The Nineteenth Byte

Mega Man

19:02

Any tip to contact someone via SE or Github?

Dennis

On SE, you only have chat.

Mega Man

I know, can I contact people who aren't on the chat via chat?

DJMcMayhem

If you ping them @username they'll get a notification later

Mega Man

How much repo to make private chats?

Downgoat

only mods can make private chats IIRC

Mega Man

19:08

argh

Conor O'Brien

have you tried skype?

Mega Man

I do not think Skype could run smoothly on a Raspi...

Conor O'Brien

ah

Dada

@MegaMan Maybe leave a comment on an old post?

Mega Man

I know something else, but I don't have my password for that thing... codingame.com

Dennis

19:09

@MegaMan Private chats can only be created for moderation purposes. All interactions on SE are meant to be public.

Mega Man

@Dada Do contact someone? Why not XD?

Kritixi Lithos

@MegaMan I tried blowing a 10*10*10 area full of TNTs on Minecraft on a Raspberry Pi 2, the lag was real

Mega Man

@Dennis Hm. It's ok, I think

@KritixiLithos True lag for the win...

I do not think saying hello in comments is a good idea

Loading codingame.com on a raspi... Mh...

Conor O'Brien

int superSeriousFunction(string seriousName){
    // hello from a comment!
    cout << "bwahhh it's " << seriousName << " haha!";
    return 42;
}

3

Martin Ender

@ais523 regarding the reusability of your prolog answer: meta.codegolf.stackexchange.com/q/4939/8478

Kritixi Lithos

19:14

@ConorO'Brien You didn't return int :(

Conor O'Brien

@KritixiLithos that's why it's a serious func

ais523

I answered it both ways because it's not clear what should happen in a situation where the function is an infinite loop

Mega Man

Does anyone know how to get a link to a username on codingame.com?

Conor O'Brien

:P fixing

Martin Ender

@ais523 and also meta.codegolf.stackexchange.com/q/4782/8478

ais523

19:15

and thus you can't run any code after it anyway

I could adjust the program to quit on the first solution but it'd make it longer and also less interesting

Kritixi Lithos

Nice! I use 42 too for returning ints on C++!

Mega Man

@ConorO'Brien If you want to be serious, you need to return int32 because of compatibility purposes :P

Conor O'Brien

@MegaMan huh? since when?

ais523

I guess this is one of those problems where I don't really care about competing on a level playing field because I doubt there will be any other submissions anyway, but if people really care, I can add some extra code to handle it

Mega Man

@ConorO'Brien Since the day I started to hate Visual C++ XD

Conor O'Brien

19:16

@MegaMan Visual C++ sounds painful o_o

ais523

simplest would be to exit the interpreter entirely just after submitting a solution; what are the rules for full programs in languages where the default input method for a full program is to open a REPL?

Martin Ender

@ais523 I'm afraid terminating is required unless the challenge allows infinite loops. you could still include the infinite loop version separately to show a program that finds all solutions.

Mega Man

@ConorO'Brien I did never use it ;)

ais523

the thing about the way Prolog works is, a full-program submission is hard to define because after you compile and run your program

Mega Man

icedove&

ais523

19:18

the program opens up a REPL

so is it acceptable to ask the user to run a particular function in that REPL to give it its input?

Martin Ender

hm good question

I have no clue to be honest

ais523

half the reason we have the rules we do is to avoid having to worry about this sort of thing when the interesting thing is the algorithm

but the way Prolog works, often I need more code to handle I/O than I do for the actual program, even when all reasonable input formats are allowed

because Prolog thinks in terms of constraints rather than data

Mega Man

I'm just notifying someone to keep track of him/her: @TuukkaX

Martin Ender

@ais523 time to port to Brachylog ;)

ais523

I disagree, I've tried to use Brachylog

it is bad at any thing remotely complex

it's still shorter than Prolog but it's way too verbose for a golfing language

and it's sufficiently bad at handling data structures that I'm not even sure it'd be shorter

for a program like this which makes heavy use of tuples

Mega Man

19:24

verbose==evil python==good python==verbose -> good==evil???

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@DJMcMayhem congrats!

DJMcMayhem

Thanks. :)

Up next is electorate, which I'll probably get with a week or two

Eᴀsᴛᴇʀʟʏ Iʀᴋ

cool!

what's that one for again?

