Haha

@Rojo the last step is to be outgoing, and it has its own guide!

haha

@Leonid To me it was a really a clumsy strategy from Wolfram to attack Hierarchical data representation first and than (maybe in the future) go to Tabular one. If ByteCount is right Dataset get 19x more memory due to it line verbosity!... For me is a paradox to marketing V10 as ready to DataScience if it haven't capacity to efficiently handle the most basic data format, that is the tabular one.

Another point is about Pivot, the function was remove from Docs these days. There is no ways to simple Pivote data in Mathematica, and Excel is still (unfortunately) the best tool to do this.

Wait more 2 year for this is a lot of time... It's a pity

I think Mathematica is being spread too thin :), trying to do too many things at the same time. Not good I think. Should just have concentrated on basic science and engineering and graphics, like Matlab does. But what do I know.

@rm-rf The expert user would giggle like he just got a pony for his birthday and break out the spelunking tools. And keep good notes for a new section in his book :D

@WReach Sooo.... when is your book coming out? :D

@rm-rf "the garbage collector ate it, sir!"

@rm-rf @acl :) There is something to be said for being strongly reachable...

@kirma Ahh, but a most interesting rant.

@kirma It is a characteristic of many (all?) high-level languages that they have opaque performance models. It is scary to think that C's performance model is getting murky -- what hope is there for Mathematica? :)

@WReach The best thing is "weltschmerz" which is a German word no one would use here.

@kirma for, if and = map nicely to conventional von Neumann architectures, but are decidedly less friendly for modern out-of-order processing, caching, pipelined, parallel, superscalar microarchitectures. But, if we squint a bit, they seem well suited for functional-programming in the Backus style. There's still hope?

@halirutan I have at least two pieces of evidence to the contrary :D

@WReach I maybe read it one or two times in books, but no one I ever met has used it in real life.

@kirma the microarchitectures are suited for FP, that is -- not for, if and =

any one knows of a trick to rotate group of graphics at once, instead of putting Rotate around each? I am doing this now"
Graphics3D[
 Rotate[
  {
   {Opacity[.1], Cuboid[{-w/2, -d/2, -h/2}, {w/2, d/2, h/2}]},
   {Red, Arrow[{{0, 0, 0}, {.6, 0, 0}}]},
   Text["x", {.65, 0, 0}],
   {Red, Arrow[{{0, 0, 0}, {0, .6, 0}}]},
   Text["y", {0, .65, 0}],
   {Red, Arrow[{{0, 0, 0}, {0, 0, .6}}]},
   Text["z", {0, 0, .65}],
   Sphere[{0, 0, 0}, .02]
   }, thetax, {1, 0, 0}
  ],
 Axes -> False, AxesLabel -> {"x", "y", "z"}, PlotRange -> {{-.6, .6}, {-.6, .6}, {-.6, .6}},

But I get the standard error that Rotate is not a graphics3D primitive due to this.

I forgot now the trick to do this. I think I've done this before.

@Nasser Wrap your graphics objects in GraphicsGroup

@Pickett thanks. Tried it, but still get the same error. Never knew about GraphicsGroup. Will go read more about it now.
Framed@
 Graphics3D[
  Rotate[GraphicsGroup
    [
    {
     {Opacity[.1], Cuboid[{-w/2, -d/2, -h/2}, {w/2, d/2, h/2}]},
     {Red, Arrow[{{0, 0, 0}, {.6, 0, 0}}]},
     Text["x", {.65, 0, 0}],
     {Red, Arrow[{{0, 0, 0}, {0, .6, 0}}]},
     Text["y", {0, .65, 0}],
     {Red, Arrow[{{0, 0, 0}, {0, 0, .6}}]},
     Text["z", {0, 0, .65}],
     Sphere[{0, 0, 0}, .02]
     }], thetax, {1, 0, 0}

That code works for me.

@Pickett well, for some readn, it does not work here. V 9.01. But will keep trying

I'm using v.9 as well... What values do you use for w, d and h?

