The default for smartview is just the false way around: if one sees a "bad" plot one still could change to smartview = true. But the default hides the 'truth' in cases.

That should work fast on a medium sized machine, using Maple 18:

st:=time():
H22*x*(cos(epsilon*((1/2)*L-x)/R)+kappa*cosh(epsilon*((1/2)*L-x)/R)):
evalf(%): expand(%):
int(%, x=0 .. L); #simplify(%);
`seconds need` = time() - st;

                      seconds need = 0.577

@vv thank you

@vv 

Do you really need the limit? combine(J1+J2); eval(%, a=0); simplify(%); # plot(op(%)); should do.

Edited: https://de.wikipedia.org/wiki/Cauchyscher_Hauptwert#Substitution_i._Allg._nicht_erlaubt (sorry, it is in German) says that changing variables is crucial - would you mind to give a sketch?

I am not familiar with Canadian copyright and "Störerhaftung" = (Breach of Duty of Care ?), but understand concerns towards that. Usually modifications of original work are covered by CR.

It may help readers of such threads to leave a note "deleted due to <copyright>" or similar.

Those interested in details certainly may have a guess how to proceed without involving Maple Inc.

Hope that helps.

Besides the 'hardware=true' I sometimes got trapped ignoring branches of LambertW and usually I go like this:

fr=1.64*10^6*E^2*exp(-8.5/E); # avoiding indexed variables
convert(%, rational);         # matter of taste
RootOf(%, E); 
allvalues(%);

One can see 2 results, depending on an index. And sometimes one had to choose or find the correct index.

eliminate(eqs, x);
                                     2
                       [{x = a y}, {a  y - y}]

#or

solve(eqs,y, parametric);

I would like to know the suggested way, if you do not mind.

On a second thought: I guess you want to consider decimal fractions and dharr already had a good suggestion for that, http://www.mapleprimes.com/questions/209994-How-To-Get-Precision-Dynamically#comment225492

I said that I do not care for that (neither seeing a reason nor a demand, just  to have a short way)

But you can measure the length (if you are sure what you expect for something like 0.001 or 0.31e-22) or you can work with

SFloatMantissa(a)*10^SFloatExponent(a);

                       1111111111121213123131
                       -----------------------
                       10000000000000000000000

and use that

nice workaround!

"it is easy" (for me) obviously means "with some work". Like other parts for the initial task.

Maple should catch that, so I suggest to submit a SCR pointing to that weakness.

It is easy (and actually in my contrib) to see that the crucial term is
always negative, from which it then follows:

a^2*(a^2+4*k) = a^4+4*a^2*k < a^4+4*a^2*k + 4*k^2 = (a^2+2*k)^2, all positive

Now take sqrt and then get  sqrt(a^2*(a^2+4*k)) -a^2-2*k < 0

@mskalsi 

I had to re-install version 18 some weeks ago and had similar problems through the upgrade (on Win 7, Maple = 32 bit). For that the following helped:

• de-install again
• install the initial version 18.02 18.0
• through installation allow to check for updates

Then it worked for me.

In the classical interface the command interface(version) gives 18.0, but in the standard interface I get 18.02

Hope that helps

Aside the answers and Maple's needs in Math I always prefered the notation, especially in (multi-) linear Algebra, finding it a proper way to write some thing down. For example the canonical morphism to a bidual

 x |---> ( f |---> f(x) ) , M ---> M** = Hom(M*, R)