Mega Man

Sounds like something moderation-based

DJMcMayhem

600 votes on ~~questions~~ challenges

17 to go

Mega Man

19:30

Then go and vote ;)

New Main Posts

0

Q: Generate all combinations of given list of elements, sorted

Make a Code that takes a list and an number as input, and now generates all possible combinations with the length of the number. For example, with the list {0,1} and the number 2 : 00 01 10 11 Your program hasnt to expect characters twice or more often in the list, such as {0,0,0,0,0,1,1,5,5}...

Mega Man

Bye :)

Conor O'Brien

@DJMcMayhem you won't get it. you need to have a high % of questions voted as opposed to answers

Dada

go unvote the answers you voted on then!

DJMcMayhem

@ConorO'Brien Yeah, 25% or more. I'm at 30-some

Conor O'Brien

19:42

oh lol

I'm at 19%

DJMcMayhem

Haha. You probably deserve electorate more, you've got like 5 times as many votes as me

Kritixi Lithos

Ranked by month, I have the 3rd most number of votes casted

Conor O'Brien

@DJMcMayhem yeah, but that also means I need 5x the votes to offset it :P

Martin Ender

@ConorO'Brien oh hey, looks like you'll overtake me in votes pretty soon

Conor O'Brien

\o/

BlueEyedBeast

19:58

@DJMcMayhem only 30% on questions? I've got 40%

Alex A.

Figured you folks might enjoy this: leftoversalad.com/c/015_programmingpeople

13

El'endia Starman

Now that I've used Sublime at work for a few days and learned some of the keyboard shortcuts, I can see why they named it that.

DJMcMayhem

@AlexA. that is beautifully accurate

PhiNotPi

@El'endiaStarman because you're transitioning directly from solid to gas?

El'endia Starman

Yes. Why else?

Flp.Tkc

20:09

@ETHproductions Is it cool for me to write & submit some code-golf answers in cubix? Also, is there anywhere to get a permalink on the online interpreter?

H Walters

"Cubix is a South Korean animated television series created by Cinepix."

Alex A.

Obviously

Flp.Tkc

@HWalters ethproductions.github.io/cubix

H Walters

@Flp.Tkc Interesting

El'endia Starman

@AlexA. Comic 19 is...dang.

Kritixi Lithos

20:15

Most of the comics reveal a sad reality of life

Alex A.

@El'endiaStarman D':

Kritixi Lithos

They're all sad... Now time to go to sleep

DJMcMayhem

@El'endiaStarman thanks a lot

ErikGolferエリックゴルファー

@AlexA. ...are you okay? You seem to want to go to your corner and cry. And please, do not do what ^ says (holding tears worsens the situation), you can freely cry at TNB.

TuxCopter

Go on a leftoversalad.com page and type flappyBird() in the dev console

flawr

20:19

TuxCopter

There is a textual flappy bird

El'endia Starman

@flawr Oooooh resonances of a guitar body?

ErikGolferエリックゴルファー

@TuxCopter 1337

flawr

@El'endiaStarman =)

Alex A.

Java vs. JavaScript

4

H Walters

20:27

She should have just asked Java to hold an object

ErikGolferエリックゴルファー

@HWalters Yeah, what will she get when she wants her coffee back then? 0.00000000000001% the coffee, 99.99999999999999% trash, or something more awful.

H Walters

Everything in Java is an object

Except primitives

ErikGolferエリックゴルファー

I mean, Java does not know which object to give when she asks for her object back.

betseg

hey

H Walters

Don't understand... Java's referenced based... it's really good at holding objects

ErikGolferエリックゴルファー

20:34

I don't think "an object" references anything lol.... that is, unless Java.

betseg

Nov 2 at 14:15, by betseg

JavaScript is just scripting extension to Java anyways.

ErikGolferエリックゴルファー

^ Wrong! 0/10!

JavaScript is not an extension to Java AFAIK.

betseg

you took that seriously? ಠ_ಠ

Alex A.

JavaScript and Java are unrelated. The creators of JavaScript changed the name from LiveScript to JS to capitalize on the popularity of Java at the time.

Dennis

Also, JavaScript does not look like Java.

ErikGolferエリックゴルファー

20:37

...you took that seriously? ಠ_ಠ×2

New Meta Posts

0

Q: Should PPCG have an official, objective, list of rules?

