@Verbeia Will there be a main time? We are all from different countries. Maybe we can at least set a UTC starting time :-)

@halirutan thanks for explaining that ;) I still don't get why image processing thing generates something that looks like chemical reactions

@VitaliyKaurov I guess numerical errors are highly amplified since (a) we subtract every time the original signal and (b) everything is rescaled.

I tried the same by making a gradient filter with a Fourier transform with no success. The Fourier was too stable and showed only symmetric patterns.

@halirutan yeah but there is a lot of senselessly looking error accumulation like blowing up exponentially or similar. this is different - it is a pattern. info about Fourier is very interesting though

was rescaling included in Fourier too?

@VitaliyKaurov Yes, basically the same approach.

@halirutan i am really intrigued by that. how did Simon came up with that? and if I remember correctly someone was talking about Belouov Zhabotinski reactions. They are decribed by Ginzburg Landau complex field PDEs - which of course have spatial gradients, i wonder if this is they key

@VitaliyKaurov Simon said he found it by accident, but he wanted to make something with wave propagation.

@VitaliyKaurov What I believe is that the one important part are the boundary effects of the gradient filter.

@halirutan interesting, thats pretty cool - so simple. Because analytica PDEs that give this - ike Ginzbur-Landau are crazy looking - check eq. 3.2 here

chaos.utexas.edu/manuscripts/1067542721.pdf

I hope @SimonWoods gets time to explore this more

boundary effects are interesting idea. but i started it from random image and behavior i think manifested right away in the bulk

maybe i am wrong though

@VitaliyKaurov Try to use different Padding settings. You'll see that the behavior changes completely.

E.g. with padding periodic I get a stable pattern after some seconds.

@VitaliyKaurov I was intrigued by the patterns as well... it seems like any image will end up with something similar.

(see the previous two messages for the starting point and intermediate step)

@rm-rf Try Padding -> .2.

The constant source coming from the boundary prevents your spiral I guess

@rm-rf cool. did u see like two things being there and then one just dies? is there some sort of anihilation mechanism - like there is always odd number of spirals and clocl- and counter clockwise things kill eachother. this is a typical scenario in some systems

@halirutan i will try the padding. real interesting

@VitaliyKaurov Yup

@rm-rf And "Reversed" seems to end in this

generally guys, the craziest thing i think is that this is real-valued field. IN Ginzburg-Landau equations field is complex - so phase can lead to formation of vertices. Here i have no idea where they are coming from ;) I hope @SimonWoods reads this and explores more ;)

@halirutan do u have a snipet code for that?

@VitaliyKaurov It's still the same:

a = DiskMatrix[10, 200]; Dynamic[
 Image[a =
   Rescale[a -
     GradientFilter[a, 2, Method -> "Sobel", Padding -> "Reversed"]]]]

@rm-rf Yes, my 8 spirals have turned into 4 too.

I guess it is just stable a bit longer until it finally converges to one

@halirutan symmetry. start that from random - u get a singel thing

@VitaliyKaurov @rm-rf What I tried additionally was constantly sourcing the original with to some degree

img = ColorConvert[
   ImageResize[ExampleData[{"TestImage", "Lena"}], 200], "Grayscale"];
a = o = ImageData[img, "Real"]; Dynamic[
 Image[a =
   Rescale[.1 o -
     GradientFilter[a, 2, Method -> "Sobel", Padding -> "Reversed"]]]]

thre are other real field models other then G-L, like CAs:

demonstrations.wolfram.com/IdealizedBelousovZhabotinskyReaction

demonstrations.wolfram.com/GreenbergHastingsModel

@halirutan heh that's a bit funny looking ;)

@rm-rf how u got that image?

OK guys, I'm out for now. Pretty late here. See you all later.

@VitaliyKaurov That was with halirutan's suggestion to use 0.2 padding

yep @halirutan is right it boundary pumps purturbations, i wonder is there setting for periodic boundary conditions

that would kill the boundary

but i guess not

@halirutan It is a UTC starting time - midnight to midnight UTC 13 September. It's a whole day event (we have 1000 unanswered questions so we need the time!) so drop in whenever you want.

we got easy conversion ;)

wolframalpha.com/input/?i=UTC+12+am+in+CDT

sorry wrong link - here:

wolframalpha.com/input/?i=UTC+12+am+vs+CDT

@VitaliyKaurov Wrong link

wolframalpha.com/input/?i=UTC+12+am+vs+argentine+time

@belisarius are you there?

@VitaliyKaurov yep

@belisarius something went silly there. I reported it - sorry ;)

@VitaliyKaurov Almost everything is either silly or fabled here

@belisarius can't argue with that. Just posted Simon's boids in 3D: community.wolfram.com/groups/-/m/t/122701

@VitaliyKaurov Ha! A flock of dancers

nice

Row[{Column[{a, b}], Column[{a}]}, Alignment -> Top]

How do I get the second a at the top with the first?

