Wolfram Mathematica

5:48 AM

@sacratus Hi.

@sacratus If you look at

FullForm[MessageName[#, "usage"] &]

you see why your last suggestion will not work. Because the above is a Function

Function[MessageName[Slot[1],"usage"]]

When you now do

MessageName[#, "usage"] & @@ MakeExpression["$CloudRootDirectory"]

you replace the head of

HoldComplete[$CloudRootDirectory]

(which is what you get when you evaluate the MakeExpression part)

So the next evaluation step of Mathematica is to turn this into

Function[MessageName[#, "usage"]][$CloudRootDirectory]

Exactly at this point $CloudRootDirectory is evaluated, because usually Function does not hold its arguments.

When I used

Function[s, MessageName[s, "usage"], {HoldFirst}]

I told Function to hold its first argument.

9:03 AM

fyi: Summary of New Features in 10.1 reference.wolfram.com/language/guide/…

2

@Nasser nice, is there any estimated release date?

blochwave

9:35 AM

As always, I get this email "Mathematica 10.1 is scheduled to release soon, and your site will
be one of the first to get this upgrade.

Mathematica 10.1 adds many enhancements to performance, coverage,
and convenience--as well as several major new algorithmic
achievements such as image identification and user-defined
grammar rule deployment."

10:43 AM

Again this rolling upgrade. I suspect it's not coming too soon to me, because as Home Edition user I might overload their support channel!

@blochwave Is it a free upgrade?

blochwave

11:00 AM

"================
EXISTING USERS
================
Upgrading to Mathematica 10.1 will require a new activation key.
Users will be prompted to enter a new key regardless of whether
this is a new installation or an upgrade. Existing Mathematica
users who have been assigned a key through the Wolfram User
Portal will be automatically notified of this release, as
follows:

* Anyone using a key that is flagged in the Portal as "Auto
Upgrade = Yes" will be sent a new activation key for Version 10.1
and given the option to download the upgrade immediately. They

afraid that;s all I have

@blochwave Eh. Interesting.

11:18 AM

@Kuba I do not know. I checked my premier service portal area, and it is not there yet. May be it will be announced soon.

Use of Rule in GrammarRules seems a bit counter-intuitive to me: reference.wolfram.com/language/ref/GrammarRules.html

@kirma 10.1 should be free upgrade if you have premier service, otherwise, I would not know. Normally free upgrades are given to those with premier service only.

@Nasser Yes, I suppose in about $WolframWeek.

@Nasser Point being here: it's ages since there has been minor, but not purely a bug-fix release.

5.2 was released 2005.

I think there wasn't Home Edition back then.

11:36 AM

@kirma yes, home edition is new thing. Not sure when it started, may be with version 9? you are right. Last time there was a .x release is 5.2 in 6/20/2005

@Nasser Interestingly enough I see a Home Edition order invoice in my mail archive dated early 2012, which would be 8.0.4 or so. (Until then I used university license and student license, which I could actually still use - but I don't really care about hassle involved.)

@kirma yes, according to this wolfram.com/mathematica/quick-revision-history.html home edition was around version 8.

12:14 PM

my quick and dirty animation of 3 links robot arm (RRR arm). Will do six link arm next:

the equations of motion are very complicated to get actually, much harder than making the manipulate itself. Used NDSolve to solve them.

I'd say amount of content in 10.1 is rather modest, and that should be reflected in upgrade cost, to an extent.

@kirma How much does WRI charge for the update to 10.1 from 10.0.1 for non-premier-support users?

12:30 PM

@halirutan I have no idea. :o

@kirma Ah, OK. I thought you knew.

Major-to-major version upgrades seem to have been 95 euros plus VAT. It would seem mostly unnecessary hassle to charge less than 50 euros for a minor upgrade...

