Don't you mean communists? We all want everyone to have no stars

so the troller was a Philanthropist?

Meh, one definition for fascism in MW is:

> a tendency toward or actual exercise of strong autocratic or dictatorial control

Such as a mod nuking my stars instead of letting the people decide what should be starred :P

I vote to star everything

Unfortunately, llamas don't get to vote. Humans only. (Sorry Alex)

@Geobits That's speciesist!

What about me? I mean a subatomic particles is way too small to use a keyboard. Of course I'm human ;)

I agree! We should do something about it. I didn't make the rules, but we could form a lobby or something to get them changed.

@Doorknob you don't know who the troller was yet ?

@Optimizer Yet or ever, but it doesn't really matter.

(If it becomes a recurring issue of course there are things that can be done.)

Agreed. If it's not an everyday thing I don't see the point tracking them down.

Ninja'd? Ugh.

Aww it would be fun though

also, there should be some star limit, right ?

@Optimizer There is a limit.

oh

.starLimit = properties.enoughToPissOffAMod;

5?

2^31-1

20.

Before SE just crashes horribly

(Per day, applies to each room separately.)

Welcome, SuperScript! Nice avatar, btw.

He's like a neutral Geobits!

@Doorknob Is there really a 20 star limit per day/room? You know I'm gonna have to test that some time...

@Geobits Yeah, it's probably documented on meta.SE somewhere.

(which means you don't have to experiment here ;) )

there is always sandbox

Awww, but.... Martin, he's being mean to me!

On a serious note, I'm about to head out. Going to the Tampa Comic Con, have fun!

wearing Deadpool costume?

So it is Tampa! :D

@BetaDecay No, I don't live there. But a Comic Con within reach is one worth going to. Staying at a hotel :P

@BetaDecay damn, your plans to dognap his dog are ruined

Haha how would I smuggle a dog on a plane though?

you can take them normallly

most airlines allow

Yeah but Geobits would find me while his dog was in quarantine

Probably

Yeah, he'd just look for the two floating red and green numbers. Impossible to miss.

or he'll just go search in basement. Nothing can be lower than that

Haha I'll wear a cloak

@MartinBüttner what is the shortest Pyth sequence to make something unique?

e.g. TESTING STRING -> TESING R

err

I wanted to notify @Maltysen, not @MartinBüttner, sorry

shortest I have right now is u+G-HG"TESTING STRING"k

@orlp for the four square?

I did oxzN{z

ah yes, that's better than mine

mind if I steal it?

go for it

@orlp since { no longer actually uses sets, we should have it be in this order from before

@Maltysen you mean in Pyth5? in Pyth5 { will just be 'uniquify'

yeah, that. It should preserve initial order

yes

not sure yet what { should be on reals though

maybe make that double

if getting rid of y

well, there's quite a lot of things to still be decided

the old y will go on P for powerset, was my thought

since the old P is now on T

ah I see

also, for the rationals, I was thinking to get the numerator, denominator representation just as a list of two elements

not two separate functions

k

so relying on j\/ for stringification

depends

if a challenge really wants you to print in that format, yes

@Maltysen I suggest you take a look at quickref.txt

in Pyth5, to see exactly what is where right now

@orlp the binary string seems useless, its the same as CM", only a one char save for an entire symbol.

@Maltysen It's one character shorter, it makes sense on that symbol, and it's not exactly the same

for example \r\n becomes just a newline in regular strings, but stays exactly \r\n in binary strings

@Geobits Have a great time! I do hope you'll check in with us while you're away though. We get worried.

We don't get worried. Geobits has built in Sat Nav.

Haha

Mandatory cranial implant when he joined the site.

@Maltysen an important difference between Pyth 5 and 4 is that Pyth5 has a proper lexer, and thus has absolutely no problems with any non-printable characters in (binary) strings

so embedding data has become quite a bit easier for Kolmogorov complexity questions

@orlp Proper lexer? Versus what?

@AlexA. an ad-hoc parser that assumes the input is an UTF-8 string :P

Is that not a safe assumption?

no, we might have null bytes and what not inside binary strings

Okay. I meant barring stuff like that.

the thing is, you can't bar that

you have to lex the file as a binary format - you can't assume any encoding because it will mess up the binary strings

@orlp I can because I don't mess with unprintables. :P

Yeah that's fair

Pyth 5 sounds like it will be pretty cool.

Any syntactic differences?

@AlexA. a ton

a lot of stuff got changed / moved around

So it will be mutually unintelligible with Pyth 4?

I'd argue that readability went up a bit

obviously most of the functions are unknown to people

That was a poor choice of words on my part.

I guess I mean compatibility

but now a, b, c, d and e are the temporary variables used in lambda functions and stuff

yes, it's not backwards compatible at all

I think nearly all Pyth4 programs need changing to run in Pyth5

Changing in a way that reduces their byte counts? ;)

For the most part byte counts will stay the same

Ah okay.

Will there be a built-in for the mean?

yes

it's in Pyth4 now too though

Really?

since 18 hours, yes :P

.O

Oh, okay. So too recent for Jakube to have used in a recent Pyth solution of his where he computed the mean manually.

I was surprised seeing that because I assumed there'd be a built-in

it's a reactionary addition to that challenge, I think

Ah, okay.

but most of the improvement in Pyth5 are in documentation, readability, interpreter quality and making future-proof backwards-incompatible changes

for example every lambda function now FIRST has the collection, THEN the lambda function

(like map)

also, every lambda function now uses the same variables: a, b, c, d, e

No wonder I couldn't get map to work right the last time I used it. I was using it backwards.

if you look in the documentation you see this

 m 2 map        q:  List from mapping over sequence using lambda d -> k -> b.
                r:  Convert to range using U before mapping.

the : stands for implicit lambda function

oh wow

haha those docs are outdated because they still state old variables

sec gotta edit those out

Will Pyth 5 be extensible via some kind of Pyth module system or will all features going forward be part of the core?

We don't have any plans for some module system

what would you mean?

Like if someone were to desire Pyth functionality that wouldn't necessarily make sense to have as part of the core. I can't think of a really good example (because perhaps there isn't one) but say a module for base conversion-related functions, supporting negative bases, counting ones in binary representations, stuff like that.

I don't know, man. I'm just thinking aloud.

meh

Or if someone wanted to compute p-values from an F-distribution

if it's useful enough we can add it to the core

if not, whatever

That's probably a more maintainable approach than trying to make a module system.

One unexplained downvote on my question and an unupvote on its only answer. Wonder what happened...

:(

Btw @Dennis, I love the reference to the Wheel of Blame in your challenge.

Let's spin it to see who downvoted the question. :P

Hmm. It claims it was Geobits.

Then again, it claims that Geobits is responsible for everything.

I'm beginning to think that the wheel is rigged.