The Nineteenth Byte

zyabin101

12:00 PM

@MartinBüttner How much it took to compute?

Martin Büttner

a few seconds

there's only about 65k states

rendering takes longest I think

Here's the code:

Graph[With[{state = Partition[IntegerDigits[#, 2, 16], 4]},
    UndirectedEdge[#,
     FromDigits[
      Flatten@CellularAutomaton[{224, {2, {{2, 2, 2}, {2, 1, 2}, {2, 2, 2}}}, {1, 1}}, state, {{{1}}}], 2]]
    ] & /@ Range[0, 2^16 - 1], ImageSize -> 2000]

(where you'd use DirectedEdge if you actually wanted to work with the graph)

zyabin101

This one POV-Ray scene was much easier.

gist.github.com/schas002/f17a27b8e8cf81238ae17a8f60f70df7

flawr

12:34 PM

2

Sp3000

12:48 PM

@El'endiaStarman Variations of Life feature request: load grid by copying and pasting a grid into somewhere? (e.g. if I generate a grid with a program and want to copy it over)

Oh nvm, I guess that's what the permalink's for

Mars Ultor

1:26 PM

@Sp3000 You could try to convince him to shorten it though, it's usually far too long to fit in one chat message

zyabin101

@MarsUltor You're back!

Did you read the UGL room?

Nathan Merrill

is there anybody who knows Haskell in the room?

zyabin101

@NathanMerrill Erm, nope. No one... No.

trichoplax

@zyabin101 Do you have a keyboard shortcut for that exact phrase?

zyabin101

@trichoplax Nope.

trichoplax

1:40 PM

chat.stackexchange.com/…

:P

Sp3000

Not sure I can speak on behalf of the room, but I think it'd be more correct to say that the people best equipped at answering Haskell questions aren't currently in the room, although the rest of us can most certainly try

Nathan Merrill

I'm still not sure how gradual typing should work with function overloading.

maybe I simply remove function overloading

for example, in many languages with gradual typing, you could type:

def func(a, b)
    return a + b

and it'll only accept types where a and b can be added

alternatively, you can do stuff like

def func(a, b: NonZero)
    return a/b

which ensures that b isn't ever zero

that said, if I have two functions that accept similar types, is that a problem?

def func(a: Int, b: IntGreaterThan3)
    #magic math

def func(a: Int, b: Int)
     #other math

I see 3 ways to resolve the overload:

1. Resolve it based on the type of the variable at compile time. Benefits: fast

2. Resolve it based on the type of the variable at run time. Benefits: flexible

3. Disallow it

Sp3000

They're similar but still different types, right? Or can they be used interchangeably?

As in, can something be both an Int and an IntGreaterThan3?

Nathan Merrill

Yes. IntGreaterThan3 is actually more like Int<GreaterThan<3>>

I could even do something like Int<GreaterThan<3>, LessThan<7>>

all of the parameters in the type are restrictions on the type

Mars Ultor

@NathanMerrill Not a good idea IMO

I like C#-style contracts more

Nathan Merrill

1:52 PM

@MarsUltor that's gradual typing.

Mars Ultor

Not entirely sure writing IntGreaterThan3 is a good idea

Nathan Merrill

A really good example is for the division example

Mars Ultor

I think it's better to have assert(b > 3)

Sp3000

I feel like if it's Haskell you could just let it pattern match to the first one that works at runtime, but I don't know if that's what Haskell really does

Nathan Merrill

def divide(a: Int, b: Int<NonZero>)
    return a/b

@MarsUltor assertions are at runtime. The benefit of gradual typing is that its at compiletime

@Sp3000 regardless of what Haskell does, what would you do?

or, expect?

Mars Ultor

1:58 PM

@NathanMerrill How do you assert b > 3 at compile time?

Nathan Merrill

whoever calls it have to ensure that its greater than 3

so, for example, if its user input

if they do a if (a > 3)

then that will work

or, if they were passed a number that was greater than 0, and add the constant 3 to it, that'll work too

or, even better, if they a number that was greater than 0, and a number that was greater than 3

Sp3000

If it were my own language, I'd probably disallow it just for the sake of code clarity? Maybe

But if it helps:

46

Q: Overloaded method selection based on the parameter's real type

I'm experimenting with this code: interface Callee { public void foo(Object o); public void foo(String s); public void foo(Integer i); } class CalleeImpl implements Callee public void foo(Object o) { logger.debug("foo(Object o)"); } public void foo(String s) { ...

There's Java, apparently

Nathan Merrill

yeah, Java does option 1

@Sp3000 would you disallow overloading with incompatible assertions?

def (Int<GreaterThan<3>>)
    #Stuff
def(Int<LessThan<3>>)
    #Stuff

(in the above example, calling it with just an Int would be disallowed if static method dispatch is chosen)

Sp3000

Hmm that'd probably require a way to check whether they were incompatible, and I'm not sure I can think of an easy way of doing that

Nathan Merrill

yeah

Sp3000

2:11 PM

e.g. if you had Int<PerfectNumber> and Int<OddNumber>

Nathan Merrill

actually, there's an easy way to do this

static method dispatch

and if multiple parameters match a call, then you can't call it?

no, that doesn't make sense

if multiple methods match, you have to explicitly cast it

quartata

By the way, C can also do this with _Generic in C11:

Nathan Merrill

I think that makes sense. It means that if I have an integer, and I add 3 to it, I'm not surprised when a different method is called

@quartata you can put values in a type?

aka, String<Length<5>>?

quartata

void foo(void* point) {
_Generic(long: something, double: something else, *point)
}

no I mean call something else based on real type

Nathan Merrill

so does _Generic store nulls in every other value?

er, nulls don't exist

but it uses up the memory for all of the types, I'm assuming

quartata

2:20 PM

The somethings are code snippets. I think it does nothing for a missing type

Not much of a C expert though.

zyabin101

How can I search eye macarenas like >_>?

Chat's search does not support the raw >_>.

flawr

Yes the search is quite strange.

E.g. it would not find my previous message if I would look for "strang"

Nathan Merrill

2:41 PM

I'm writing a paper, and I keep on using markup to do text emphasis

-.-

flawr

.

._.

=)

@NathanMerrill Is there no markdown -> latex converter?

Nathan Merrill

not using latex

zyabin101

Nathan Merrill

its not that kind of paper

zyabin101

I do POV-Ray, a third time!

https://gist.github.com/schas002/4718c764985953675f6258b24dc84b91

Nathan Merrill

2:44 PM

just using google docs

New Main Posts

0

Q: About a 16x16 (or 32x32, 64x64, etc.) Franklin Square that can´t be more magic

I made with pen, paper and a calculator the most magic 16x16 (and from there a 32x32, 64x64 etc.) Franklin square. All the horizontals and verticals add up to 2056, all the half ones (of wich there are horizontally as well as vertically 5 ones) add up to 1028, all the fourth ones (of wich there a...

puzzle-solver

Sherlock9

3:22 PM

@El'endiaStarman Hey El, I've got some blasphemy for you: eev.ee/blog/2016/01/04/…

Eh, that was poorly worded. Sorry, I just thought you might this article interesting, since you like Paul Graham, and this guy... well, I was wondering what your reaction to an opposing viewpoint might be

New Main Posts

0

Q: Form tiles in a square ring

Given the input tilesX and tilesY create a method that would make a square ring from the tiles. The function must order the tiles in a ring like this: tilesX and tilesY are always positive integer values, cannot be 0 tilesX and tilesY are not always equal, expect values like: 2x5, 6x3 etc. C...

Transcript for