@casperyc It can be made to plot either all the points, or just the later ones, sure. It's ok to have A be local, but of course for efficiency Array A must be created just once outside the loop. There are more efficient ways to build the PLOT structure, yes -- more on that later. And it is possible to have the procedure emit the PlotComponent (right below) any eventual output. And distinct calls to your computation routine could emit and use their own distinct PlotComponents. The return value could also include the final plot, sure.

It'll take me a while to construct a more thorough example with these ideas. Perhaps later this evening...

@rightClick I see. If I try it on the output of the following command then I get the problematic behaviour you described if I have typesetting level set to "standard". But it seems to work ok if instead I have that level set to "extended".

`𝓍__2`(s) = 1/(s^2+3*s+2);

There are at least two ways to set the typesetting level to the extended mode. One way is to use the menubar choice Tools->Options-Display and change the dropmenu "Typesetting level" from being "Maple Standard" to being "Extended". Another way is with the command,

interface(typesetting=extended):

I will submit a bug report against this problem.

It works for me, in Maple 18.01 on Windows 7 64bit, using both 2D Math and 1D Maple Notation modes for input as well as using both "standard" and "extended" typesetting modes.

I used x[2](s) as the left-hand side of the equation, which I would have expected not to matter here. Your italic x seems larger. Just for interest's sake, what did you enter it as?

acer

@Alejandro Jakubi Ok, it is sold by Maplesoft (and I edited my comment accordingly). It's still a bug in the IDE add-on if it claims that valid Maple syntax is invalid. Do you disagree?

Note that I have at no point claimed that such add-ons are de facto at fault if their behavior goes against that of the Maple product proper. In this particular case it seems to me that the described behavior of Maple is fine, and the syntax is valid.

The one accepted by Maple itself is correct.

The Maple IDE is an add-on product. This appears to be a bug in the add-on product.

acer

@Joe Riel It would be great to see such a useful package be included directly in the next major releases of Maple.

@jzivku I downloaded your linked file BioCalc.zip and unzipped. This produced a directory named BioCalc, which contained another zip file named Biocalc.zip.

Now, in Maple 18.01 on 64bit Linux, I get,

questions:=MapleTA:-Import(cat(kernelopts(homedir),"/Downloads/BioCalc/Biocalc.zip"));
Error, (in MapleTA:-Import) module does not export `questions`

What should I being doing instead, to access the questions directly in Maple 18.01 using the Grading:-Quiz command?

@Carl Love These approaches of conversion to seq (or list, or listlist then flattening) following explicit Matrix transposition can break down for large size Matrices. For example, with a 1000x1000 Matrix M with entries between 0 and 9 say, the system may take much more time and possibly even hit stack limits.

Using a single sprintf call with tailored rtable (Matrix, Vector, etc) formatting such as printf("%{ns}ld\n",M) holds up much better at larger sizes. It can also walk entries by column with a minor modification, ie, printf("%{tns}ld\n",M) to act on the data columnwise (as if transposed).

@Carl Love There is a need for better handling of oscillatory integrands taken over a semi-infinite range, or an oscillatory integrand over a finite range which can be mapped to the former. Suggestions in the past have included something like the Levin-u transform for numeric summation, possibly taking the addends as the integral  over each cycle. A quick google search reveals this hit of a feature new in Mma 8. (I've tried splitting some examples with NextZero, then using evalf/Sum, but it needs a more serious effort I'm sure.)

Also, there are some examples of complex-valued integrals which can be easily split (manually) into real and imaginary components with evalc and then fed separately to fast methods such as _d01ajc. But the system doesn't seem to handle that automatically.

@nm Note that without square brackets around the arguments to Vector the first argument will specify its size. So Vector(91, blah) produces a Vector with 91 entries.

@Markiyan Hirnyk You forgot to restart between the two cases, so in your worksheet some computations are remembered and your second call is artifically appearing faster than it is.

Also, it's a mistake to re-execute with !!! rather than line by line, and to display the plots. On avg the second example is slightly faster (about 109ms versus 125 ms on my machine, over many separate runs). This difference might be due to the fact that the `map` case has an extra custom operator call ( as i->i*sin(i) ) for each entry's construction. And the `map` case produces two lists (one of which is temp garbage) rather than just one, so there is a memory used and alloc difference as well. But these effects might become negligible as the examples becomes much larger and involved.

restart;
with(CodeTools):
Usage(plot(map(i->i*sin(t),
           [seq(i, i = .1 .. 1, 0.2e-2)]), t = -2*Pi .. 2*Pi)):
memory used=131.28MiB, alloc change=41.86MiB, cpu time=1.90s, real time=2.09s, gc time=46.80ms

restart;
with(CodeTools):
Usage(plot([seq(i*sin(t), i = .1 .. 1, 0.2e-2)], t = -2*Pi .. 2*Pi)):
memory used=131.58MiB, alloc change=41.12MiB, cpu time=1.89s, real time=2.10s, gc time=46.80ms

The small size of the example makes the variation relatively huge, between runs. So without a larger run, or averaging, it's too easy to get an erroneous conclusion.

@Markiyan Hirnyk I ran your worksheet in Maple 18.01 and it displayed z not z~ as output.

@Preben Alsholm I wonder whether this kind of output -- or something similar -- might relate to why inttrans is still table-based. Perhaps there are recursion or efficiency issues. I'm just wondering out loud -- I don't know.

@Markiyan Hirnyk You cite only the first portion of a sentence, out of context. The whole sentence qualifies why loading is not wanted.

The full sentence by the OP was, "I do not like to load a package using with() and then use its commands and functions, since I then lose track knowing from which package a function or command being used in the code came from when I look at the code later on." So my suggestion to retain the explicit long indexed form in the input function call does directly address the stated need in the latter clause. The last sentence also links the desire to avoid package loading with retention of the fully qualified name, ie, "I really do not like loading packages as I said, and like to keep the name of the package attached to each function to help me know where each function is coming from."

Anyway, it does work for me,

restart;
with(inttrans,laplace):
f:=t->piecewise(t<0,0,t>=0 and t<z,t,t>z,z):
r:=convert(f(t),Heaviside):
r:=inttrans[laplace](r,t,s);

                                                                                 z
    r := laplace(t Heaviside(z - t), t, s) - z laplace(Heaviside(z - t), t, s) + -
                                                                                 s
eval(r,z=0.5);

            0.5000000000 (2. - 1. exp(-0.5000000000 s) (s + 2.))
            ----------------------------------------------------
                                     2.                         
                                    s                           

                                                   1.      
                 0.5 (1. - 1. exp(-0.5000000000 s))     0.5
               - ------------------------------------ + ---
                                  1.                     s 
                                 s                         

kernelopts(version);

          Maple 18.01, X86 64 WINDOWS, Mar 28 2014, Build ID 935137

@nm One difference is that the enclosure of the plots within a GUI Table affects the stretching differently w.r.t the vertical and horizontal directions. This is partially why I referenced my response to your earlier Question, since that mentioned some properties of the Table.

The help-page plot,options also describes the size=[w,h] option. It mentions an exception for w in the case of a plot array (which is what BodePlot returns).

I would say that the whole 2D plot aspect ratio business is much easier and sensible outside of a GUI Table (especially one with default re-sizing properties). I have been investigating a command that acts much like plots:-display(Array(...)) but offers up-front programmatic control over all the GUI Table properties - I'll try and add a note here if I find time to polish it up and post for feedback.