Wolfram Blog: New in the Wolfram Language: PlotThemes for Gauges

12:51 PM

This might be of interest to people here: Z3 prover is now available under MIT license (that is, you don't need to pay $10000 or so for commercial use any more, etc.).
http://www.reddit.com/r/programming/comments/30ejp5/z3_is_now_really_open_source_mit_license/

Wolfram Blog

2:08 PM

posted on March 27, 2015 by Tim Shedelbower

The first gauge I remember was a blue wrist watch I received from my parents as a child. Their hope was probably to correct my tardiness, but it proved valuable for more important tasks such as timing bicycle races. Today digital gauges help us analyze a variety of data on smart phones and laptops. Battery [...]

2:40 PM

Does someone know whether there is a built-in alternative to URLSave for Mathematica 8?

I want to download (not extract) a zip file to the local disk completely from within Mathematica.

You can use this master.zip for testing purpose if you like. AFAIK there is no solution, but I might misremember.

2:57 PM

@halirutan My guess would be to try Import, but manually specify a file format like Byte or String to override the .zip interpretation. Then write the results to a file manually.

3:09 PM

@MichaelHale If it works with an URL this could probably the way to go. Thanks

@halirutan You may try file = Quiet@FetchURL[url, FileFilters -> {}], where FetchURL is from "UtilitiesURLTools". This is what I was using in my ProjectInstaller.

@LeonidShifrin I was searching exactly for this! I just couldn't find it in the docs

@LeonidShifrin I searched for URLTools and nothing came up

@halirutan I've spent some time to find it back when I was writing ProjectInstaller. I think I had to hack the code to find it.

@halirutan Actually, ProjectInstaller code might have other pieces you may be interested in, because it was basically solving a similar problem to what you seem to be after.

@LeonidShifrin Do you use FetchURL for all versions, even >8?

@halirutan I have to admit that I simply did not have the time to update that stuff, so yes, I do.

3:13 PM

@LeonidShifrin I guess this is where I saw FetchURL anyway. I was looking through your code back in the days

@halirutan It is quite bloated code, actually. It should be separated into several smaller modules,because part of it is general, while other parts are more specific to the project's structure I had in mind etc.

@halirutan It also involves Java, because I used Java code to extract the archive, not being aware of ExtractArchive at the time.

@halirutan Sorry, it was displayed incorrectly - I meant, Utilities`URLTools`

@LeonidShifrin Yep, the moment you gave me the larger Utilities hint, I knew how to find it again :-)

@LeonidShifrin I have finally improved the SE-Uploader and pulled of a very extensive README

@halirutan I think, I learned about this function from Oliver's post on SO (or SE?)

@halirutan Wow, that's really impressive! That's how it should be done.

@LeonidShifrin Look what we can do now:

Import["http://goo.gl/NaH6rM"]["http://i.stack.imgur.com/EBnmo.png"]

(I meant the notebook-sharing, not the content of this randomly selected notebook :-)

@halirutan This is pretty cool! Very very nice.

@halirutan I wonder whether it is likely that SE staff would eventually block this possibility somehow - although, this doesn't seem very likely.

3:28 PM

@LeonidShifrin Yes, I have considered this possibility. I would be interested which SE sites post the most images.

Btw, the upload content is restricted to 1MB anyway, because then, imgur automatically converts the PNG images to JPEG, and lossy compression is not so nice if you want to transfer data :-)

@halirutan Presumably those, whose main topics are very visual, like may be woodworking (should ask @rm-rf), or may be home improvement or travel or something like that.

@halirutan Yep, that's a good point.

@halirutan Yesterday, while answering a question on Russian - speaking Mathematica forum, I have constructed a tiny package to measure elapsed execution time in cells when they are evaluated. Not sure how useful this is, but it was fun, and I actually started using it myself.

