In this case it is equivalent to A[i,1..n]. It is a convention used in rtable (Vectors, Matrices) for accessing all the elements of a row/column. You can also access a partial row in the same way: A[1,2..-2]. The -1 means the last value of the index, -2 the previous to last index, etc. This doesn't necessarily work in Arrays, which permit actual negative indices.

Hi Vladimir, I apologize for saying you weren't paying attention. Actually, considering that it took me a few minutes to recall what I had previously posted, I wouldn't be surprised if you had forgotten it. Note that, as Jacques points out, the Maple Statistics package is not going to be the fastest gun around. That being said, there are ways to make things go faster, so we need to learn how to utilize them. By creating all the samples at once, you only have to pay the overhead of invoking the Sample procedure one time.

You weren't paying attention in my response to a previous post of yours. Go see very-slow-evaluating-matrix-computation. As an example, here is your test3 rewritten:

test3a :=proc(N,n)
local i,a,X,A;
    use Statistics in
        X := RandomVariable(Normal(0,1)):
        A := ArrayTools:-Reshape(Sample(X,n*N),N,n);
        for i from 1 to N do
            a := A[i,1..-1];
        end do:
    end use;
end proc:

This is an order of magnitude faster.

It works with the Linux release. The location of the Standard GUI initialization file file has changed, it is now ~/.maple/11/maplerc. I don't have Windows and cannot test that. I don't recall whether my previous preferences (from Maple 10) were updated when I migrated to Maple 11.

You might try command-line Maple. I believe it uses a different postscript driver. At least I got completely different results (inspecting the encapsulated postscript file) from the Standard gui and command line Maple. I haven't tried the Classic worksheet (Maple 11).

The with(plots) statement is not necessary; plot3d is not part of the plots package.

The with(plots) statement is not necessary; plot3d is not part of the plots package.

With what version of Maple are you having problems?

Explaining just how you did that might be of use to others. Here is how I'd go about it. To generate a list of the named members of a repository, in this case, the Maple 10 main repository, do:

lib := sprintf("%s/lib/maple.mla", kernelopts('mapledir')):
members := map2(op,1,march(list, lib)):
named := map(s -> s[1..-3], remove(s -> s[1..2]=":-", members)):

Rather than creating a huge list of the inert forms of every procedure, I'd do the following:

for nm in named do
    val := convert(nm,name);
    try
        inert := ToInert([eval(val)]);
        if hastype(inert, _Inert_EQUATION(_Inert_FUNCTION(_Inert_NAME("Normalizer"), anything)
                                          , _Inert_INTPOS(0)))
        then
            printf("%s\n",nm);
        end if;
    catch:
    end try;
end do:

That does print the names of the 13 procedures that Jacques mentioned; however, I could not get it to work in Maple 11, which generated stack errors. Since I have a shell utility for directly searching maple archives (I'll post it someday; it currently has problems with Maple 11 repositories), I can speed this up considerably by limiting the search to those members that contain a call to Normalizer:

members := ssystem(sprintf("mgrep -h Normalizer %s", lib))[2]:
members := StringTools:-Split(members)[1..-2]:
named := remove(s -> s[1..2]=":-", members):

With that, the search takes about one second, compared to a couple of minutes for the full search.

In the help browser, look up mtn. The help page states (correctly) that the Maple Technial Newlsetter is no longer published. In the search results window is a list of the numbers of most of the published issues (some from the first year appear to be missing). Clicking on any one brings up a page with the relevant abstracts for that issue.

While I haven't tried Krusader, I can quickly get to any directory using emacs. The problem I have is that from there I cannot launch a Maple worksheet that opens a new tab in the existing (already launched) Maple GUI. I could open a new Maple GUI, but prefer not to. That isn't the case with, say, firefox, where I can open an html file (or URI link) in a new tab in an existing firefox application.

Ya gotta type faster 8-). I don't mind multiple posts; it ensures cross-checking. This add/sum problem is coming up regularly, about every other day.

Ya gotta type faster 8-). I don't mind multiple posts; it ensures cross-checking. This add/sum problem is coming up regularly, about every other day.

I'm not quite sure what your question is (your secondary question about the shape confused me). That is, are you asking for a structure different from piecewise, in which to express the result? That is, would you be happy with

piecewise(a<>0,2,b<>0,1,c<>0,0,-infinity)

as the result? Or is it not the shape so much you are interested in but an ability of Maple to compute the result (in whatever form). Intuitively, piecewise seems a bit of a kludge.

A method---not sure how---to be able to easily open worksheets from multiple locations on the file system. The file-open menu remembers the last location, which isn't always convenient (but other times is) and is an annoyance when it is different from currentdir. Maybe this is already doable and I just haven't figured it out on my linux box. Is there a script that one can run that opens a worksheet in the existing (active) Standard gui? For example, I use firefox as my browser. From emacs (or mail) I can click on a URI link and a new tab in the existing firefox is opened with the appropriate URI. I'd like to be able to do something similar with Maple. Another possibility would be for the File-Open (and Save) menu to have a history mechanism that remembers the last few directories (not just worksheets) from which a worksheet was opened. That would solve a lot of problems. I have some rather deeply nested directory trees and it is a pain to continually traverse them using a mouse. Occasionally I'll insert symbolic links in my home directory to frequent locations, then I just have to click on the home icon and the symbolic link to get to the directory of choice, but this seems a bit hackish. Oops. I think this belongs in the wish-list. Sorry about that. Alas, I don't know how to move it and replicating it seems pointless.