Oops. Found: stackoverflow.com/questions/8377960/…

@MichaelHale so you are doing it with frame?

@MichaelHale Grid[{{Column[{a, b}], Column[{a}]}}, Alignment -> Top]

Ah, that too.

@MichaelHale Grid[{{a, a}, {b, SpanFromAbove}}]

In the actual code I don't know how many elements will be in each column, so the first two are better.

Now to make a Slider with two draggable tabs on it.

@MichaelHale Take a look at this: Experimental`IntervalSlider[Dynamic@x, {0, 1}] But I think you can't adapt it for your purposes unless it fits them already :)

That is pretty cool. I just started with:

It works fine if I click on parts of the bar, but I can only drag them a little bit before I lose the grab because they redraw.

What is splits

Just the pair of numbers for the position of the two tabs.

@MichaelHale You want those tabs independent?

I think IntervalSlider would be perfect if I could make it bigger. ImageSize doesn't seem to work like it does for Slider.

And I was planning to generalize it to support three tabs too.

@MichaelHale, you'll want to remove the outer Dynamic and wrap Dynamic around the second arguments of Slider and around the right-hand side of ImageSize. That will fix the dragging problem you're having.

Ah, yes. I had just tried Slider[Dynamic@x, ImageSize -> Dynamic[100 x]] and it gave me hope.

Thank you.

@JohnFultz Hello. Just by the way. Why some Options works with Dynamic and other doesn't?

@Kuba The difference is that some option values get translated directly into FE option values. Others affect kernel computation, but don't have a direct translation into the FE. And, as I've lectured many times before, Dynamic is useless unless it's displaying, directly, in the FE.

For example, look at the Plot function. Something like ImageSize goes directly to the FE. So Dynamic translates well there. But PlotPoints affects the underlying mechanisms of Plot itself, not anything the FE is actually doing. If you want to change PlotPoints based upon a Dynamic setting, then you need to recompute the entire Plot. So Dynamic needs to wrap the entire Plot.

@JohnFultz Yes, that makes sense. Thank you. I thought so, but I've faced an example where I thought that the option "could" be Dynamic but it failed. Unfortunatelly I can't recall that example now so I will just ask you about it next time. Thanks :)

@JohnFultz Just one more thing, have you seen this question about possible bug? mathematica.stackexchange.com/q/32035/5478

@Kuba There are a few cases of options that one would think would translate directly into the FE but don't for various arcane reasons. Several of the options to Grid are such an example, for reasons which are...well...messy (essentially, the kernel options for Grid don't correspond to the FE options for GridBox).

@Kuba Oh, the coincidence. Do you know what this chat is delaying me from doing? It is delaying me from putting the fix to THAT VERY BUG into the central source code repository.

@JohnFultz I'm sorry :P

@JohnFultz And about Options, I will keep that in mind, thanks.

@Kuba There, it's in now. Any one who's on the beta tester list can try it out in the next build.

@JohnFultz Ok, so if you don't mind I will quote you for other awareness. Or maybe you can put an answer there.

@Kuba, Added (see comment to Rolf's answer). I really don't like to post that kind of stuff in an answer, though. It just seems capricious to give an answer that says it's a fixed bug, but no sorry, you can't actually get the version it's fixed in because it's in beta.

@JohnFultz Thank you. Well, knowledge that this is a bug and will be fixed is important so an answer is good place for this. Rolf's answer would be enough but unfortunatelly he doesn't put any reffrences :)

@Kuba well, I tried it in the beta-test build I have, but I also thought that it does not make so much sense to put that into an answer. I wished there would be more beta-testers and the NDA would not be so strict.

@RolfMertig I didn't have any reason to not believe into your statement but you have to admit that it was automatically creating question "how do you know it" :)

I always have doubts about negation in english... 3 times in one sentence, is that ok? :p

@Kuba : I thought of linking to Lao Tzu, but that would not have been very constructive either. My mother language is german, not english, so I don't know. Where are you from?

@RolfMertig I quite close :p from Poland.

@Kuba stop by one day in Berlin.

@RolfMertig Thanks for invitation :) I will :)

@RolfMertig But I don't know when :) I'm constatly running out of time :p

I voted for this old question to be closed as a duplicate, to keep the site tidier. Please check and vote as you see appropriate. Closing as duplicate links them together.

Is anyone here?

@Szabolcs would you please take a look at this question and answer what is needed to fix that line width problem?:

I know you faced something similar with the Image Uploader, and the OP confirmed that the Uploader worked correctly, so you must know the solution. Then I can finally delete that stub answer of mine and give R.M some peace.

Hello @ybeltukov! :-)

While playing with the linked lists stuff I crashed Mathematica. I'd like to know how far this extends. I'm using version 7 on Windows. What do you use?