This is something that's been bothering me for a while, and is really impairing my use of the site. Any competition needs rules in order to be fair, and this has been established repeatedly in Meta posts (e.g. this post talks about the need for objective winning criteria). It needs to be complet...

discussion meta site-health

BlueEyedBeast

Now muddyfish again!

ErikGolferエリックゴルファー

Can you change your parent site to PPCG, please? You show up as "BlueEyedBeast" currently.

Alex A.

His chat profile's parent site is PPCG

ErikGolferエリックゴルファー

Yeah, but, he must "change" it to PPCG again in order to change the chat visible name.

BlueEyedBeast

20:42

Done

Nathan Merrill

@NewMetaPosts I feel like this is a really important question to ask. It's bothered me that my question of "where do we actually stand on programming languages" didn't actually change anything

Nathan Merrill

21:02

@ais523 I'd really recommend formatting/clearing up your post a bit. It was a tough read, IMO as it's basically a wall of text (and reads like a rant).

I agree with your post, though

ais523

it is a rant, in some respects

Nathan Merrill

a rant isn't a way to get change to occur :)

ais523

PPCG's lack of objective rules lead me to ragequitting chat a few days ago, then being upset for hours and irritable for days

the only reason I come here is because I like the puzzles and enjoy solving them

everything else about the site, I find terrible and even offensive in some respects

I would have joined the site years ago if not for the fact that I hate it so much

Fatalize

if you ragequit because of answers on PPCG I'd say the problem is not primarily PPCG

ais523

the problem is PPCG's infrastructure

I have to come here because this is where the challenges are, and where the competition is, and where I can interact with other users

Nathan Merrill

21:07

I'm just saying that you need to communicate your frustrations better

a "I hate this and that and I quit" post doesn't change much. We've actually had those before

ais523

I agree with you, but I'm not sure I can actually do anything about it

I'm not sure I am necessarily going to quit, but I thought explaining the reasons why I hate it might potentially lead to change

there's a specific post right now whose legality is in dispute, and which is on an answer-chaining question and thus is holding the whole thing up

El'endia Starman

@ais523 Do consider how much of that infrastructure we can actually change. PPCG is essentially a hackish solution to having code golf and other types of challenges on the Stack Exchange model, which was designed for something else entirely.

Nathan Merrill

right, you can totally do that, but add some bolding to important facts, or make a numbered lists, or something

ais523

there's no objective rule dealing with it

Nathan Merrill

@El'endiaStarman he's specifically talking about our site-specific rules which can be changed

ais523

21:10

@El'endiaStarman I know, it's Stack Exchange itself that I hate, not anything about the PPCG community who are doing a moderately good job of working around it

but the site-specific rules are a potential workaround that isn't currently in use

ErikGolferエリックゴルファー

That's a different point.

ais523

atm we're using SE's default method for determining what the rules are

which is only good if people aren't skirting the edges of the rules all the time

and the nature of a competition site, as opposed to a programming QA site, is that you have to skirt the edges

on Stack Overflow, if your question or answer is close to being invalid or rulebreaking, just edit it to bring it further from the edge

on PPCG, you can't do that because it probably costs bytes

ErikGolferエリックゴルファー

PPCG isn't a Q&A site. That might explain a lot about your situation.

Nathan Merrill

@ais523 one potential problem with nailing down all of the rules is that there are lots of rules that would really difficult to nail down unambiguously. For example, (in recent memory), we allow input/output via the stack. What exactly is the "stack"?

New Main Posts

0

Q: Sum digits till Square

Given is any integer x > 0 and any base y > 3. Sum all digits of x (if written in the set base). Multiply this by the highest possible digit (is always base -1). Repeat until this value is (y - 1) ^ 2 Searched is the count of iterations and the steps. Example 1: x= 739 y= 7 searched: (7 - 1...

code-golf combinatorics base-conversion

Fatalize

21:14

It is impossible to set objective and exhaustive rules given the number of different languages there are here

Nathan Merrill

I wouldn't say impossible, but we'd have to get really good at making statements that are broad enough to include languages but specific enough to ban loopholes. A good example is Peter's post about what a programming language is

ais523

@NathanMerrill IMO the best way to nail down input/output arguments objectively is to a) have a very standard sitewide standard for which input and output is given (e.g. stdin and stdout, arrays have such and such a formula, etc.); b) have an official list of wrappers that convert from that standard to things that are "more suitable" for the languages in question; c) require wrappers to be posted to Meta and get a certain number of upvotes to be usable

our current method is full of loopholes, but I can't know whether they're worth abusing or not because I can't know whether they comply with the rules or not

