@Szabolcs better not rush to that conclusion when you've not seen and experienced their site. I think they're just moving towards making that a "more official" support site for non-premier users. I look at it as the equivalent of Meta Stack Overflow for Stack Exchange or Atlassian Answers for JIRA/Bitbucket/Sourcetree, or Kiln support and FogBugz support for FogCreek software etc.

Every one seems to be doing it Q&A style these days. No reason that their site and our site can't co-exist. Afterall, the above sites and Stack Overflow seem to be able to co-exist without eating the other up.

Has anyone here used Mathematica on a cluster?

The cluster integration package seems to want to start each kernel as a separate job which is inconvenient and wasteful. But I'm not sure how to start multiple subkernels on multiple computers from a job file...

Graphics`Mesh`PolygonWindingNumber[CountryData["Australia", "Polygon"][[{1}, 1]]] gives -2

That means that the polygon is wrong and somewhere it folds on itself, right?

(I meant [[{1},2]])

Yeah

@Rojo Tasmania?

Graphics@CountryData["Australia", "Polygon"]

@Verbeia No, the big one

Graphics@CountryData["Australia", "Polygon"][[All, 2]]

It looks right visually, but there seems to be a little curlicue around Sydney Harbour - maybe a rounding problem?

Also in the south east corner near Lakes Entrance.

Found it - just north of Sydney on the east coast there is a little bay that is reversed.

i.e. the line seems to cross there.

:D

anyway, gotta go again.

@Verbeia Thanks, bye

@belisarius, negative numbers with prime Abs are primes?

@Rojo There are no negative numbers. It's the parents.

@Rojo hullo btw

@Rojo how is your sleeping schedule going?

@belisarius Hallo

It's less worse

Thanks

All good on your side?

@Rojo "less worse" doesn't sound very optimistic :(

@Rojo Well, I live in Argentina. Do you want the complete story?

@belisarius Please, no

Let me escape for a while

@Rojo yeah, better that way

@Rojo I've beaten you by 1000 chars in the code golfing for nextprime[]

@belisarius Cheater, you made me add more for efficiency, and now you compete in golfing

@Rojo hehe .. the old trick

Pizza will get me sleepy

Brb

@Rojo So @rcollyer is now a WR employee. I'll not discuss my personal affairs with nobody here in the future. I don't want them published in the next NKS edition

@belisarius I hate that you people made me learn an acronym for that book

@Rojo "you people" is derogatory, "you bastards" is much more precise

@belisarius I thought your life already looked like a CA30 rule after 100 steps?

@rm-rf More like this

t2.gstatic.com/…

@belisarius Well since you used a double negative: spill. What secrets haven't you told us?

@belisarius that's just a little wrong.

I'll go to sleep with 19k

@Rojo I am now:) Studying your "Code duuuuuump!":) It's still morning for me so the code in not speaking clearly yet.

@OleksandrR., I still was not able to find a way of how to select the result of one of the dependent variables of ParametricNDSolveValue. Say I write pfun=ParametricNDSolveValue[ .etc etc. ]. If I then do pfun[0.1][[1]] I will be attributing a value to the parameter that I want to estimate (and this is not what I want to do). Yet, if I do pfun[[1]] I get weird results..

@rcollyer Congrats on your (new?) job!

@belisarius, do you think you could help me?

@Sosi Sorry, I'm at a meeting now

@belisarius ok ok, sorry!

@Sosi hello, I am in a meeting too at the moment. I will talk to you later though.

@Rojo Did "we" make you learn OP and MMA as well?

@Ajasja I just sat down on my PC after waking up. The same second I sit, ping

and yes, of course

@Ajasja So the code duuump doesn't work properly?

Ohh, it uses CellObject

It can be fixed

@Rojo Good morning! :)

@Rojo Learning a lot of great stuff from your answer mathematica.stackexchange.com/a/17973/745

But it does seem quite complicated

@Ajasja I just edited, let's test it

:)

@Rojo Great, seems to work on MMA8 as well.

Great

I still don't quite understand why

man = Manipulate[a + b, {a, 0, 2}, {b, 0, 2}]

getOwnValues@man

returns an empty list.

But if I change some values in manipulate by hand and copy insert it into getOwnvValues[...] then it returns only the values that have been changed...

@Ajasja because you aren't initializing the variables to any value

so the value you see set is just "the default"

The first time you touch anything, it seems the manipulate automatically changes to explicit values

@Rojo Aaa

Got it! Thanks

man = Manipulate[a + b, {{a, 1}, 0, 2}, {{b, 0}, 0, 2}]

getOwnValues@man

One could change the code to extract those default values as well. But then we would have to think of what the default values are for each control type

Right

@ Rojo One thing i'm still struggling with. How come there are no context localization problems?