@ybeltukov That last question is for you. Are you still here?

@Mana can I help you? Just having a look around?

@Mr.Wizard oh, I lurk here

pay me no mind

@Mana Okay then; enjoy. :^)

:v(

That's a pretty fun face.

Hello @Mr.Wizard!

I'm using version 9.0.1 on Linux. I have the ability to access other versions on Linux.

@ybeltukov Ah, hello. I thought perhaps you were having connection problems.

This may crash Mathematica. Would you please run this code and tell me if it completes?:

SeedRandom[1]
ll = Fold[{#2, #} &, {}, RandomInteger[99, {500000, 2}]];
MemberQ[ll, {86, 99}]

No, it kills the kernel.

Well then, it seems we have a problem that runs from at least version 7 through 9 and is also cross-platform.

I think the stack size is exceeded. I didn't expect this to happen with a kernel function like MemberQ. It also happens with iterative tests such as t7 and b1.

How to format here?

@ybeltukov When you paste multiple lines of code you should see a new button to the right of the comment field appear that says "fixed font" or something like that; if you click it you should get code formatting. Or you can manually indent the lines four spaces. You can also use backticks for inline formatting.

Thanks!

@Mr.Wizard when I tried evaluate your code in the terminal it returns segfault.

@ybeltukov Good to know; thanks.

After ulimit -s 65536 it works fine!

So this is clearly a stack overflow problem

@Mr.Wizard I only had time for this now, will expand if necessary: mathematica.stackexchange.com/a/32197/12

@Mr.Wizard now it also works in the front end :)

@ybeltukov I've never had to change that parameter. Is there a downside to increasing the maximum stack size (what I assume that does)?

@Mr.Wizard Yes, this command increases the stack size but I've never played with this parameter before.

I'll be back in 1-2 hours.

@ybeltukov Okay; thanks again for your help.

a few months (ok, maybe almost a year) ago, there was a question about getting the outline curve from a plot of points. So if all the points fell roughly within a hexagon, the outline would end up looking like the a hexagon. Does anyone know what to search for to find that answer?

@tkott Very likely the thread used a convex hull. Searching for this should be a starting point.

@tkott Was it 3D or 2D?

well, I need 2D, but any example will do. Looks like that search string is bringing up some good answers

just need to look through them. Thanks @halirutan

@tkott No problem.

Hi

how I can send private message to a member?

@barznjy You can't.

@rm-rf what time zone r u at?

@VitaliyKaurov PST

@rm-rf did I understand correctly that the "party" will start here in 1.5 hours?

@VitaliyKaurov I believe Verbeia meant it to last all day long on the 13th (UTC). So yes, it starts in 1.5 hrs

what party?

@rm-rf ok thanks ;)

@barznjy chat.stackexchange.com/transcript/message/11187596#11187596

@VitaliyKaurov btw, I see 1759 deleted questions... so we crossed the 10k mark a loong time ago :)

@rm-rf you could celebrate 10,000 in each base as you pass it. Of course, 10000b2 was a while ago

@rm-rf I know, but we didn't celebrate a year of the site or a year of graduation, and it's cool that the 10,000 undeleted question mark gets passed on Friday 13th.

And it seemed so much more positive than 1000 unanswered questions :)

Yay happy 10000 question celebration spooky friday 13th everybody!

@JacobAkkerboom 15 minutes to go by my computer?

@Mr.Wizard around?

Hello @Rojo

FAST!

@Mr.Wizard Hello

I had my sound on, for a change. :-)

haha

you guys are soo lame.. you have to come late to a party! This is cool!

:-)

On the protecting some definitions only question, a couple of nights ago I had drafted something while I couldn't sleep, but didn't get to test it. Of course I won't understand my code now but i wanted to test it at least

@Mr.Wizard You had some issues with the efficiency of the old version. What exactly?

@Rojo The overhead that it added to a simple hash table look-up. I'll have to prepare an example.

The time taken if there's a big list of "protected definitions"?

@Rojo Actually I think there was only one. I'm tired right now so I'm going to be slow. Let me find the question first...

@Mr.Wizard Great. If you are interested, here's the notebook I had shared, that I'm opening right now to see if what I did the other night made sense and test the speed dropbox.com/s/7wsfbl6rc7k4sl5/AHH.nb

@Rojo I'll look, but don't expect any magic. :-p

@Rojo Okay, compare these timings (use timeAvg):

fix[f, f[0] = 0]

Do[f[i] = i^2, {i, 10}]

f /@ RandomInteger[10, 50000] // timeAvg

and:

Let's see

g[0] = 0;
Do[g[i] = i^2, {i, 10}]
g /@ RandomInteger[10, 50000] // timeAvg

The point is that f keeps reevaluating. This wouldn't matter with complex definitions of f, and I still like your way best, but it does limit its utility a bit.

@Mr.Wizard I see