Fatalize

@NathanMerrill And then Dennis created this proving that Peter's definition of programming language is dubious at best

because it's actually too hard to define to not get any abuse

ais523

for a while, I believed Jelly (the language) is abusing a loophole in I/O methods, but the lack of rules meant that I couldn't prove it; I've since changed my mind and decided that it probably isn't, but again, with no rules it's hard to be sure

that said, I think it's possible to abuse the same loophole far more effectively

possibly even to the extent of having a valid zero-byte answer for every question

@NathanMerrill anyway, I've calmed down a bit; could you be more specific with what makes my rant hard to read (in particular, is it the answer, question, or both?), and I can potentially try to fix it

Martin Ender

@ais523 thanks for bringing this up. I also find our rule situation very unfortunate, but I don't really have any good ideas for how to solve it.

noɥʇʎPʎzɐɹC

21:24

I don't see a problem with our rule system.

ais523

I'm not sure I do either, but I thought it was worth at least starting the discussion and giving my best guess at a solution

Martin Ender

the only thing I can think of is making the rules a lot stricter, e.g. by allowing only full programs and stdin/stdout communication. I'd actually be in favour of this, but I know it would lead to a lot of backlash from the community. with anything like functions, you'll always get problems in terms of objectivity when it comes to language semantics. if nothing else, I can always design a new language that defies the concepts we've based our definitions on.

ais523

actually, the reason I think we have to allow functions and dubious communication methods is because we have competition between languages here

Martin Ender

@ais523 re your new meta post, I haven't read all of it yet, but we do have time of posting as the default tie breaker

ais523

and some languages need more boilerplate than others, and the boilerplate isn't that interesting

I believe the fix is actually to define new language variants with less boilerplate, though

Martin Ender

21:26

competition between languages is pointless with or without rules accommodating those languages more conveniently

Fatalize

I know ais523 has answered with Prolog, and in that language restricting I/O to STDIN/STDOUT makes no sense

and for other languages it's the natural way

ais523

yes, but in Prolog, restricting output to strings and lists and integers (and concepts that even exist in other languages) makes no sense too

even with PPCG's huge permissiveness about I/O

I often have to spend more bytes complying with I/O formats than with the actual algorithm :-)

Martin Ender

the other solution (and I've suggested this in the past) is to go the other way and get a lot more fine-grained by setting up rules for submissions per language

ais523

I don't mind that, though; the existence of golfing languages mean I don't have to worry about winning

ETHproductions

@Flp.Tkc Of course :-) Just beware that it's quite difficult to use for all but the simplest challenges.

Conor O'Brien

21:28

@ais523 tbh if you're golfing to win and not using a golfing language you're setting yourself up for disappointment

New Meta Posts

0

Q: Should the victory condition for a challenge always ensure that it has only a single winner?

Back when I first posted a challenge, it was closed as unclear, on the basis that – despite having an objective victory condition – the victory condition allowed ties. After searching Meta, I can't find a rule about this anywhere, so I think it's about time to start a discussion on what is and is...

discussion winning-criteria

noɥʇʎPʎzɐɹC

@ConorO'Brien not if it's python

Fatalize

@MartinEnder here

Conor O'Brien

@noɥʇʎPʎzɐɹC you're still setting yourself up for disappointment.

Martin Ender

@Fatalize wow, I would have had no idea where I had said that, nice work

Fatalize

21:29

Well I asked that so I remembered :p

ETHproductions

@Flp.Tkc And no, not yet, I generate them manually with JS's btoa. I'll add a button for that when I have time tonight

Conor O'Brien

@MartinEnder per language may be good, but perhaps too fine. Maybe there could be general categories and a process for appealing to a finer granularity on a per-language basis. (E.g., there could be a general golfing language category, or a general programming language category)

ais523

CMC: given two ranges of values on positive integers (e.g. 3<X<6, 4<X<8), find an integer that belongs to both ranges (e.g. in this case the answer would be 5), or return something clearly distinct from a valid answer if there's no answer available