Where is a converted to something like $CellContexta$20139$$` in your code?

@Ajasja Those things appear in the formatting of the manipulate

As soon as the cell is read, I use MakeExpression

so they go away

Aaa I think I understand: Each time you load a controller set the manipulate gets reinitialized?

And the values are injected directly into an expression?

@Ajasja Look at getCellManipulate and replaceCellManipulate

getCellManipulate for example

In English, it says: turn the Cell expression into a nice MMA exprssion with MakeExpression, and then return the first Manipulate

Nice MMA expression means: something quite similar to what you usually write as input Manipulate. Manipulates are barely evaluated. They just display weird

And, every time I load a file, whatever manipulate is in the cell gets its controls and bookmarks replaced

so, you could add another control, remove, and modify the manipulate, and you can (or should be able to) still load a state from a file

@Rojo I think I'm beginning to understand. But is the initialization section of Manipulate reinitialized as well? (Checking now...)

@Ajasja That's a good question, didn't think of that

@Rojo Yes, it is run each time you load

man // addLoadSaveButtons

@Ajasja that's a problem

right?

@Rojo Well it could be a bit problematic.

But since all the bookmarks are saved as well, it shouldn't be too much of a problem

@Rojo And the buttons to save and load could be inside the manipulate as well (if I understand correctly?)

@Ajasja What exactly is the problem with initialization being rerun? Because I am not sure how to do anything to prevent it

Yes, buttons can be put where you like

it should all work anyway, as long as the buttons are in the same cell

If you move the buttons away from the cells with a manipulate (say palette), then we should replace EvaluationCell by First@SelectedCells, and something similar for v8

@Rojo No, it probably is not.

@Rojo Why do own values have to be wrapped in Hold?

@Ajasja Imagine x=8. If you don't hold it, then the global x is assigned 8

If you store it, say {x, 8}, then if the global x has value 2, it becomes {2, 8}

But you don't have to do that inside RuleDelayed?

@Ajasja Right, OwnValues usually stores them HoldPattern[x]:>value, so x as well as value are being held

but since I'm reading them from a Manipulate, I felt like it was simpler or clearer to store them in a Hold than to build that up

@Rojo Aha, so this "Variables State" :> {a = 0.37,b = 1.795} should be OK

Yeah

(instead of "Variables State" -> {Hold[a = 0.37], Hold[b = 1.795]})

Right

@Rojo This has been very enlightening! I always keep forgetting that Manipulate is just an expression.

@Ajasja Glad it helped. Feel free to edit the answer if you go on fixing stuff on your own

@Rojo OK will do. (Through to be honest, we'll probably not use that in production -- code that you don't understand completely is hard to maintain)

@Ajasja What are you two doing?

@Rojo An interface that is used to extract data from some experimental images.

So each set of experimental images will probably have a set of bookmarks (with all the parameters used for fitting and colouring the images).

@Rojo As a side not, do you perhaps know how/if it's possible to go from "a" to FE`a$$241 or whatever is the local equivalent?

For example that

Manipulate[Print["a = ", Symbol@"a"]; Print[Unevaluated@a, " = ", a];
 a + b, {a, 0, 2}, {b, 0, 2}, Button["change", {a, b} = {1, 2}],
 Button["change by sting", {Symbol@"a", Symbol@"b"} = {1, 2}],
 LocalizeVariables -> True]

change by string would work.

@Ajasja Nice

@Rojo I don't know what your field is, so I was being very general:)

Err, I don't think you can know. Unless perhaps you use the very varaibles to see what has been appended to their names. You have a small example of what you need that for?

@Ajasja I am not sure what my field is either :)

@Rojo Yes, then the simple solution (mathematica.stackexchange.com/a/17976/745) could work with LocalizedVariables->True

Let me see

@Ajasja I had almost forgot that you weren't the one asking the question

Well, I'm more at home here than my collage:) But I'm trying to get him to become an active MMA.SE member as well :)

@Ajasja but

the bookmark list is always localized. Why do you say it doesn-t work?

@Rojo When I printed the TypesetbookmarkList$$` it always contained localized names.

{"\" untitled \"" :> {FE`freq$$321 = 3., FE`Fun$$321 = Sin,  FE`plotLabel$$321 = "Plot Label", FE`showAxes$$321 = True,  FE`showPlotLabel$$321 = True}}

And when manipulate is reinitialized this localized names are no longer valid.

