Wolfram Mathematica

« first day (1464 days earlier) ← previous day next day → last day (3317 days later) »

kirma

04:54

@SquareOne I think you have become overly sensitive to the concept for a while. There's a name to that concept... which I don't remember now.

Above pattern is not exactly a nephroid though, although it's very close to one.

Nasser

05:13

I was wondering if anyone here checked Julia? I just found the most confusing thing about it. When A is a vector or matrix, and one types B=A, then when A changes later, then B also gets changed! So one has to type B=copy(A) to actually make a copy, not a reference. I can see this easily leading to many bugs, unless one types copy() each time. Which makes the code more ugly to read. I have no idea why they designed it this way.

3 hours later…

SquareOne

08:37

When I said "nephroid" I actually meant [Ranunculoid](http://mathworld.wolfram.com/Ranunculoid.html) of course ;). Check this `Manipulate[Show[{Graphics[{
ParametricPlot[
scale*{6 Cos[t + s] - Cos[6 t + s],
6 Sin[t + s] - Sin[6 t + s]}, {t, 0, 2 \[Pi]},
PlotStyle -> Red, DisplayFunction -> Identity][[1]]
}], g1}, AspectRatio -> Automatic, ImageSize -> 200,
Ticks -> None], {{scale, 0.14}, 0.0001, 1}, {{s, 0.9424}, 0, Pi}]` where `g1`is your graphic.

2 hours later…

Szabolcs

10:58

@Nasser I think you asked about this before 1-2 years ago, but the question was maybe about Python ... (which behaves the same way).

Jul 1 '14 at 23:01, by Nasser

I just un-installed Julia from my Linux after I saw this:
julia> A=[1,2];
julia> B=[3,4];
julia> A=B;
julia> B[1]=7;

now guess what A is?

julia> A
7
4

So changing B has indirectly changed A. It is not copy by value as any normal language will do. This is so confusing.

4 hours later…

kirma

14:38

I have a large set of sets of items. Every set has a probability of getting sampled. How to efficiently generate random samples of N sets conditioned so that every set in them has a non-empty intersection with the union of the rest of the sets in the sample?

I mean, of course, that these samples would be unbiased in comparison to actually drawing N random sets, and then discarding the result if the condition is not fulfilled.

halirutan

15:35

@kirma I did something similar long time ago and new I couldn't resist to play once more with it.

kirma

halirutan

Obviously imgur doesn't like so much

imgur.com/gallery/rfHzteR

kirma

I also considered animating it, but the interesting nuances, like those lines very close to an integer value wouldn't have really appeared on animation...

halirutan

@kirma Above, I tricked it because I only animated fractions and not real values. Another apporach is to to make the animation of the parameter not moving linearly in time. You can make it stop at some fractions. I even created an HD video of it. If you like you can watch it on YouTube.

kirma

I also considered such "cheats" :)

kirma

16:02

I guess my subset problem above is a variation of random spanning (sub)tree of a graph...

1 hour later…

Szabolcs

17:06

Hi everyone!

Do we have an equivalent of ImageResize that works on arrays (matrices) rather than images? ListInterpolation comes close but it's still not direct. Anything better?

Also it's pretty slow to construct the new matrix from ListInterpolation ...