Martin Ender

@ConorO'Brien I don't know. Jelly has very different semantics from Pyth, which is very different from Brachylog, which is again very different from CJam

Conor O'Brien

@MartinEnder true as that may be, it could allow for conformity among newer golfing languages

Martin Ender

21:31

re objective rules, one of our most successful and more recent rules is how we determine what's truthy and what's falsy. this definition doesn't actually work in a lot of esolangs, and I can't think of any good general objective rule that would cover all of them in a reasonable way.

undergroundmonorail

@TuxCopter i.imgur.com/AxTP4Uw.png

;)

ais523

(the above is designed to make a point about Prolog; it looks like a reasonable question, but in Prolog, you'd idiomatically answer the entire question in the input format rather than in the program itself)

H Walters

Golfing to win is tantamount to golfing for points... and the point structure doesn't seem to be optimized for golfing even for code-golf posts

TuxCopter

@undergroundmonorail q_q

You have a better score than me on my own challenge

Mitch Schwartz

it's always good to compare two answers that are solving different problems, and choose a winner

Conor O'Brien

21:32

@ais523 how are ranges inputted? pairs of numbers? 4 different, ordered numbers?

ais523

@ConorO'Brien that's the point I was trying to make with the question :-D

"any reasonable input method" is what I'd write on PPCG

Conor O'Brien

oh, cool

Fatalize

I don't see why that's a problem

ais523

the point is that as soon as you define how the ranges work, you're giving an unfair advantage to some languages and an unfair disadvantage to others

Conor O'Brien

so just leave it to any reasonable input method

ais523

21:33

OTOH, it's still an interesting (if fairly easy) problem algorithmically

Nathan Merrill

@MartinEnder I'd be ok with a bunch of rule sets for input/output. For example, lets say we had a "Rule set" that said "Truthy is 1" and "Falsy is 0". Then languages could "sign up" for that rule set. Basically, for each language, you'd have a list of rule sets that apply to it.

that way, you can reuse most rulesets for most languages

and you don't have to create new definitions every time a new language comes around

Conor O'Brien

@NathanMerrill +1

Fatalize

@NathanMerrill I basically argued that here but people don't agree

ais523

huh, I just remembered how booleans work in INTERCAL-72

there's a consensus that you use 1 and 2 for true and false; however, it's usual to change which is which dynamically in different parts of a program

because it's much easier to negate them when you generate them than when you actually use them

so which is truthy and which is falsey? my opinion here would be "you can do either but there's a penalty of 1 bit" (which is the amount of information you need to specify which is which"

then, if you submit two programs, one to handle one way round, one to handle the other way round, and they're the same length

then you can prove that the output format isn't being used to save bits, so you can remove the penalty

as it is, though, I think the fundamental problem in codegolfing is "this operation just happens to be shorter in this language" tends to determine the winner

Conor O'Brien

I don't think winning the challenge is the end-all to enjoying the challenge.

H Walters

21:39

Is it bad for swapped truthiness to count?

Dennis

@TimmyD There's now Powershell (with more aliases) and Powershell (coreutils) (reduced aliases). I've added the aliases according to this. Others might be missing.

ais523

@ConorO'Brien nor do I; if I did, I'd have left the site ages ago

OTOH, I find trying to define the rules an interesting problem in its own right

noɥʇʎPʎzɐɹC

CMC: Solve P=NP

Conor O'Brien

@ais523 right. but I'd go so far as to say that the winner is irrelevant

@noɥʇʎPʎzɐɹC very original, funny, and relevant joke.

ais523

I think the competition is relevant, and the winner is a method of defining the competition

Fatalize

21:40

@ais523 12 bytes

ais523

however, actually being the winener doesn't matter much

noɥʇʎPʎzɐɹC

@ConorO'Brien is that sarcastic?

Conor O'Brien

@noɥʇʎPʎzɐɹC Yes it is.

noɥʇʎPʎzɐɹC

@ConorO'Brien why ;_;

ais523

@Fatalize neat, and it makes a very good point about Prolog (and Brachylog) in the process

noɥʇʎPʎzɐɹC

21:41

It is original

Mitch Schwartz

because we are all friends, and that's what friends do

Nathan Merrill

@Fatalize its funny because I actually agree with both of you. A large problem with what you propose is that you propose that each language has a single "natural" input/output. That said, I actually think it'd make more sense for each language to have multiple options when it comes to what is more natural

Fatalize

Would be 6 bytes if the intervals weren't strict

Conor O'Brien

@noɥʇʎPʎzɐɹC not really

Nathan Merrill

Aka, in Java, if you are passed a list, you can be passed a List<T> or a T[]

noɥʇʎPʎzɐɹC

21:42

@ConorO'Brien I wasn't aware of that. You have accused me of reposting.