@Pickett now this
{{w, .7, "width"}, .1, 1, .1, Appearance -> "Labeled", ImageSize -> Tiny},
{{h, .7, "height"}, .1, 1, .1, Appearance -> "Labeled", ImageSize -> Tiny},
{{d, .7, "depth"},

Framed@
 Graphics3D[
  Rotate[GraphicsGroup
    [{
     {Opacity[.1], Cuboid[{-w/2, -d/2, -h/2}, {w/2, d/2, h/2}]},
     {Red, Arrow[{{0, 0, 0}, {.6, 0, 0}}]},
     Text["x", {.65, 0, 0}],
     {Red, Arrow[{{0, 0, 0}, {0, .6, 0}}]},
     Text["y", {0, .65, 0}],
     {Red, Arrow[{{0, 0, 0}, {0, 0, .6}}]},
     Text["z", {0, 0, .65}],
     Sphere[{0, 0, 0}, .02]
     }], thetax, {0, 0, 1}
   ],
  Axes -> False, AxesLabel -> {"x", "y", "z"}, PlotRange -> {{-.6, .6}, {-.6, .6}, {-.6, .6}},
  SphericalRegion -> True, Boxed -> False, ImagePadding -> 2, ImageSize -> 400

Yep, definitely works for me.

@Pickett strange. May be because I have everything inside Grid? as in Grid[{{ Framed@....}}] but it should not matter. But will keep debugging.... I am making a quick demo for the Cloud to try it, which does not use Manipulator...

@Nasser Ok, I hope you solve it.

@Nasser That code works for me as well, as does the original code. I do, however, get the error message you cite if thetax has no value.

@WReach oh, just put any value for theta. It is the current rotation angle. Like thetax=Pi/4 or anything. I am making a spin control demo, and need to rotate an object around the axis to check for stability. so I update the equation of motion, and get new angle each time tick...

@Nasser Right. If I give it a value, I see the expected graphic. If it is unbound, then I get the error message you see.

@WReach I see! you mean you see the same error as me when thetax has no value?

@Nasser Yes.

@WReach that was it!! thank you. I had not yet set thetax in the code from the solver before plotting, forgot about it. Strange error message for this. Oh well. Thanks everyone.

@Nasser Glad I could help :)

I solve the equation of motion and forgot to update thetax

Error messages in Mathematica really needs to improve and be more specific. one would expect an error such as "Rotate: Angle of rotation is not set" or something. But "Rotate is not grahics3D primitive"?

Ok, this is not complete and have bugs I am sure. BUt I just deployed this here wolframcloud.com/objects/91a71492-d424-4f6c-a5b6-306f592a5124 I selected the manipulate cell, and did share cell, and selected CDF and general URL. THe above is the result. But when I click on buttons nothing happens. The sliders seem to work.

I do not think this is working.

I tried again. Can someone please try this one wolframcloud.com/objects/abf245af-86ed-498a-aa7f-69129a7ac6ec

if you move the sliders for the angles, the box moves!! but the buttons do not work.

Yes, all the sliders work, except buttons. May be buttons not supported yet in the cloud.

@Nasser I think you're right. The cursor doesn't change appearance upon hovering the button either, so it seems like they haven't paid any attention to buttons at all.

@Pickett I just posted question here in Community on this: please see community.wolfram.com/groups/-/m/t/283411?p_p_auth=s2Hl4Gkr

I think it is a bug in the Cloud. No one tested the buttons.

@Pickett No button support. No Manipulator support. This not not going to work!

@Nasser I can't imagine that the developers don't know that buttons don't work, but still I think it's a good idea to ask on the community site to raise awareness that there is some discontent about it.

@Pickett testing all of M functionality in new environment, is a huge task. Given how complex M is. And with the internet in between, this makes this task much more complex I think.

@Nasser As you say there are many things missing. Locators don't work either, for example. I think it's a WIP.

@Pickett yes, WIP indeed. I'll stick to normal CDF

They should not have released it in this case, and waited untill it is more complete!

@Nasser It's very common to release substandard products, and there are many advantages for WRI in doing so. I don't see any particular downside to it, at least now I know what they're up to.