@halirutan All you have to do to start is to load the package

Import["https://gist.githubusercontent.com/lshifr/71977c7c0ea7469c5bcf/raw/CellT‌imeTrack.m"]

And then execute CellTimeTrack[True]

Then, any cell with a computation will get a counter on the right in seconds, for example

i = 1;
Do[i += 1, {100000000}]

To exit this mode, one executes CellTimeTrack[False]

3:44 PM

@LeonidShifrin Can you check your URL and repost it as code? (4 indents work here too (or Ctrl+K))

Import["https://gist.githubusercontent.com/lshifr/71977c7c0ea7469c5bcf/raw/CellTimeTrack.m"]

@halirutan Sorry about that, somehow the first one was invalid.

@halirutan which is weird, since they look exactly the same.

@LeonidShifrin I guess you have entered some magical byte code

@halirutan Well, may be

@LeonidShifrin I get an error after the loop is finished

@halirutan Which platform do you use?

3:48 PM

Otherwise it works perfectly

@LeonidShifrin Linux, V10.0.1

@LeonidShifrin Sorry, I mean V10.0.2

@halirutan Thanks, this is an important data point. I only tested on Mac. When I get my hands on a Linux machine, I will try to fix that.

@LeonidShifrin Let me ask you something. Maybe you'll have an idea how to solve this:

@halirutan One thing I like about it is that it handles Aborts correctly, and generally does not require any additional infrastructure (new cell, etc) - I can just execute the code normally. I can see many cases where this would spare me from wrapping stuff in AbsoluteTiming

@halirutan Sure

@LeonidShifrin I will definitely look at its implementation.

@halirutan Well, it is nothing special. Just tiny fun problem, which might be somewhat useful.

@halirutan So, what's the stuff you trying to solve?

3:54 PM

@LeonidShifrin I'm not sure whether this link works, but let me show you a discussion of Vitaly about animated GIF and manipulate:

goo.gl/JgRwkt

I had to shorten it. It won't fit into this message box.

The basic thing is that he exports a Manipulate as movie and re-imports the frames to re-export them as gif.

This is IMO one of worst things you can do, because movie compression is made for movies and plots can be handled differently.

Here you see the GIF that I created which is not only very small, but has a better quality:

@halirutan How did you do this?

@LeonidShifrin A tool that I always use which takes the screen and encodes the GIF directly.

@LeonidShifrin The question (which is completely unrelated) is now, do you know whether it is easily possible to clone a Manipulate so that all actions are reflected?

@halirutan Ok, so this is an external tool, which has nothing to do with Mathematica. Makses sense. But how did it know which portion of the screen to work on?

@LeonidShifrin You can just select the region.

@halirutan Oh, I see.

@halirutan So what exactly we would mean by "clone" here? Manipulate doesn't have a history of how it reached a given state, does it?

4:00 PM

@LeonidShifrin I want to make captures (as images) of a Manipulate while using it and everything inside Mathematica

@halirutan So, basically, rasterize at run-time?

@LeonidShifrin One way would be to keep track of the value-changes, yes. But I would like to do it life

@LeonidShifrin Yes

@halirutan It might be possible. If we refer to EvaluationCell[], we could read its contents, and rasterize it, I think.

@LeonidShifrin Or using NotebookFind should be possible too, right? If we would tag the cell for instance

@halirutan Yep, but it would probably be slower, and is a bit of a round-about method. One could perhaps also wrap the entire Manipulate in Dynamic, that would track and rasterize it.

4:05 PM

@LeonidShifrin Yes, that was exactly what I was thinking of, but in a short test it didn't work as I expected it.

@halirutan Do you have some simple example you tried handy?

@LeonidShifrin First I tried just

Dynamic[t = Manipulate[Plot[Sin[x (1 + a x)], {x, 0, 6}], {a, 0, 2}],
 TrackedSymbols :> {t}]

(I tried without TrackedSymbols first, but none works)

@halirutan Give me a few moments to play with it a bit

@LeonidShifrin Yes. For me it came down to how clone the $CellContext variables used in one place to another place and reflect the changes.

rm -rf

A quick question @halirutan: You use IDEA on Ubuntu, right? Do your fonts in IDEA look crappy (thin, anti-aliased) on Ubuntu compared to OS X? If so, do you know how to fix it (if it can be)?