Fatalize

@NathanMerrill What annoys me the most is people claiming e.g. "lmao" and "rofl" are both valid truthy values because they evaluate to true in some language

Though I've noticed that challenges recently generally ask only one value for truthy and one for falsy to avoid that

Conor O'Brien

@noɥʇʎPʎzɐɹC accusing me of what you may.

ais523

oh gah, I just realised that Help, WarDoq! probably works for the polyglot challenge

Nathan Merrill

@Fatalize in some languages, I'd allow that

H Walters

@NathanMerrill Or (T,T,T,...T)

noɥʇʎPʎzɐɹC

21:43

@ais523 ?

ais523

@Fatalize almost everything is truthy in Perl; I often exploit that via printing a copy of the input, or something similar, for true

Fatalize

@NathanMerrill I would only allow one truthy value

Nathan Merrill

for example, in python, if you wanted to return "lmao" as a truthy value, I'm ok

ais523

@noɥʇʎPʎzɐɹC well we're collaboratively (answer-chaining) making a polyglot; Help, WarDoq! seems like a language that could genuinely be worked into it

Fatalize

people returning the result of a computation as a truthy value are not returning a truthy value

Nathan Merrill

21:45

its actually a really common practice to return something other than true/false even when its going to be used as a truthy/falsy value

that's one of the strong points of python

H Walters

In production C, 0's often used for truth and nonzero for false (most especially when dealing with methods that return error codes)

Nathan Merrill

I'm completely ok with that

Fatalize

I'm ok that it's common. I'm not ok that you can return a lot of different values that are supposed to all be "true"

ais523

@Fatalize well, given a challenge like "does a string of letters contain any letter other than x", I'd call s/x//g (+1 for -p) a valid solution in Perl

Fatalize

You choose one for your answer, not an infinite set of them

Mitch Schwartz

21:46

what's most important, though, is the emotion behind it all

ais523

if you change it to "a string of letters and digits" it no longer works

Conor O'Brien

as another example, if the object was to decide whether or not a char was in a string, (c,s)=>1+c.indexOf(s) is shorter than (c,s)=>c.indexOf(s)>-1

ais523

because 0 is falsey

noɥʇʎPʎzɐɹC

Your lang might be turing complete, but can your language perform every operation a RAM operation can perform in O(1) time in O(1) time?

bf isn't "RAM-speed-complete"

ais523

"A RAM operation"?

Nathan Merrill

21:47

@Fatalize I totally disagree. For example, lets say I had an iterator object with a "hasRemaining()" function. But instead of returning true/false, it returned the number of iterations left. I could totally see that existing in real life, so I'd be ok with it be valid on this site

in essence, its not the multiplicity of values that is the problem, but when users use the broadness of rules (that were intended for other languages) in their own language

Martin Ender

@ais523 finally managed to read through your other meta post. I like your suggestion, but I find the wording a bit too strict: there is an optimal answer to every code golf challenge, it's just usually very very hard to prove that you've found that optimal answer.

Nathan Merrill

aka, accepting a unary number

Fatalize

@NathanMerrill "Instead of returning true/false, it returned the number..." so you're saying yourself that you're not returning a truthy/falsy value

Martin Ender

and this is the main reason why code golf works so well as a winning criterion: in code bowling there's no upper bound. you could "improve" your answer endlessly, but there's also usually no challenge in doing so.

ais523

@MartinEnder I agree that that's a problem, and would welcome improvements; I was having difficulty wording it myself

noɥʇʎPʎzɐɹC

21:49

Random-access machine

In computer science, random-access machine (RAM) is an abstract machine in the general class of register machines. The RAM is very similar to the counter machine but with the added capability of 'indirect addressing' of its registers. Like the counter machine the RAM has its instructions in the finite-state portion of the machine (the so-called Harvard architecture). The RAM's equivalent of the universal Turing machine – with its program in the registers as well as its data – is called the random-access stored-program machine or RASP. It is an example of the so-called von Neumann architecture and...

