@hirnyk 

Certainly, Maple 13.02 yields those unevaluated limits. The results that I have mentioned were found by using Maple 14.

The difference arises because Maple 14 includes the new integration method FTOCMS, that evaluates limits by using MultiSeries:-limit, and it succeeds in this case.

@hirnyk 

Certainly, Maple 13.02 yields those unevaluated limits. The results that I have mentioned were found by using Maple 14.

The difference arises because Maple 14 includes the new integration method FTOCMS, that evaluates limits by using MultiSeries:-limit, and it succeeds in this case.

@hirnyk 

Probably it would be better to give assumptions by splitting the domain ksi^2-eta^2>0, like:

int(exp(-((ksi+I*eta)*x)^2),x=0..infinity) assuming ksi>eta,eta>0;
                                  1/2
                                Pi
                           ---------------
                           2 (ksi + eta I)

int(exp(-((ksi+I*eta)*x)^2),x=0..infinity) assuming ksi<eta,eta<0;

                                   1/2
                                 Pi
                          - ---------------
                            2 (ksi + eta I)

@hirnyk 

Probably it would be better to give assumptions by splitting the domain ksi^2-eta^2>0, like:

int(exp(-((ksi+I*eta)*x)^2),x=0..infinity) assuming ksi>eta,eta>0;
                                  1/2
                                Pi
                           ---------------
                           2 (ksi + eta I)

int(exp(-((ksi+I*eta)*x)^2),x=0..infinity) assuming ksi<eta,eta<0;

                                   1/2
                                 Pi
                          - ---------------
                            2 (ksi + eta I)

@bthur 

I find that there are some differences. On the one hand, Maplesoft employees do have blogs that call blogs, unlike  the rest of the users. On the other hand, the Employee Guidelines do pose on them some additional constraints. 

I would suggest, look if you keep a local backup...

@chiwhalee 

For instance:

J:= Int(Int( f(x)*exp(I*x*p),p=-infinity..infinity), x=-infinity..infinity):
applyrule(Int(A::algebraic*exp(I*x::algebraic*p),p = -infinity .. infinity)=
2*Pi*Dirac(x)*A, J);
value(%);
                     infinity
                    /
                   |
                   |          2 Pi Dirac(x) f(x) dx
                   |
                  /
                    -infinity


                              2 Pi f(0)

@JacquesC 

A save page complete, of the page as it stands, including heading and right column implies too much unnecessary stuff. Frequently, forum sites offer a printer-friendly html version, with the thread content but excluding the "decorations". Besides, I would appreciate a pure text version of the thread for maximum efficiency in local archive and search.

Also note this width problem.

Also note this width problem.

Chances are that you could use the equation

Dirac(p):convert(%,Int)=%;

             /        infinity                  \
             |       /                          |
             | 1    |                           |
         1/2 |----  |          exp(_k1 p I) d_k1| = Dirac(p)
             | Pi   |                           |
             |     /                            |
             \       -infinity                  /

in some simple cases by pattern matching techniques:

applyrule(Int(exp(_k1*p::algebraic*I),_k1 = -infinity .. infinity)=2*Pi*Dirac(p),
Int(exp(_k1*(a-b)*I),_k1 = -infinity .. infinity));

                          2 Pi delta(a - b)

As a side comment, during betatesting, I have asked the option to save whole threads in text form. Will has promised to add this feature in the future.

@Axel Vogt 

It was fixed in Maple 13.02. Perhaps, a version specific bug tag (like "bug Maple 12") could be more useful.

@Will

I would find extremely useful the ability of making searches after logical combinations of words, tags, dates, etc. Something like: find title integration and not body Bessel and tag bug and date before 2009, you get the idea.

This goes off topic, but I have to say that I disagree with the current voting system of voting to users. During betatesting I have asked the option of not being voted and not voting others, but I have been told that I cannot avoid being voted. Anyways, I will keep not voting anyone else.

I have asked also that the list of voters be public, but this request was also denied.

I would prefer open votes to articles, not individuals, with a brief explanation of the reason for the vote. Among other benefits, it would avoid the need of guesses or conspiracy interpretations.