There is a line in the concluding paragraph of the paper: "... and we completely trusted the results by ...". This mindset produces the ranting about CASs making mistakes. Do they? Of course they do. Do I make mistakes? (Well, never mind...)

The point is not to be lenient; bugs can and should be exorcised even if occasionally at great expense. But as a user of any tool it my responsibility to verify it is (a) the appropriate tool and (b) the results it produces are correct in the context of the problem posed. And, incidentally, running the same algorithm through two different CAS, while maybe useful, is not a proof of correctness, not by far.

No CAS can replace thinking. No program can be "trusted" if the results are mission-critical. In that sense the paper reveals more of the attitude of the authors than of CAS systems. And, are we really that surprised that their bug report did not have any discernable effect? To be fair here, these systems are complicated and implementing a fix may not be a trivial matter in this case. (Added in edit: Somewhere I read Wolfram said they fixed this one in Mma 10.<something>. But now apparently other, similar ones are showing up. Strange)

Just my $0.02

Mac Dude

Edit: Fixed up the English in places and a whole bunch of typos. Added comment about Wolfram supposedly fixing the bug in latest release of Mma 10.

I am not an expert with sed, but I assume the semicolon trips things up.

But, why don't you read in your parameter file in script.mw? If you want to read in valid Maple assignments, you can use "parse()" on the string you just read in to execute it as a Maple statement.

M.D.

@Fairouz A cursory look in the Web would make me believe that you can embed a Flash movie or a Quicktime movie in a pdf, at least if you use recent versions. But you are on your own here as far as I am concerned, I have never done this.

M.D.

@andhaga I am not sure about your diff(V(y),y) term... y is defined  as y(t) above and Maple won't differentiate against a function without some specific incantations. Using a name (y) both as function and as variable is stringly discouraged, you are going to confuse yourself and maybe Maple also.

As far as correctness is concerned (besides the issue with y): neither your de nor the solution offered look like a pendulum equation to me. But there is truckloads of info on pendulum equations on the web, I'll leave it to yo to figure out what it should look like.

M.D.

Your problem starts at the line x:=0. This causes Maple to read x(t) to be 0(t) to be 0. You cannot use x as function and as a value at the same time.

if x:=0 is really meant to be the initial value x(0)=0, then ignore that for now; solve the general case and then insert the initial values.

If you wan to to it numerically, the IVs go into the dsolve(...,numeric) command. RTFM.

M.D.

The function is called Heaviside (capital H). Also, your Heaviside(0)-Heaviside(1) makes no sense; I guess what you want is Heaviside(x)*Heaviside(1-x); which is an impulse 1 unit long.

With these changes dsolve returns a solution but it cannot evaluate it. I don't know what is wrong, and I don't have time to trouble-shoot it, my apologies.

Mac Dude.

Is it even clear there is a solution? Faced with a situation like yours I suggest you spend the time understanding what your system is about. It appears you are mixing linear and multilinear terms, with some complex constants thrown in. You can, e.g., try to solve subsets of increasing complexity and see what you get. You can try to solve for some variables if they appear to show up in only a few eqns. and thereby reduce the number of variables in the final solution attempt. Even better if you can break down the whole problem into self-contained pieces you can solve individually.

Just throwing a large system at any CAS is (in my experience) not usually a promising strategy even if, at times, one gets lucky. Of course, you can always run the thing overnight or a weekend... again, I have not had much luck with that approach.

Mac Dude.

@ Well, I think it is time to make that phone call. I don't want to play with my own installation to verify it, but I think soap.dat is needed for a single-user installation (which is what I have and presumably you as well). I assume customer service can verify that you have a "legal" license and maybe let you reinstall or help you in other ways.

M.D.

@Arilson Mar Can you describe a specific problem to tackle?

M.D.

@ I checked my own installation. The licensing files appear to reside in

/Library/Frameworks/Maple.framework/Versions/15/license/

(and you notice that here I have a relatively old version). In this directory I have two relevant files (plus two text files):license.dat and soap.dat. License .dat is installed by the installer whereas soapfile appears not to (judging by the installer log). You may want to check whether you have both of these. License.dat sems to be mostly a list of packages etc. that are included. Soapfile is an xml file that amongst others has the "purchase code" (the code you have to type in upon activation) as well as another number that may encode the host you activated on. Both of these are readable with a regular editor (Emacs, TextWrangler or similar).

If license.dat is not there I assume you'll need to get it back from your backup. You did backup before the system upgrade, right? In fact, you could even try to restore the whole license directory.

If that does not work; then maybe something more is bollixed. You could try to reinstall Maple (assuming that your license allows for more activations).

Good luck,

Mac Dude.

@trigo1 The addition of two (or more) sinusoidal functions with different frequencies creates terms with sum and difference frequencies (often called beat frequencies). So if you have omega1 and omega2 you'll get in general 4 frequencies (the originals and the beats), each one having its own amplitude (related to A and B).

Maple's facilities to ferret that out are not strong or intuitive. Playing around a little with your formula it seems that working with the exponential representation (convert(expr,exp)) allows for more strength in manipulations using Maple's facilities. If you want to stay with sines and cosines you may have to roll your own sum and difference rules although I seem to recall there are some packages that help this along.

If A=B, the original frequencies vanish and you are left with the two beats (trivially shown by sum and difference rules for sine functions).

This link has a number of Maple worksheets possibly useful in this context:

http://www.peterstone.name/Maplepgs/trigfcns.html

So, it can probably be done, but you are unlikely to end up with a simple, pleasing answer.

Mac Dude

@birdiestreaming Just to be clear: you HAVE to adhere to the rules of Maple if you want reliable results. You said that just fixing the spelling of Pi gave you a result. But how do you know it is the correct result??? There are a good number of cases where Maple will give you a result, but not the one applicable because the problem was posed in the wrong way.

Caveat Computor! There is little or no room for sloppyness here, esp. when you are just starting out with Maple.

Mac Dude

I would use the Physics:-Vectors package to tackle this. See the docs on the physics package and its various subpackages (here esp. the Vectors subpackage). Once you spell "Nabla" correctly (with a "b" before the "l") it'll know what it means and knows how to do grad, div and curl.

Mac Dude

I don't think there is a built-in checkpointing procedure in Maple (checkpointing = writing state of a calculation to disk so you can resume from that point later). Assuming you have a kind of loop, you can save intermediate results; this thread may be helpful:

http://www.mapleprimes.com/questions/202512-How-To-Save-The-Session-Of-Maple-Worksheet#answer209792

You could e.g. wrap the save-state code into a proc & call that periodically.

If it is the intermediate state of one operation you want to save; I don't know how to do that.

M.D.

@rit With a shared kernel, ALL variables defined at the Maple prompt ARE global to both sheets (I just checked, as I do ot usually run in that mode).

Mac Dude