Also, hi @LeonidShifrin!

4:11 PM

Hi @rm-rf! It's been a while :)

@rm-rf Look here

@rm-rf I'm not sure whether this is thin but this is how it looks

(I always felt that Apple does a bit too much blurring for the fonts)

@halirutan Indeed, looks like a tough cookie. I have a feeling that the solution is somewhere close, but so far wasn't successful. I'll try to play with it more. Will let you know if I manage to get to something interesting.

@LeonidShifrin The dumb thing is that I clearly remember John Fultz talking about dynamic variables and how they can be shared, but I don't find the reference any more.

@halirutan Assuming that this works, what would you like to have at the end - do you want the rasterized images to accumulate dynamically, as we move sliders etc?

@halirutan Yep, I also seem to remember something like that. Will try to recall where was that dicsussion.

@LeonidShifrin Yes, and since the rasterized images don't have the compression artefacts, it will give a perfect GIF

4:21 PM

@halirutan Ok, I see. Alas, I have to go now, but I will spend some time on this later.

@LeonidShifrin No hurry. It was just an idea. See you later and nice talking to you after such a long time.

@halirutan Same here!

@rm-rf Gtg now, but hopefully there will be another time soon!

rm -rf

5:25 PM

@halirutan Yeah, that's somewhat how it looks. For some reason I like OS X's font rendering (maybe because I've been using it for quite some time now)

Szabolcs

5:53 PM

@rm-rf I absolutely hated OS X's font rendering when I first started using it, but the retina screen fixed that completely. Later I had the chance to try Windows on a high dpi screen, and OS X looks better there. On low dpi screens (say ~1200 px wide 15"), I still prefer Windows for readability, even knowing that its rendering is not accurate ...

When the pixels are so small that I cannot see them then it's the accuracy and beauty of the glyph shapes that matters most.

6:28 PM

@halirutan do you want to capture whole manipulate or only the body?

@Kuba It was only an idea because I think the approach of exporting a manipulate to mov first and then create the GIF is the wrong way. Furthermore, you can only use the automatic moving mode of the manipulate and you cannot show specific things which renders the whole approach useless for most of the cases.

Another thing makes me preferring an external tool: I cannot catch mouse-movements or other things like how I type in a normal notebook.

@Kuba But just for the sake of argument: Would it be possible to have a function like

@halirutan I completely agree :)

Record[Manipulate[...]]

which gives me a record and a stop button and stores frames of what I'm doing in the manipulate.

Ok, but whole picture or only the body? Controls too?

(I mean in its essence, somehow the Export needs to do this too)

@Kuba Controls too. That's one point that makes it harder.

6:34 PM

@halirutan I see

@halirutan will pin you if I find something useful. Don't know how much time I can spend though :)

@Kuba As I said already to Leonid, there is no hurry. It was just an idea and the way how this can be implemented is more interesting than the result itself.

@Szabolcs I feel exactly the same. On my Retina Macs, I love how everything looks. On older machines with normal resolution, I would probably prefer the thin and edgy fonts of Ubuntu.

6:56 PM

@halirutan I just quickly tried Pause@5; Table[Pause@1; Rasterize@SelectedCells[], {10}] Then I would select the cell with the Manipulate and use it for 10 seconds. The Rasterize often catches it in the middle of an incomplete state though.

FinishDynamic didn't help.

@MichaelHale That was what I was afraid of. With NotebookFind this problem will probably become worse.

