Wolfram Mathematica

« first day (2961 days earlier) ← previous day next day → last day (1520 days later) »

Chris K

12:12 AM

@JimB One day I should learn stats

3 hours later…

user76284

3:28 AM

Does anyone know why Integrate[Boole[PrimeQ[Floor[t]]], {t, 0, 20}] yields 0?

b3m2a1

3:44 AM

@user76284 Not sure but it might just not be sophisticated enough when it comes to such a weird integral. On the other hand even I can integrate that one:

weirdInt[n_] :=
  Total@Differences@
    Append[Select[Prime@Range[Ceiling[n]], # < n &], n];
weirdInt[20]

18

Not sure if you can always do the Append but it worked for your case

Ah this might work as hack:

weirdInt[n_] :=
  Total@Differences@
    If[Boole[PrimeQ[Floor[n - .01]]] > 0,
      Append[n],
      Identity
      ]@Select[Prime@Range[Ceiling[n]], # < n &];
weirdInt[20]

4 hours later…

Carl Lange

7:37 AM

@b3m2a1 Sadly I haven't seen anything about it in the livestreams or the prerelease. But perhaps if the video export is so much improved, I don't care about gifs...

C. E.

7:54 AM

@b3m2a1 What is the alternative to rasterizing and then stitching?

b3m2a1

@C.E. there is no real alternative, it's just super slow in the FE since the FE rasterization is so brutally slow. Sometimes you get better performance by pre-rasterizing, since then the FE doesn't pick some weird rasterization options, but the stitching is painstakingly slow

1 hour later…

Carl Lange

8:57 AM

@b3m2a1 @C.E. Yes, the stitching itself is extremely slow. I normally generate my frames ahead of time with a ParallelTable or something, but Export["my.gif", frames] still takes forever (when compared with gifsicle or imagemagick or whatever).

And the output gif is normally 50MB+, when compared with gifsicle which creates a 45KB gif instead with no obvious loss in quality.

C. E.

@CarlLange @b3m2a1 yup, I agree about the slowness and the size issues. It would be great if they could speed it up somehow.

6 hours later…

Kuba

3:09 PM

@C.E. @CarlLange @b3m2a1 would appreciate if you could judge this, I guess it is plain wrong but it is hard to tell for me :) mathematica.stackexchange.com/q/215357/5478

C. E.

3:23 PM

@Kuba I'm not familiar with FormFunction etc. but I if you want to send things in the body then you should make sure that it's sending a POST request.

kirma

Certainly so, because GET requests don't have bodies at all, by definition.

3 hours later…

Rolf Mertig

6:50 PM

@kuba: you could try using the experimental $CookieStore = Automatic

CA Trevillian

7:10 PM

What's the difference between {{x1,y1,z1},{x2,y2,z2},...} and {pt1,pt2,...} when it comes to Tube? It does not seem to have any rhyme or reason. I can wrap my list of point coordinates in Line and it works fine. BUT when I try to use the same list in Tube it does this:

Tube is on the left & Line is on the right. What gives? I can ask a question with specific data, also, if this is recommended. Just seems like a simple misunderstanding of how Tube works. I thought I figured out the problem, but I really did not.

kirma

@CATrevillian Frankly I have trouble understanding your question. Can you provide a small example that misbehaves?

CA Trevillian

@kirma honestly the only way I find it breaks is with this data set that, for whatever reason, is enjoyed by Line, but detested by Tube. Would posting the data set here be productive? By productive I mean its a 100 length list of machine number triples & I don't know if it will post here due to length

kirma

Hmmh. I wonder if there's a misbehaving point in the set.

CA Trevillian

That would be ignored by Line but treated vastly different in Tube?

kirma

@CATrevillian Just speculating. If I were you, I'd try to trim the dataset from beginning and end and see if at some point the Tube version starts behaving better.

CA Trevillian

7:28 PM

Good idea...ah Nope, whole thing makes a big-old curled sheet. ;(

It's not taking into account the z-axis part of the data...hmm

kirma

Line and Tube shouldn't really behave too differently. I suggest asking a question, if you can make test data/input compact enough.

CA Trevillian

I'll do this, and maybe just take an internal cut of the data, since, from my fiddling per your recommendation, it doesn't matter what part of the data is used! That way I don't have a huge code block of boring numbers haha!

CA Trevillian

7:51 PM

Maybe someone has some good input, or can tell me what I have done wrong:

Q: Why does Tube treat this list of machine number triples differently than Line?

I have a set of data that Line enjoys quite well, however, Tube seems to throw something of a tantrum when I deliver the same set of data. The data: data = {{-1214.2199935658107`,-848.2705321806565`,-0.00001416666666666664`},{-1244.2463057872262`,-804.5024679750907`,-0.000013333333333333296`},{-...

2 hours later…

CA Trevillian

9:38 PM

Thank you, @kirma

Carl Lange

9:48 PM

@Kuba Does it actually work to keep persistent data in a variable like that between diffeent FormFunction calls? I get the sense that data will need to be kept elsewhere (a CloudObject perhaps) and written to & retrieved after submit

But I might be misunderstanding your intent

Is there some problem with this answer that I'm not seeing? I'm unsure of why it has gotten such a poor reception when the question has been viewed plenty of times. there are no other answers and it seems to me like it answers the (very sparse) question. https://mathematica.stackexchange.com/questions/215333/how-to-test-upload-speed/215342?noredirect=1#comment550760_215342
I'm genuinely asking - it feels as though I've missed something really stupid and nobody wants to tell me :)

(I really don't want to come across like I'm fishing for votes - please don't take it that way!)