is it ok to ask question on the main board about the CDF plugin in the browser?

I want to ask how to check if it is installed. It is really a javascript question, but I assume it has to use some CDF plugin API

But I want to find out BEFORE I do the cdf.embed() call.

@Nasser Well, there has been a precedent

So I would risk it:)

@Ajasja, thanks! will look at it. I need to actually do 2 things: find if the cdf plugin is installed,and also find the dimensions on the fly.

@Nasser I was refering to the fact that this question is basically JavaScript only as well, but was not deemed off topic

So I would guess it's at least worth asking.

Well, gotta go (this was just a drive-by visit:)

I want to ask a question on the main board about how to make M looks better on Linux. The fonts and all my UI's build on windows are all broken when I open them on Linux using version 9. I wonder if this will be ok to do or people will think it is not a good question to ask.

@Nasser: why not use the function from cdfplugin.js, i.e., : var pluginIsPresent =
  function () {if (typeof ActiveXObject != ' undefined') {
       //
        IE compatible browsers try {var control =
           new ActiveXObject ("Mathematica.Control");

          if (control) {return true;}} catch (e) {}} else if \
(navigator.plugins && navigator.plugins.length > 0) {
       //
        Mozilla and WebKit browsers for (var i = 0;
          i < navigator.plugins.length;
          i++) {if (navigator.plugins[

@RolfMertig, I just copied it and will try it, Thanks. I still need to find the dimensions, and I think the other link above shows how to do that part. (I am not good in Javascript)

@Nasser What's wrong with M on Linux? Maybe it depends which Linux you use. I prefer Fedora these days.

@RolfMertig, it is BAD, real BAD. I moved to Linux, but I can't stand M there. The fonts are BAD. All the demos I did on windows are messed up. sliders wrong sizes, things are off the edges, etc.. I can show screen shots, but I could find if it s a font problem or what.

I use Linux mint 14, 64 bit.

I am using a TRIAL version of M, since I am only allowed one licence and I could not install my windows M. So downloaded a trial one for 30 days to check linux first.

I might have to keep my windows PC, just to use M there.

and do everything else on Linux. Will use a USB stick to move files around

I just do not know if the question will be ok on the main board.

I was fraid someone will get ubset

@Nasser: which one? XFCE? I used Mint in the past, but did not like it since I could not upgrade between releases (they used to recommend to reinstall everything, which I don't like). Why don't you try spins.fedoraproject.org/scientific-kde ?

@RolfMertig, yes! XFCE

I'll show screen shot in a minute as an example

W|A runs on Linux. So it is possible to do things right ...

@Nasser can I get the source of your demo? I can have a look later.

Here is one from a demo I submited sometime ago. same code. I just opened it in Linux. Look at he left panel, how it is all messed up. all fonts are wrong size. I'll post same exact code but on windows now.

Please select the SOLVER option to open the same screen I just uploaded above so you can compare.

Here is another screen shot (I am on linux), the top image is from windows, and the bottom image is same but on Linux). see how much different the UI looks

Size of UI objects is the problem. Since size of fonts and controls are different, then all the sizing I did on windows and spend so much time to get right, now is all broken.

Does anyone know what is the default MatrixPlot ColorFunction?

@rm-rf Re the Copy Hyperlink crash. I fixed that problem in 9.0.1 (coming real soon now). Also fixed a problem in Copy Address (embarrassing that we somehow broke both...but coincidental...two separate problems).

@Nasser One problem seems to be that Linux substitutes "Courier" for "Courier New", which is not the same. Theoretically "Liberation Mono" should be the right replacement, but I tried and it improved a bit, but there are other more serious problem with the control layout. I have no time to look into this. Please contact Wolfram Support.

@JohnFultz Did you also fix that strange message one gets from SetOptions[Cell, CellFunction :> f] (* i.e.,: An improperly formatted directive with head Symbol was encountered. *) ?

@RolfMertig Apparently not. I don't recall having seen a bug report on that.

@JohnFultz Hm. Maybe I did not send one. Not that it is very important (I posted an answer on the WolframCommunity and it is mentioned as a comment in that answer).

John and Rolf, if you're not too busy: is there a faster method to keep track of a fixed number of values by dropping the last one and adding a new one? As we know Append is slow on long lists but I cannot think of a better way. I mean something roughly like:

big = Range@1*^8;

Do[
  big = Append[Rest@big, RandomInteger@99],
  {10}
] // AbsoluteTiming

Is there some trick to avoid reallocation?

@Mr.Wizard Why not Do[big[[-1]] = RandomInteger@99, {10}] // AbsoluteTiming ?

