Historically, in earier Maple versions, subs was used for such calculations instead of eval. Using eval for that is relatively recent innovation. In many cases, eval=subs, just with changing the order of arguments. The eval help page doesn't contain corresponding examples though.

Also, showstat can be used for viewing the code,

showstat(`eval/diff`);

The result of substituting x=H and y=H can be predicted from the following,

diff(v(H,H),H,H);

       D[1, 1](v)(H, H) + 2 D[1, 2](v)(H, H) + D[2, 2](v)(H, H)

That's just the way Maple works. If you do eval without the quotes, Maple changes the variable from z to H, doing exactly what it was asked to do. It is nothing wrong with Maple here. Another way of dealing with that is changing order of assignments - i.e. assign "a" in my example after assigning "v",

v:=z->z^H:
a:=eval(diff(v(x),x),x=H);
                                     H
                               a := H

If using quotes is impossible, that just means that the algorithm should be revised making that possible. The problem is not in Maple - the problem is in using eval in incorrect way.

This is a result of so called "premature evaluation". It can be avoided by adding unevaluating quotes,

a:='eval(diff(v(x),x),x=H)';
                              /d      \|
                         a := |-- v(x)||
                              \dx     /|x = H
v:=z->z^H:
a;
                                   H
                                  H

Also, it is not the best way of doing that. It is better to use D,

restart;
a:=D(v)(H):
v:=z->z^H:
a;

                                   H
                                  H

Nice pictures. Both look like Physics in Art rather than Math though.

That also can be done as

isort:=L->ListTools:-Transpose(sort(ListTools:-Enumerate(L),(a,b)->a[2]<=b[2]))[]:

Indices,Values:=isort([4,5,3]);

               Indices, Values := [3, 1, 2], [3, 4, 5]

The following 2-line procedure is slightly faster,

Isort:=proc(L) local a;
a:=sort([$1..nops(L)],(i,j)->L[i]<=L[j]); [seq(L[i],i=a)],a end:

Values,Indices:=Isort([4,5,3]);

               Values, Indices := [3, 4, 5], [3, 1, 2]

Another problem,

icoeff3((a+b)*(c+d));
                                  4

Another problem,

icoeff3((a+b)*(c+d));
                                  4

I should use procname in it instead of icoeff4 :)

I should use procname in it instead of icoeff4 :)

That should work OK. Also, for efficiency reasons, subsindets[flat] can be used instead of subsindets.

That should work OK. Also, for efficiency reasons, subsindets[flat] can be used instead of subsindets.

That has several problems, too. In particular,

icoeff2(3);
                                  1

That has several problems, too. In particular,

icoeff2(3);
                                  1