Fatalize

which is not what a challenge asking truthy/falsy means

noɥʇʎPʎzɐɹC

ie being able to jump in memory

Nathan Merrill

@Fatalize right, but 0 is falsy, and everything higher is truthy.

Conor O'Brien

@Fatalize but numbers are truthy

Nathan Merrill

and that is very natural in python

Conor O'Brien

21:50

in some languages*

noɥʇʎPʎzɐɹC

@NathanMerrill italics are cool

Fatalize

True/false is two values

Conor O'Brien

truthy does not mean true, likewise for falsey

Fatalize

0/[1;inf[ is not

H Walters

@NathanMerrill How about find in C++ STL containers, which return iterators if the element isn't found (but nothing you can interpret unless you get the container's end iterator to compare to)

Nathan Merrill

21:51

you don't want to force python users to return literally True and False

I think that's a rather bad idea

Fatalize

You can return anything you want

But 2 values

not as many as you want

Choose whatever you want

Nathan Merrill

then you're going to end up with a worse situation where users will return 0/1;inf[ for true and 1/0;inf[ for false

ais523

what about Prolog? in early implementations, returning false and crashing were indistinguishable

undergroundmonorail

@TuxCopter i.imgur.com/rS4vtZ4.png we probably have exactly the same solution haha

Nathan Merrill

@HWalters doesn't C++ auto cast those iterators to true/false?

ais523

21:52

(since then, certain errors have been declared serious enough to terminate the program rather than merely failing)

H Walters

@NathanMerrill No

TuxCopter

@undergroundmonorail It's the obvious solution

undergroundmonorail

yeah

TuxCopter

And I think it's optimal

undergroundmonorail

for power usage, sure, but people have done ~Y5 and 4 lines

TuxCopter

21:53

wut

Nathan Merrill

@HWalters well, I don't know C++, but the point is that we would have a list of "acceptable" values for each language. As to what C++ is allowed, I have no idea

undergroundmonorail

@TuxCopter i.imgur.com/7REkyvI.jpg check it out

TuxCopter

Wow

H Walters

C++ has truthy values for sure; just raising the point that there's a potential "cheat" if you don't (in this case, can't) specify explicitly which value is falsey

Nathan Merrill

oh I agree. I was specifically talking about Python for my example

ais523

21:55

incidentally, it's not just codegolf where this sort of situation comes up

noɥʇʎPʎzɐɹC

@everybody What if it is a fastest-code challenge, and I create an object that evaluates to either true or false when evaluated, e.g. using @property in Python?

Nathan Merrill

in Java, returning 0/1 instead of false/true would not be allowed

noɥʇʎPʎzɐɹC

Then I can do it in 0ms!

ais523

mathematicians have been stymied for ages trying to figure out rules for what counts as Turing complete

(the loophole in question being the use of an infinitely long program in languages where all programs are inherently infinitely long)

@noɥʇʎPʎzɐɹC I believe the length of time it takes to determine truthiness should count as a penalty on the answer

arguably the compile time should too, otherwise you could construct a lookup table of all possible answers at compile time

noɥʇʎPʎzɐɹC

@ais523 C++ is instantly disqualified.

H Walters

21:57

(Also in C++, string has a find that returns std::string::npos if not found, which is roughly largest possible signed value cast as unsigned)

ais523

write a faster compiler :-D

Nathan Merrill

@ais523 just hardcode your lookup table :)

noɥʇʎPʎzɐɹC

^

I have never allocated memory in C++, and yet I have written not-small programs. Am I C++ing wrong?

Nathan Merrill

there's a reason why good fastest-code challenges near-impossible to make a lookup table for

I believe for C++ that's fine. But not ok with C

H Walters

@noɥʇʎPʎzɐɹC You never use strings?

noɥʇʎPʎzɐɹC

21:59

@HWalters Not really.

ais523

@noɥʇʎPʎzɐɹC It's very hard to avoid allocating memory; creating any local variable does that. However, in modern C++ it's entirely possible and reasonable not to allocate any variable explicitly, and rely on libraries and language functionality to do it for you.

noɥʇʎPʎzɐɹC

I mean I do string buf; for simple stuff.

ais523

even int i; allocates memory, if it appears inside a function

Transcript for