@John Fredsted 

You just run the executable file maplew.exe in the subdirectory bin.X86_64_WINDOWS (assuming that you are running the 64-bit version).

@John Fredsted 

On Linux I have got interesting information about some problems of the Maple Java processes using the VisualVM tools. As I see in the release notes, Windows 7 is supported, though I have not tried it on this platform. Also on Linux, launching the Standard GUI from the console, a large amount of informative warning and error messages are sent to this console when Java problems occur (quite frequently). You may try launching the GUI from the Windows command prompt. Perhaps you get something.

Rather than enjoyable to use, after these changes, I find that this site has deteriorated sharply (even more). I keep finding every day comments that have disappeared. Those that remain frequently are shown in upset order. Links for them do not work. The list of regression bugs is very long, and many promissed new features do not work. Frankly, I am very disappointed with the management of this site.

@ecterrab 

Clearly I mean GRTensor. Certainly, your answer appeared in between while I was writing mine.

@taro yamada 

The IntegrationTools package was introduced in Maple 10.

@John Fredsted 

Why they are not, and have never been, implemented as exports of the LinearAlgebra package is a good question. I think that the person that should answer it is another Maplesoft employee, Dave Hare, project lead of this package.

@nm 

There are many. Just for one example, Total Commander has German, English and French. And obviously there are many technical forums in other international languages like Spanish. Many persons do not know English, or not well. And it is rude for them leaving them out.

@Markiyan Hirnyk 

Where do you see Exact/table? Do you mean perhaps the local (private) procedure int:-Exact? In any case, int(tan(x)^(1/2)*sec(x)^2,x) is found by the elliptic integration algorithm (not a table), which is different to the algorithm being used in the case of Carl's workaround.

@Carl Love 

I do not get the same weirdness in both cases:

> int(tan(x)^(1/3)*sec(x)^2, x= a..b, AllSolutions);
                         (1/3)       2
               int(tan(x)      sec(x) , x = a .. b, AllSolutions)

> int(tan(x)^(1/3)*(1+tan(x)^2), x= a..b, AllSolutions);
                { (-infinity - infinity I) (a - b)        a < b
                {
                {                0                        b = a
                {
                { (-infinity - infinity I) (a - b)        b < a

Maple 17.02 on Linux 32-bit.

@taro yamada 

OK. Some details on the usage of dividend might help. With a single argument, the expression to be transformed, it divides its numerator and denominator by the numerator, and expands them. With three arguments as in the example above, the first argument is again the expression to be transformed, the second argument is the factor that will divide its numerator and denominator, and the third argument is a list of two procedures, applied respectively to the divided numerator and denominator.

@Joe Riel 

If lies are needed, then the problem is the design...

@Joe Riel 

Neither 1 is of type `*` nor 0 is of type `+` as the original expressions. They seem the result of a transformation in the kernel changing the type, equivalent to:

> `*`();
                                       1
> `+`();
                                       0

These cases of type change are not properly documented in ?select, thus generating confusion, in my opinion.

@stevens1111 

Classic means the Classic GUI, the traditional GUI. Later, for Maple 9, Maplesoft introduced a new GUI, that was branded the Standard GUI. As it was somewhat problematic, the tradicional GUI was kept and rebranded as "Classical". Actually, it was kept till now, except for the Mac platforms and only in a 32-bit version. Additionally, Maple has a third interface, a command line one (CLI).

Now, negative exponents are produced both with the Classic GUI and the CLI by default, and by the Standard GUI, when setting the extended typesetting mode, e.g. using the command interface(typesetting=extended). No need of hacks.

@nm 

I agree with you, this is a missing functionality. But there is a more basic problem to be addressed before: how to tell the region in the complex plane? That is, how to specify it? Consider e.g. the example of the linked thread: the annular region |z|>3. If you then go solving this inequality, you will hit the default interpretation stated in ?solve,ineq :

In general, variables and parameters will automatically be assumed to be real-valued in the context of one or more inequalities.

Then:

solve(abs(z)>3,z);
  RealRange(Open(3), infinity), RealRange(-infinity, Open(-3))

Really, a good standarized representation of regions in the complex plane (disks, curves, etc) is missing.

@Carl Love 

I think that this problem is an unavoidable consequence of the system architecture adopted for the current versions of Maple, namely a Java subsystem grafted upon the historical C subsystem. Then stopping correctly would involve stopping both subsytems simultaneously. Something most unlikely, I think. Moreover as the Java subsystem runs on a virtual machine (besides with possible high latency problems). 

In other words, solving this problem would involve coming back to a unified architecture. Something unlikely in the foreseable future. This is one of the reasons why I have moved to using the CLI through an IDE, whose button stops instantly.