The Nineteenth Byte

Jan Dvorak

17:00

@ETHproductions Object.defineProperty(foo, "bar", {set:function(x){this.bar = x; Object.defineProperty(x, "goat", {set:...})})

Downgoat

function(){var a = 1;} what object is A on?

Step Hen

@ETHproductions not quite everything, 1.foo = "bar"; is invalid :P

ETHproductions

@JanDvorak But what if you don't know that someone is going to set the goat property?

Downgoat

^ also you'd have to move goat setter inside other setter

Jan Dvorak

So? Then the setter won't be called.

Step Hen

17:01

@JanDvorak The original question was how to tell if an object was changed at all, I believe.

musicman523

javascript

Downgoat

@StepHen well that is syntax error but 1..foo is accessible

Jan Dvorak

Well, that's not as easy. But you can enumerate properties and attach setters to all of them.

Downgoat

You can't enumerate properties that don't exit

Jan Dvorak

Yeah, you'll miss those. Was that a requirement?

Downgoat

17:02

Yeah because empty object doesn't have any properties

ETHproductions

We want to know if anything changes on an object. For example, write a setter that detects when this code is run, without using <expression>: foo[<expression>] = 1;

Jan Dvorak

Setters won't do that job, then

Downgoat

What you could do is attach a global detector to Objects proto

Jan Dvorak

You could attach a proxy in your setter

CalculatorFeline

If you can run a function on yourself frequently enough, you can check for changes there.

Jan Dvorak

17:05

@CalculatorFeline bad idea

CalculatorFeline

True.

ETHproductions

	for (var index = 0; index < arr.length; index++)
		if (index === arr.indexOf(arr[index]))
			perm(arr.slice(0, index).concat(arr.slice(index + 1)), len - 1).map(function(b) {
				result.push([arr[index]].concat(b));
			});

@StepHen Happy? :P

At least now I don't have to do +index every time

Downgoat

If I make a functional operator TC proposal does anyone else want to help make?

Step Hen

@ETHproductions Yes :P

ETHproductions

@Downgoat I'm not sure what this functional operator you're talking about is

Downgoat

17:07

@ETHproductions basically (+) == (a,b)=>a+b

ETHproductions

Hmm

Downgoat

Cc @ConorO'Brien

ETHproductions

Would (~) be invalid then?

Jan Dvorak

Eh, it's not as groundbreaking if you're not ... wait, this is TNB. Go ahead.

Downgoat

Why would it? That would be unary functional op

ETHproductions

17:08

But then how do you get unary (-)? :P

Jan Dvorak

(-) might be both unary and binary based on how you call it

good question though

Downgoat

Either 1) call 1 arg does unary and with two does binary 2) (:-)

Cheddar does borh

ETHproductions

> borh

Jan Dvorak

What does 1-arg call do with (*) ?

Downgoat

* is not unary arg

Jan Dvorak

17:10

Hence the question

Downgoat

Btw ... is not an operator in JS but often described as one

ETHproductions

Error I'd say

Because a* is a SyntaxError

Jan Dvorak

One issue is that (?:)(a, b, c) can't be equivalent to a ? b : c

similarly with (&&) and (||)

ETHproductions

I wouldn't allow (?:) if I were in charge

The other two though, idk

Jan Dvorak

Why not? The desired behavior is clear.

Step Hen

17:11

I still don't understand where this'll be used

Downgoat

^^ we can still have but not short circuit

Jan Dvorak

not short circuit = not identical

Leaky Nun

tfw you look at VSauce's old videos and realize that indeed everyone has a dark history

Downgoat

Since JS is becoming example of functional programming would be great for functional program

ETHproductions

@StepHen sum = array => array.reduce((+), 0)

for 1 example

Jan Dvorak

17:13

@Downgoat Syntax error: failed to parse the grammar structure of this sentence

New Main Posts

3

Q: Fibtraction - Fibonacci but with subtraction

The Fibtraction sequence (as I call it) is similar to the Fibonacci sequence except, instead of adding numbers, you subtract them. The first few numbers of this challenge are: 1, 2, -1, 3, -4, 7, -11, 18, -29, 47, -76, 123, -199, 322, -521, 843, -1364... The sequence starts with 1 and 2. Ever...

totallyhuman

hi

Phoenix

o/

totallyhuman

My controversial sandbox post is doing well on main :D

Step Hen

@totallyhuman good :)

Jan Dvorak

17:22

The sandbox is fundamentally broken :P

Step Hen

@JanDvorak We need a meta-Sandbox to fix Sandbox

Jan Dvorak

There already is one, and I'm not sure it helped

ETHproductions

>

á: function(x){var p=[],u=[],f=function(z){...};var l;return f(this).map(function(z){return z.slice(0,x||z.length)}).â()},

Note the extraneous var l;

Old me was incredibly dumb :P

Step Hen

@ETHproductions var p=[],u=[] => var p = u = []

totallyhuman

Welp first downvote

Downgoat

17:29

@NewMainPosts as ETH said is near dupe of lucas but with - instead of +

Any reason why not dupe?