Also this is only a small part of the entire product, and the manipulates with only sliders that I tried worked nicely. It's already a pretty good product in some respects.

@Pickett I'm more worried about Mathematica 10. Mathematica tends not to get many updates once it's released. If M10 turns out to be buggy, that's fine, but then they should keep doing point updates regularly.

Web-based applications tend to get regular updates, it's easy to roll them out. So the cloud is a bit different.

@Pickett You said locators do not work in the cloud. Can I add this piece of information to my post at community? I can either just say I read this at stackexchange or just mention it as is. Though to see what you think.

@halirutan I'd say Finnish equivalent of weltschmerz, "maailmantuska", a literal translation, is relatively commonly used in .fi.

Usually somewhat sarcastically, but nonetheless.

@Murta re: dataset stuff: have you tried python pandas for doing pivot table stuff. I find it pretty awesome ... really pandas in general is amazing (for what I have used it for). It absolutely kills me to see how preliminary Dataset is. All this cloud stuff I am indifferent to (as it seems expensive and niche to me) but modern access to tabular data (pandas, R like) was why I was waiting for v10 ... it is getting hard to be a Mathematica fan these days

Upgraded Programming Cloud to Explorer level. I don't see any sensible improvement on getting things evaluated, though.

@Szabolcs We can always hope

@Nasser Sorry, I had to go. I'm not an expert on the locator thing, but if you try for example LocatorPane[{1, 1}/2, Graphics[{Gray, Disk[]}]] it's the same as with the buttons. The locator appears but it's inactive.

@Szabolcs I agree that might become a problem.

@Pickett He's probably alluding to the fact that many bugs don't get fixed

for example this

Yes, I got that. I said "might" because MMA 10 isn't released yet.

@Pickett If lots of non-interface bugs get fixed, it'll be the first time

So I mean, it won't become a problem. That's been the situation for years

Each version adds new stuff on top of the whole creaky structure :)

Hmm with all the complaining I do, I am starting to feel like on of these gentlemen

@acl I see what you mean, I should have expressed myself more carefully on such a sensitive and emotionally fraught subject ;)

I hope they will start to release regular bug fixes though. I'll try to be optimistic.

@acl I did not follow the link (yet), but I guess you´re a Muppets fan?

... right your are!

@Murta It was not my decision to do it this way. Initial design was different, we had tabular data as the main workhorse, while hierarchical data was planned to be fully supported later. The design was then changed by the powers that be, and since then I am no longer the lead developer of Dataset. I can understand the rationale behind the current design, it might indeed be better as a general design for the Dataset, but I agree with you about the readiness to do data science and efficiency.

@YvesKlett where else could that link have led to...

@Murta I think these issues will be eventually sorted out, and hope this will happen pretty soon. You won't need to wait until the next major release for that - I think this is a matter of several months, not years.

@acl yeah, not really a tough prediction (only if grumpy cat had a twin).

@Gabriel It may sound strange, but I feel your pain. We did have these designs at some point, so hopefully it won't take too long to add the efficient tabular data functionality. If it had been under my complete control, tabular data manipulations would've been there from the start.

@LeonidShifrin I have been intently watching everything about Dataset (even since it was called a DataStore :) So I know that very smart people (who seem very aware of the other products that do this type of thing very well, unlike the current marketing push). It just absolutely kills me as I find myself having to go back to python/R increasingly to have a nice interface for this kind of thing.

I moved to mathematica because when it does something well it is just such an amazing environment to do science. V10 is just starting to feel like science is becoming less of Wolframs focus which just kills me. Still I will keep hoping that if not for v10.0 then v10.1 will slow my move to python!

Content from this question keeps getting rehashed by different users: mathematica.stackexchange.com/q/51692/131

Either this is some really mangled homework or one user with multiple accounts?

@YvesKlett "same.Follow me answered. What does it matter to you?"

haha

@acl oh dear.

And it has a reverse question mark at the end. Right-to-left script?

@Yves Did you already flag it? Moderators can see IP addresses.