@halirutan @MichaelHale Try this, first create:

Pane[
 Manipulate[{x},
  {x, 0, 1}],
 BoxID -> "TEST"]

then run:

boxes = {};
RunScheduledTask[
 boxes = {boxes,
   MathLink`CallFrontEnd[
    FrontEnd`BoxReferenceRead[
     FE`BoxReference[InputNotebook[], {{"TEST"}}]]]};
 Print["recording..."]
 ,
 {1, 10}]

And play with manipulate for those 10 seconds

at the end run:

CellPrint /@ Cell /@ BoxData /@ Flatten@boxes

This is not final and has pros/cons but it is a way to start

Michael E2

7:12 PM

@halirutan You may be thinking of DynamicModule Wormholes.

@MichaelE2 Yes. This is what I remembered.

@Kuba Not bad for a start!

Something just crashed my Mathematica :-)

@halirutan somehow after second run it stores only the same state. don't have time to test it now much more.

@halirutan not me :p

No problem. Nothing important was open

Szabolcs

@YvesKlett I'm glad you find it useful :) If you have any specific comments or pain points, let me know!

Michael E2

@halirutan This is something I did prompted by Vitaly's solution: mathematica.stackexchange.com/questions/76832/…

7:20 PM

@MichaelE2 But it follows more or less the automatic controlling approach, right?

Michael E2

@halirutan Yes, it does not follow the mouse. The basic idea was to inject values into the Manipulate variables, and let the front end render it. Maybe you could save the variables' values under live manipulation and render at leisure.

@halirutan p.s. on the other hand if you know where manipulate box is then you can take SS:mathematica.stackexchange.com/a/77989/5478

and crop it

@halirutan @Kuba @MichaelE2 Expanding on the capture input suggestion some:

Column@{Manipulate[Plot[Sin[a x], {x, 0, 6}], {a, 1, 5},
   LocalizeVariables -> False],
  Dynamic[Refresh[a, UpdateInterval -> 1], TrackedSymbols :> {}]}

Or I guess you don't need the Refresh, just put the UpdateInterval as an option to Dynamic.

7:43 PM

@MichaelHale @halirutan I was playing earlier with the code to inject such record button to Manipulate. I'm at the point that I can inject a button that automatically prints variable state, which is nice. But again, I don't know if I manage to find more time for this soon

ClearAll[addrecorder];
addrecorder[man_] := Module[{boxes, vars, body}, boxes = ToBoxes[man];
  {vars, body} =
   Cases[boxes,
       Verbatim[
         RuleDelayed][#, _], \[Infinity]][[1]] & /@ {"Variables",
     "Body"};
  vars = vars /. RuleDelayed[_, x_] :> Hold[x][[{1}, ;; , 1]];
  ToBoxes[Button[1, Print@ReleaseHold@vars]];
  boxes =
   boxes /.
    HoldPattern[
      Typeset`specs$$ = {x___}] :> (Typeset`specs$$ = {x, {Hold[
          Button[2, Print@ReleaseHold@vars]],
         Manipulate`Dump`ThisIsNotAControl}});

@Kuba Wow, pretty cool. @halirutan As far as I got:

input = {}; output = {}; recording = False; Column@{Manipulate[
   Plot[Sin[a x], {x, 0, 6}], {a, 1, 5}, LocalizeVariables -> False],
  Dynamic[If[recording, input = {input, a}]; a, UpdateInterval -> 1,
   TrackedSymbols :> {}],
  Row@{Button["Start", recording = True; input = {}],
    Button["Stop",
     output = Rasterize@Plot[Sin[# x], {x, 0, 6}] & /@ Flatten@input;
     input = {}; recording = False]}, Dynamic@output}

Q: NotebookWrite in EvaluationBox forces Input CellStyle

8:13 PM

@MichaelHale Thanks

p.s. why those functions are released before testing :(

0

There is a promissing new function EvaluationBox[] and nice last example in documentation section for it: Button["date", NotebookWrite[EvaluationBox[], ToBoxes @ DateString[]]] (Click) The problem is that this string is in "Input" cell now! Why it was converted? Quite a problem ...