ETHproductions

@StepHen That actually is very wrong, as any changes to p will then affect u

Step Hen

@Downgoat We don't have a Lucas sequence post, as far as I know?

@ETHproductions oh right, yeah, don't listen to me

@Downgoat Oh wait we do, but it was closed as a dupe of something else

Downgoat

Yeah :P

Step Hen

But I don't think the current post is a dupe of codegolf.stackexchange.com/questions/64476/… though

totallyhuman

;-; ok

Guess that's that

ETHproductions

17:34

Wait, I thought someone discussed this in the Sandbox and came to the conclusion that it's different enough

Downgoat

From totallyhumans quote he is saying old one sucks but is still dupe

totallyhuman

It was controversial

ETHproductions

ah ok

Downgoat

we can discuss here about improving the Fibonacci challenge though

totallyhuman

Not gonna bring any activity to it though

Downgoat

17:35

The challenge being old also does not stop it from being dupe

ETHproductions

For every operator returning a boolean (e.g. <, >, in, instanceof) JS should have a version with a preceding ! that returns the opposite

So that we can change e.g. if (!(x instanceof Array)) to if (x !instanceof Array)

Downgoat

yes pls I hate doing if(!(a instance of T))

Does anyone else here on es discuss?

trichoplax

@ETHproductions I like the idea of !> and !< instead of <= and >=, and the others all sound more readable that way too. Suggest it to W3C

Downgoat

*TC39

trichoplax

Is that the new standards committee?

Downgoat

17:44

They do ECMAScript, JS as a language W3C only does the browser specific stuff

ETHproductions

OK so how do I propose something like this?

trichoplax

@Downgoat Ah I see. Thanks

Downgoat

Though TC39 does have representatives from most browsers

@ETHproductions first make a markdown file describing spec

trichoplax

I guess the other way is to talk one or more browsers into implementing it to gather support

Downgoat

then you'll need a champion to present it to TC39 comittee

ETHproductions

17:46

...I was afraid it wasn't any simpler than that -_-

trichoplax

I use !> and !< when writing by hand, because it feels more intuitive, and then when typing up as code I change to <= and >=

Downgoat

^ I always have to spend like a minute thinking what weird side effect will happen when switch

Might also be because I am dumb

trichoplax

Maybe implement it in other languages so people get used to it and it can spread out into more and more languages until JS naturally follows

@Downgoat Yes it's definitely a source of delay and error. I'd love to be able to just leave it as !<

The only reason I put up with the risk is that the ease of working with it on paper seems to justify it

Random question: Roughly what proportion of time working on a JS project would you expect to be "getting it to work on IE"?

ETHproductions

Question: NaN <= 0 is false, but NaN > 0 is also false. Should NaN !> 0 be false (as in NaN <= 0), or true (as in !(NaN > 0))?

Leaky Nun

@ETHproductions what the hell

ETHproductions

17:54

Hey, JS isn't my fault :P

trichoplax

NaN === 0 returns false and NaN !== 0 returns true. I'd guess anything that's not a number isn't going to be equivalent for !< and >=, whether it's a string or NaN or whatever.

Step Hen

term !> term should be syntactic sugar for !(term > term)

ETHproductions

^ that's kinda what I thought

But I think JS just has all comparisons with NaN return false, so I don't want to break that

trichoplax

Isn't the NaN weirdness part of the floating point standard? One thing that can't be blamed on JS...

@ETHproductions !== returns true, so there's no hope for consistency

Step Hen

@trichoplax Yup

ETHproductions

17:58

oh... darn

Step Hen

@trichoplax NaN does not equal any number, so NaN !== num returns true

trichoplax

I agree the !(a>b) seems the most intuitive meaning for a!>b

Step Hen

so what about !<= and !>= XD

Lembik

hi all

trichoplax

@StepHen That was going to be exactly my argument for !> returning false, until I tested in a browser and found NaN !== 0 returns true...

Lembik

18:00

if you haven't seen it, my meta question is up re: time limited code golf questions

waiting for your opinions :)

Step Hen

@trichoplax Mistyped sorry

trichoplax

Excellent - off to have a look...