@RolfMertig Well, this is slightly faster but still not quick:

big = Range@1*^8;

Do[
  big[[;; -2]] = Rest @ big;
  big[[-1]] = RandomInteger@99,
  {10}
] // AbsoluteTiming

Actually about a third faster so it's a start. Are you thinking of a different formulation?

To be clear I want a circular buffer of sorts, not an accumulating list, or I'd just use Internal`Bag.

@Mr.Wizard I remember that Roman once described linked list implementation in Mathematica. Somewhere a long time ago.

That's usually done with something like {1, {2, {3, {}}}} -- I'm not sure how I'd use that here.

@JohnFultz do you have any comment on this?

@JohnFultz sorry, I meant of course SetOptions[Cell, CellEpilog:>fun]

@RolfMertig This would do the same thing, but without the message:

CurrentValue[$DefaultFrontEnd, CellEpilog] = fun

@JohnFultz Great! I did try with $FrontEndSession, but not $DefaultFrontEnd

What is $DefaultFrontEnd?

@RolfMertig Actually...there seems to be some issue with CurrentValueand $DefaultFrontEnd. Maybe better to use SetOptions.

@JohnFultz Do you want to write a bug report to [email protected] ?

Options work on an inheritance model. You've seen global options inherited into notebooks, notebooks inherited into cells, etc.

It all starts someplace. That would be the hard-coded option values in the front end C code. The root of the inheritance of options.

Then one typically thinks of $FrontEnd as being the next level down, but there's another level in between. That is $DefaultFrontEnd.

So, at the global level, it's <hard-coded values> -> $DefaultFrontEnd -> $FrontEnd -> $FrontEndSession

But only values set to $FrontEnd get written to your preferences file.

@Mr.Wizard alas, I don't have a deep understanding of how the (in)efficencies of the kernel expression structures, so I'm probably no help here.

@JohnFultz I would appreciate your asking around if there is some Internal` ring buffer function. I expect I'm not the first to want it.

@JohnFultz Why is there a $MessagePrePrint and no $MessagePost? I.e., I could have answered Bretts question without resorting to CellEpilog if there would be a way to inject an action after a message has been printed. That would be also useful in general for debuggin, I guess.

@Mr.Wizard Leonid should know about Internal` and linked list and so on ...

@RolfMertig This sounds interesting. To which question do you refer?

@JohnFultz Why is there a $MessagePrePrint and no $MessagePost? I.e., I could have answered Bretts question without resorting to CellEpilog if there would be a way to inject an action after a message has been printed. That would be also useful in general for debugging, I guess.

@Mr.Wizard at the Wolfram Community someone wanted a OnDialog[Power:infy] function which popped up a Dialog (Notebook?) when a Message shows up. Can be done with CellEpilog. I really like CellEpilog.

@RolfMertig FWIW, I haven't really looked at SetOptions on things like Cell and Notebook in a long time. Perhaps they really should be setting on $FrontEndSession instead. But the current behavior seems to use $DefaultFrontEnd

As to why there's no $MessagePost, I guess it was just never needed.

@JohnFultz Thanks :) Will wait for 9.0.1

So, why am I on chat today? Because I'm spending the day hacking away in Excel (yes, really), and I'm looking for... Pity? Sanity? Escape? Well, anything that delays this little bit of torment, anyway.

@JohnFultz Poor you. I once scripted Excel through NETLink (even executed macros). Long time ago.

@rm-rf You dug that out of MatrixPlot? Thanks, I'll just use it then :) But Can you post the cut off part of the expression too?

@Szabolcs Yes, dug it out of MatrixPlot... Just follow the link; it's hidden under "Show more"

OH yes, I'm stupid today ...

in light of the comments here and votes on the question and comments, can we get some NaRQ close votes?

@JohnFultz It has been nearly a decade since I had to do that. I can empathize.

@rm-rf started it.

On["Packing"] makes the prediction bar ugly.

I tried using this as a workaround:

$Pre = Function[private`code, Internal`WithLocalSettings[On["Packing"], private`code, Off["Packing"]], {HoldAllComplete}]

Hopefully this should isolate the packing messages to the code being explicitly evaluated by the user.

@Szabolcs Did you find anything interesting in the answers to this question?

Perhaps the accepted answer needs to be cleaned up a bit? (The answer is in the Addendum and perhaps the saving to a CellTag, while useful, is a bit distracting). Any comments appreciated...

@Ajasja Could you please remind me in a few days if I forget? I'm very interested in this, but unfortunately I have a lot of work for the next 5 days.

@sjoerdc.devries could you join me in here?

Ok, I gotta leave now, but I've reopened the question, so feel free to answer.