@Szabolcs yes, flagged it now. It is not overly rude, but a waste of our time.

Does anyone know of a ready-made solution for including beautiful LaTeX generated formulae in Mathematica figures? I think I can implement it, but I don't want to spend time on it if someone has already done it.

@Szabolcs I think I used psfrag for this once upon a time.

@acl My knowledge of LaTeX is very limited. Thanks for the suggestion! It looks like it's what I need

@Szabolcs I thought psfrag would be too obvious :D

@rm-rf what is it with SphericalBesselJ?

@YvesKlett what's it to you?

@acl same!

@YvesKlett no I was trying to be a jerk like the person asking that question

I wonder how they all ended up here

@acl so was I (not to worry, my grasp of irony/sarcasm reaches that far) :D

@YvesKlett right!

OK I guess everybody here knows the answer to this but can't be bothered to write it up

What do we do with it?

@acl congratulations! You got the job!

@YvesKlett heh. No!

@rm-rf Well, what are you waiting for?

@rm-rf well at least the OP is polite and gives a coherent explanation.

@rm-rf there should be a tweet button for really bad content

@acl I was going to tell him to press F1 and check for himself, but then I looked at RidgeFilter docs myself, and it was useless.

If you're really new, then seeing RidgeFilter[<actual image> ...] is not at all helpful... where do you go from there?

@rm-rf Yes most of the docs are pretty useless if you're really new

@acl not sure. I still think the help system is great (and thought so the first time I found it).

@YvesKlett I still have trouble with it if I try to use new classes of functions.

Any thoughts on whether or not "How come this doesn't work on the Cloud" style questions are useful at this point? Seems like much of the material I've 'deployed' previously as CDFs is giving the cloud trouble.

@bobthechemist might this be better of at community?

@YvesKlett probably - I just lurk here more.

@bobthechemist I'd vote to make most cloud-stuff on-topic as it's based on the same programming language as the "old" Mathematica. But a lot of Dynamic stuff is simply not implemented for the cloud yet (just ask Nasser), so there's not other good answer than "cloud is beta, it doesn't seem to be implemented yet".

@bobthechemist If you think it's a bug, just send them some feedback using the button on the bottom right.

@Szabolcs I'm running in to problems with Manipulate, so if the community knows that dynamic functionality is still buggy in the beta form, I'll let people work instead of dealing with my rants.

@bobthechemist A lot of that functionality seems to be missing, and some of it works only partially.

@Szabolcs I certainly like have access to the "cool new stuff" but I wonder if this shouldn't still be in 'private' beta.

@bobthechemist You might want to follow this thread. Let's see if the question gets answered.

Does anyone know anything about this? community.wolfram.com/groups/-/m/t/282440?p_p_auth=xeWw69Kt

Any one have a function to do the following without the Map. a = RandomReal[10, {10}]; Min[#, {5}] & /@ a

@brama Clip[a, {-Infinity, 5}]

@YvesKlett awesome....that works better

@brama Clip is also quite fast (as I keep telling anyone who wants to listen) ;-)

@YvesKlett I am going to help your cause by "starring" the suggestion :)

Great, now we are going to have a bunch of Cliptomaniacs running around.

@bobthechemist haha....do you have a suggestion that works better?

@brama If I were you, I would do my best to avoid suggestions from those who feel it necessary to post really bad puns.

@bobthechemist huh?

Anyway, see you later...

@YvesKlett thanks and bye!!

@YvesKlett Making fun of myself

a new and exciting question!

I think I have reported four bugs on Programming Cloud now, three on geometric computation and one on other matters...

Their response time is convincing, but I sense they do employ a bit too much automation on responses at times.

I can't read many of the questions's code being posted here lately. It is all full of subscript this and subscript that. Like this Subscript[j, [Nu], 1][s_]. Terrible looking code really. I am sure it looks better in the original notebook, but even if I copy it to my notebook from the post, it is still in the InputForm and still hard to read.