Lembik

but please do read the whole question as I cover a number of points

7

Q: Do we need a new tag for time limited code-golf?

I like to ask code-golf questions but I typically don't want to see brute force solutions. The challenge I like to pose is to golf any reasonably fast solution for a (possibly hard) problem that will actually terminate in a reasonable amount of time. Examples of such questions (not all by me) in...

discussion

ETHproductions

@StepHen NaN !== x returns true for all x, haha

Leaky Nun

> you have more in common with a stranger today than you do with yourself ten or twenty years ago

Step Hen

18:03

@ETHproductions That's what I'm saying

Lembik

I am thinking it will really annoy people if I ask for fastest-code that you can tweet... right?

Step Hen

@Lembik Depends on tags

Lembik

@StepHen ok so let me be more concrete and maybe you can help

Step Hen

@Lembik If it's tagged code-challenge fastest-code dunno why anyone can complain

trichoplax

That seems a valid challenge, but as you mention that's the other way round from asking for shortest code that meets the time requirement

Lembik

18:07

I would like to ask for the fastest integer factorization code that fits in 140 characters

@trichoplax it is indeed

Step Hen

@Lembik Don't see anything wrong with that

Lembik

@StepHen cool!

trichoplax

I don't see anything wrong with either

Lembik

people may complain it excludes java :)

Step Hen

The problem is when the winning criterion is not the most important one

Lembik

18:08

well you will need to do a lot of golfing

Step Hen

When you're saying "the fastest code that is at least this short" that's fine

Lembik

so you are again having to achieve two things

trichoplax

Upper limits on size or running time will exclude some languages, but if we only have challenges that include all possible languages then we'll miss out on lots of interesting ones

2

Lembik

it's exactly the same as the smallest code that runs in less than one minute

which upset a number of people

well the same idea, just the other way round

Step Hen

@Lembik No it's not, it's flipped on its head

Lembik

18:09

@StepHen right.. the same principle

trichoplax

All code golf challenges already have at least 2 objectives: the spec and the code length

Lembik

optimise X subject to a limit on Y

@trichoplax I agree completely!

but people got all upset and I had a number of downvotes

ok here goes :)

Step Hen

You're not Sandboxing it? :P

well chat probably pays more attention than Sandbox anyway

trichoplax

In between the extremes there is "your code must be tested", which doesn't specify a time but you can't post until the program terminates. Some would require that all code is tested anyway...

Lembik

@StepHen I will sandbox so we can discuss it :)

@trichoplax I tried that .. people also got upset!

oh the sandbox isn't linked any more

from the main page

trichoplax

18:12

I think whatever you do someone will complain. I'm interested to see how meta responds, to get a better idea of how many people support each side

Lembik

@trichoplax thanks. Me too

no answers yet

ETHproductions

> Feature requests for future versions of ECMAScript should not be made in this repository. To make a feature request, post to the es-discuss mailing list. Your goal will be to convince others that your proposal is a useful addition to the language and recruit TC39 members to help turn your request into a proposal and shepherd it into the language.

@Downgoat

Lembik

err...how do I find the sandbox?

Step Hen

338

Q: Sandbox for Proposed Challenges

What is the Sandbox? This "Sandbox" is a place where Programming Puzzles & Code Golf (PPCG) users can get feedback on prospective challenges they wish to post to the main page. This is useful because writing a clear and fully specified challenge on the first try can be difficult. There is a much...

discussion sandbox

Riker

@Downgoat absolutely not, sorry

Lembik

18:13

thanks!

ETHproductions

@Riker I totally read that the wrong way at first

trichoplax

I've flagged it for a mod to add the featured tag back on

Step Hen

@trichoplax Why does Community do stuff like that? 0.o

trichoplax

Because there are more cases where it's useful to remove the featured flag than cases where it isn't :)

We're abusing the featured feature because we don't have a mechanism for permanent notices (yet)

Step Hen

@trichoplax Oh, that makes sense

trichoplax

18:16

74

Q: We're not a Q&A site. But what should be done about it?

It has come up several times recently (more than usual) that PPCG differs from most of the other Stack Exchange sites in that it's not a Q&A site. People don't come here to ask a question because they have a problem, people come here to solve recreational challenges. The most valuable contributio...

discussion site-health

dzaima

I'd say that the featured tag shouldn't be removed by a time period, but when the question stops gaining attention. Just randomly removing the tag helps nobody

Downgoat

@ETHproductions imo false

Riker

@Downgoat I'm still learning julia, are you trying to learn it also?