I see

Paste Bookmarks in the manipulate menu does however return normal names (as does Michael's E2 edited answer mathematica.stackexchange.com/a/17988/745)

@Ajasja The thing is

Ok, fortunately

that bookmarks local variable is in the same dynamic module as all variables

so you can use that in your favour

@Ajasja for example

getSuffix[Typeset`bookmarkList$$], ran inside a Manipulate, will give you what comes after FE`var$$....

in all the locals

@Rojo. Yes, but how do you get into the other direction. For example the first print does not work:

Manipulate[a + b, {a, 0, 2}, {b, 0, 2},
 Button["print", Print@getSuffix[Symbol@"a"]; Print@getSuffix[a]]]

@Ajasja its probably simpler to show with code, give me a minute

So if I load this {"untitled" :> {a = 0, b = 0}} from a package, will a go into FE`a$$58

@Rojo OK

We had lately a question about plotting the errors of a fit in 3d. There were some answers showing nice lines of the errors together with the surface.. Does someone remember this?

@Ajasja You're there?

@Rojo Now I am:)

@Ajasja I pasted it as a comment link

Thanks! Testing now. One small problem: if you save and load any empty list it gives you an error.

@Ajasja It was meant to show the local variable issue only

Aaa, ok.

Still studying that part...

And the previous issue is easily solved using

decodeBookmarkList[{}, suffix_String] = {};
encodeBookmarkList[{}] = {};

:)

@Ajasja or adding another dot in encodeBookmarkList[s : {(_String :> _List) ...}]

Same for decode

:)

@Ajasja I'm deleting the comment, don't lose the notebook :P

OK.

@Rojo Thanks again! Got to run now!

Chau

@Rojo Is that Argentinian for ciao?

@Szabolcs That's right

Ciao is not Spanish anyway

I want a ParallelCombine which generates the elements of its second argument sequentially, to save memory

Can someone try reproducing this: Evaluate Hyperlink["Google", "http://www.google.com"], right-click on the link and select "Copy Hyperlink". Mathematica simply crashes and quits on my machine.

$Version: 9.0 for Mac OS X x86 (64-bit) (November 20, 2012)

@OleksandrR. Thanks. I'm aware of this, but there is something else attracting me recently. Maybe someday I'll pick it up :)

@rm-rf Nothing unusual happened here. version 9 on win7 x64

@rm-rf All good, v9 w8 x64

Hmm... wonder if it's a mac issue then.

Can't help. You say mac, I think of burgers

@Silvia Hey

@Rojo Hi Rojo :)

I was trying to convert a Graphics3D to a Image3D object. And then I encountered the difficulty with getting the true PlotRange..

@Silvia I have no clue how

but I'll give it a shot

Ooooh, did you konw that PlotRange also works as a function? PlotRange[the graphics]. Still doesn't work however

@Silvia FullGraphics@Graphics[same as before] for hilariously sad breakage that shows the issues

@Rojo Yes I know that. I think AbsoluteOptions[graph, PlotRange] internally calls PlotRange[graph].

@Rojo Thanks a lot :)

@ssch It's interesting LOL And How can I never knew the FullGraphics function!

@Silvia found a probably useless workaround, open drawing tools and draw a point, run AbsoluteOptions[< graphics here>, PlotRange] I have no idea what is going on :D

@ssch This is now really interesting :O

get coordinates works too, but it has to be from within drawing tools, just right click->get coordinates does not work

Have to sleep now. I'll see if I can work around with your clue when I get awake :) Thanks!

@ssch I need to do it programmly :(

Probably

the PlotRange becomes explicit when you mess with it manually

@Rojo Yes. Having a look at the FullForm can confirm a explicit PlotRange is inserted.

Must be possible to automate whatever get coordinates does

So Mathematica do have a way to know the right answer, just how?

@Silvia We could inspect the drawing palette

@ssch btw right click -> get coordinates do give me right coordinates

Maybe version related?

@Silvia oh now it works for me too, must have done it wrong the first time :)

@Rojo I bet it will end up with some built-in, protected functions :)

@Silvia I have faith that whatever the button that draws a point does, can be done programatically. Some FrontEnd token ;). YOu gotta have faith

OOR, let's try it through LinkSnooper

Nothing :(

@rm-rf Once up a time, when things got wrong, I saw the AbsoluteOptions[graph, PlotRange] gave me a result like {PlotRange -> PlotRange[graph]}

@Rojo All blame on you, I'm struggling whether reading the code of the drawing tools palette or going to bed..

@Silvia Why go to bed, it's only 7:25 pm :)

but I am not sure the palette will help. It sets CurrentValue[$FrontEnd, "Current2DTool"]

Soo verbose

FrontEnd`AddUsedToGenerateSideEffectGraphicsFunctions

I see... So HAVE to sleep now. Thanks you friends! I'll see if I wake up with an answer presenting right here :)

@Silvia That's the attitude we expect from our OPs

:D sleep tight

@Silvia hehe, that seems to be how everyone finds it :)

RSolve`RSolveDegreeBoundForPolynomialSolutionsOfLinearFirstOrderODEs

That's why I'm so happy with every little improvement in the autocomplete