How can I use clip with vectors instead of using Min & Max? a = RandomReal[10, {10}];b = RandomReal[10, {10}]; Clip[a,{0,b}] is there an alternative way?

is MapThread[Clip[#, {0, #2}] &, {a, b}] my only option?

@Nasser you have been Bessel´d!

Out-Bessel'd?

@YvesKlett I am busy now making list of all dynamics control that do not work in the Cloud. Going over them one by one and trying them. found 5 controls so far that do not work.

Maple is trying to compete now with Wolfram Cloud. They are starting new web thing. I think everyone wants to go to the cloud these days.

@Nasser you should charge per item

@YvesKlett it must be really hard implementing dynamics in the cloud. It is hard enough on the PC.

@brama what´s wrong with MapThread? Looks good.

@YvesKlett I am not sure if MapThread works in parallel and I am guessing inbuilt functions work in parallel when used in a compile

@YvesKlett Please correct me if my understanding is not correct

@brama not sure parallelizing will be very effective. You could partition your lists and start multiple MapThreads.

@YvesKlett oh....but the issue is that I have to do multiple iterations on 1X180 size vectors

@brama the proof is in the pudding - benchmark away!

@YvesKlett sorry..do not understand what you mean

@brama just saying that you need to time your code with/without parallelization. Predictions are difficult.

@YvesKlett ok

Lalalal al lala la lalalalaaa

@Rojo ah, yes, the wine...

@YvesKlett :)

@Rojo or are you training for the next match?

I run on wine. I am a Windows dude trying to run on Linux

@YvesKlett The next match I have interest in is tomorrow, Brazil-Chile

@Rojo lol

@YvesKlett But, speaking seriously, I had a moment of happiness while reading a piece of work of one of my students. It is goooood

@YvesKlett What about you, how are you doing, whatś up

@Rojo congrats! Does not happen too often here, as well.

@Rojo warming up for the weekend. 25°C, very agreeable.

Had some south-american filet for dinner. Yum

@YvesKlett Nice

I had potato in a microwave

@Rojo potatoes here as well! From the oven, though.

@Rojo why the wine setup?

@YvesKlett That was just a joke. I am new to linux, and Wine is a software that acts as an interface to run some native window apps in Linux

@Rojo my experience with wine was mixed. I wanted Picasa on Linux.

@YvesKlett Oh, nice. What OS do you usually use?

@Rojo Windows. Used to do Linux, but too much hassle at work.

The Picasa thing actually worked, but was quite slow...

@YvesKlett I see. So far I miss Cubase (audio). Also I hope for Mma to work well (I have some issues but I pray for v10 to have them fixed, but I don't have high hopes). And maybe OneNote (not for me, but I shared some stuff with others)

@YvesKlett and in my case it was the othe rway around. I worked with Windows among Linux users

@Rojo that might actually be more difficult in terms of being mobbed :D

@YvesKlett Hehe, right

all them geeks getting all snotty about Emacs and such ;-)

@YvesKlett Oh, they are not such geeks

None of them uses either emacs nor vim

ah... nerds, then.

but there's some Fedora vs Ubuntu

hehe the rpm wars

Exactly

And gnome vs some other,

KDE

@YvesKlett And I now use i3 so, screw all of them

and I started with Fedora. Just by chance, I still now don't know what can be the big difference between Fedora and any other

nor do I. I mostly did SuSE (OpenSuSE later on).

@YvesKlett Mathematica is better than Wolfram Desktop :P

(let's start a war)

I cannot be bothered until Version 10 is out.

@Rojo signing off for tonight - have a great day!

@YvesKlett Thanks, good night!

It looks like we are getting lots of people who haven't programmed before asking questions

(I mean this literally, not as an insult)

@acl I guess you never seen the Matlab forum before? :) This is nothing new there.

This just means Mathematica is getting more popular!