Downgoat

Yeah I'm trying to do some 3D ANN stuff with GPU and C++ is too hard so trying to use Julia

I don't actually know if Julia has CUDA bindings though

Downgoat

18:29

@Riker huh for some reason the notification for this does not have comma

Riker

@Downgoat ye I edited

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Tweetable integer factorizer The task is as follows: Write code that factorizes integers as quickly as possible subject to your code being no more than 140 characters long. Your code can take input and give output in any way that is convenient including for example as arguments to a function. ...

Lembik

comments please!

in particular, how precisely do you measure the length of a tweet? Is is characters or bytes for example

Downgoat

@Lembik UTF8 bytes

Lembik

@Downgoat So I would say 140 UTF8 characters?

Step Hen

18:37

@Lembik Past challenges have done chars, or just done bytes like regular challenges

up to you

Downgoat

@Lembik well twitter doesn't have any specific encoding tweets use (the solid byte limit is 140) but online tweeter uses UTF8

Lembik

I may go for bytes... seems simpler

thanks @Downgoat

trichoplax

Use bytes if you want languages with different encodings to be included

Lembik

do I need more testcases?

@trichoplax bytes it is

Downgoat

I think test cases are good but need more spec on timing

is program startup time included?

trichoplax

18:42

I'm guessing no one is going to be able to hard code the existing test cases but they might be able to simply hard code what distinguishes them. More test cases might be easier than trying to objectively rule out hardcoding.

Downgoat

are you going to do like time ,/myprog

@trichoplax hardcodjng in this case is loophole abuse anyway I think and IIUC answers should still work for other numbers

Lembik

@Downgoat aha.. I was literally going to do time ./myprog

@Downgoat can you suggest some wording for this?

Downgoat

I think you should allow custom timing scripts e.g. let start=nanotime(); factorize(283844288373); print(nanotjme() - start)

trichoplax

@Downgoat Ah yes "should still work for other numbers" seems to cover it. Maybe with a specified range over which it should work?

Lembik

@Downgoat would you mind editing my question?

or @trichoplax too

Downgoat

18:46

Sorry in mobile rn so not so easy to edit

For like C, what is stopping me from using preprocessor to generate giant list of prime factors and doing lookup

Phoenix

BrainFlak CMC: swap contents of both stacks.

Lembik

@Downgoat If I am running time ./myprog wouldn't that take forever?

@Downgoat I am not sure how to allow custom timing scripts. They wouldn't be in the answer so how would they be provided?

Downgoat

I suppose the answer could say how to use special timing script and if they don't just use time

Lembik

@Downgoat that would need your help :)

it sounds complicated

currently I have "I will literally run time ./myprog on my Ubuntu system to do the timing so please also supply a complete program for me to run that includes any function you have defined."

Downgoat

Though with time I reccomend do like time [:] as a baseline because like even if I run empty program time says it takes like 0.013s on my computer

Lembik

18:52

0.013s is not a lot compared to 10 minutes

there aren't many test cases

Downgoat

89

A: Calculate the number of primes up to n

C, 0.026119s (Mar 12 2016) #include <math.h> #include <stdint.h> #include <stdio.h> #include <stdlib.h> #include <string.h> #include <time.h> #define cache_size 16384 #define Phi_prec_max (47 * a) #define bit(k) (1ULL << ((k) & 63)) #define word(k) sieve[(k) >> 6] #define sbit(k) ((word(k >> 1) >

That question says you can only hardcore up to 19 & answer must be deterministic. Maybe similar thing here?

Lembik

@Downgoat I am not sure what that would mean given the test cases

Downgoat

@Lembik latter part meaning you can't do isfactor(random()) and say it worked in 0.03s one time! Former meaning you might want to just say no hard coding any values

Lembik

@Downgoat but it does say that

trichoplax

Is the giant list of hardcoded primes a problem if it has to work for all numbers in range anyway? If "work" means return within 10 minutes

Lembik

18:56

@trichoplax I don't understand how you make this giant list quickly

oh you mean at compile time?

trichoplax

Downgoat's suggestion was to generate it at compile time

yes :)

Lembik

it would have to be too large to ever work

look at the size of the test cases!

it would never finish compiling

trichoplax

So you're happy for people to generate a smaller list of primes, provided it compiles on your machine in less than N minutes?

Lembik

sure

ah ok.. I see you want me to include the compile time.. ugh

this is a pain :)

trichoplax

I mention it for your benefit - if you don't specify then you're the one who is going to be compiling all these programs on your own machine and wishing you had specified :P

Lembik

18:58

how about languages that are compiled must take less than 30 seconds to compile?

@trichoplax thank you!

trichoplax

I don't suppose the time limit matters much more than the 140 characters limit. As long as it's specified there's no "best" limit

Transcript for