@Nasser probably

@MichaelE2 when Min[x]works in a compile (as in mathematica.stackexchange.com/a/50799/11116), why doesn't it work in a = 1; Vf = 100; cap = 2500;p1 = RandomReal[10, {10}];p2 = RandomReal[10, {10}]; demand1 = Compile[{{a, _Real, 0}, {p1, _Real, 1}, {p2, _Real, 1}}, Min[Transpose[{p1 Vf - p2, a cap - p2}]], Parallelization -> True, RuntimeAttributes -> {Listable}, CompilationOptions -> {"InlineExternalDefinitions" -> True}]; demand1[a, p1, p2]

holy crap ipython notebooks are amazing. Any mathematica users use this much? Any heads up for shortcomings from a mathematica users perspective ... I am blown away how far the python science stack has grown in the last 5 years ... been to long for me

@brama It works, why shouldn't it? Do you get an error message?

@brama Works for me (V9.0.1), too. (Transpose is unnecessary.)

@brama Can one run python code in the cloud directly without one downloading or installing anything on the PC? like an applet for example.

@Nasser I think this project does that: pythonanywhere.com

also it seems that ipython has a feature to do this ... though I am in the process of learning this, but in a pycon talk it was mentioned you could host ipython (interactive ones) from amazon cloud services

@Gabriel I've actually seen this site before, and not able to find single example to run!

@Nasser lol! Yeah I am trying to get my head around what the current features of python are in this as well ... a bit complicated to understand for me at the moment ... but damn I am loving what I have gotten working ... the latex/markdown notebooks in ipython are just amazing

I am not looking for a place to login and run python code myself. I am talking about having an python program with GUI allready deployed in the cloud to run myself and see how it works, as a user.

yeah I get that: wakari.io/nb/url///wakari.io/static/notebooks/…

this one is also cool showing the markup I was mentioning

wakari.io/nb/url///wakari.io/static/notebooks/…

@Gabriel yes, but this is not what I mean. There are many such things on the net now. You can actually edit/compile and run many thing on the web. What I mean is a deployed app to run as user. Like Java applets or CDF's, but with no additional downloads. Like the Cloud is supposed to do. I do not think Python can do this.

You can edit/compile Latex on the web. If you google "compile run languages on the cloud" you'll see many sites that does this for many languages.

you would mean like a demonstration? As for the links I have given they are live if you login and work in the way I have used wolfram cloud so far ... meaning you can edit run the code with no plugin

my understanding is that the interactive/manipulate like features are at a more basic level at the moment

@Gabriel yes, I mean an actual app I can run myself as user. like Java applets, etc... with no user having to download anything.

the cloud stuff I linked to is this

you don't need to download anything it all runs natively in the browser as html/javascript

@Gabriel That is not a GUI application. That is just a notebook that one can run. I mean a GUI application. This one also requires one to login to run, etc... and have to be on that site server.

But any way, Python is nice for OO programming.

@Nasser yeah I think you are right. Looks like a secure distribution of the interactive gui stuff is the development road for ipython 3.0

Yeah I loved python as a language ... but have only recently looked back at it for science as I am becoming increasing unhappy with the direction mathematica is going for the kind of work I do ... but I hope to be wrong! I just am really dissapointed by the every shrinking Dataset stuff and am looking at python Pandas as a way to sooth my broken heart ;) Though we shall see if it is anything more than a fling

@Nasser the gui support in the web is super basic at the moment (vs the local notebook server): see nbviewer.ipython.org/gist/certik/7854264

@Gabriel Yes, this is more what I am thinking about. Thanks. I see a plot with a slider and it works in the browser without me downloading or doing anything on my end. When Python gets all the other features in Manipulate, I'll give it a second look :)

:)

totally looks like that will take a long time as they need to port there plotting stack to d3 or something like that ... though I never thought they would get as far as they have when I left ... so we shall see if they become competitive in this space

This is cool: mpld3.github.io/examples/custom_plugin.html

mpld3.github.io/examples/linked_brush.html

maybe it won't take as long as I think

wow ... the zoom feature is sweet as well

okay last post but this is a great example with that library and a notebook:

mpld3.github.io/_downloads/mpld3_demo.html

For fun, just wrote this, took all day. Satellite spin control.

select the spin axes, the select which other axes to perturbe . If it spins on the axes with largest moment of inertia, will stay stable, else will tumble.