If you mean the compound matrix as defined in planetmath.org/encyclopedia/RthAdjugate.html

there isn't, but it's not hard to construct a command for it.

> compoundMatrix:= proc(M:: Matrix, r::posint)
      uses LinearAlgebra;
      local m,n, S, T;
      m, n := Dimension(M);
      S:= combinat[choose](m,r);
      T:= combinat[choose](n,r);
      Matrix(binomial(m,r), binomial(n,r),  
        (i,j) -> Determinant(M[S[i],T[j]]))
   end proc;

The order Maple uses for the elements in a set is unpredictable (depending on  memory addresses).  One exception, I think, is small integers, which are always in increasing order (I'm not sure how small is "small" in this context). 

The formula for the roots of a cubic equation in terms of radicals tends to be rather complicated, especially if there are symbolic parameters involved.  Quartics are even worse, and for higher degrees there may be no formula in terms of radicals.  Once you're used to it, the RootOf notation is more convenient and useful than the radicals.

But if you do want the explicit roots to be used, try allvalues.  For example:

> int(1/(x^3+x+1),x);

sum(1/(3*_R^2+1)*ln(x-_R),_R = RootOf(_Z^3+_Z+1))

> allvalues(%);

1/(3*(1/12*(108+12*93^(1/2))^(1/3)-1/((108+12*93^(1/2))^(1/3))-1/2*I*3^(1/2)*(-1/6*(108+12*93^(1/2))^(1/3)-2/(108+12*93^(1/2))^(1/3)))^2+1)*ln(x-1/12*(108+12*93^(1/2))^(1/3)+1/((108+12*93^(1/2))^(1/3))+1/2*I*3^(1/2)*(-1/6*(108+12*93^(1/2))^(1/3)-2/(108+12*93^(1/2))^(1/3)))+1/(3*(-1/6*(108+12*93^(1/2))^(1/3)+2/(108+12*93^(1/2))^(1/3))^2+1)*ln(x+1/6*(108+12*93^(1/2))^(1/3)-2/(108+12*93^(1/2))^(1/3))+1/(3*(1/12*(108+12*93^(1/2))^(1/3)-1/((108+12*93^(1/2))^(1/3))+1/2*I*3^(1/2)*(-1/6*(108+12*93^(1/2))^(1/3)-2/(108+12*93^(1/2))^(1/3)))^2+1)*ln(x-1/12*(108+12*93^(1/2))^(1/3)+1/((108+12*93^(1/2))^(1/3))-1/2*I*3^(1/2)*(-1/6*(108+12*93^(1/2))^(1/3)-2/(108+12*93^(1/2))^(1/3)))

Searching strings is very fast, as SearchText is built-in.  AFAIK there is no built-in routine to search a sequence of digits for a given pattern.  One could be written in Maple, but that would probably slow it down by a couple of orders of magnitude.

Searching strings is very fast, as SearchText is built-in.  AFAIK there is no built-in routine to search a sequence of digits for a given pattern.  One could be written in Maple, but that would probably slow it down by a couple of orders of magnitude.

All map does is apply a given function to each member of (in this case) a set. 
What I'm doing is substituting random integers for the parameters other than ke, ce and s.

The lucky guess in this case was that the answer (or at least some matrix elements) would not depend on the parameters other than ke, ce and s.  If I had received an answer that involved some large numbers,  I would have concluded that my  guess was probably wrong, but I still might have obtained some useful information on the structure of the matrix entries as rational functions in ke, ce and s.  With further work I might have been able to figure out the coefficients in those factors.
 

All map does is apply a given function to each member of (in this case) a set. 
What I'm doing is substituting random integers for the parameters other than ke, ce and s.

The lucky guess in this case was that the answer (or at least some matrix elements) would not depend on the parameters other than ke, ce and s.  If I had received an answer that involved some large numbers,  I would have concluded that my  guess was probably wrong, but I still might have obtained some useful information on the structure of the matrix entries as rational functions in ke, ce and s.  With further work I might have been able to figure out the coefficients in those factors.
 

Looking at this with Digits set to 100, it seems the imaginary part of Q(x) is in fact nonzero (approximately 1.2e-15) when x is near the zero of Re(Q(x)).  Presumably the floats in the definition of Q(x) are only approximations, and the imaginary part would be 0 if the exact values were used instead.

Due to roundoff error, for x in the interval in question Q(x) as computed will have a small imaginary part.  I think that is what is giving fsolve trouble.

By the way, when I plot it the zero is near x=14.

> fsolve(Re(Q(x)), x=12 .. 16);

         14.05766281

> Q(%);
 

-.20e-11-.180810179e-12*I

The <maple> tag fails again: that should be

-.20e-11-.180810179e-12*I

There doesn't seem to be a delete button.  You could make your file private, so others won't have access to it.  Then if you really want to delete it, you
could click on Contact and send a message to the MaplePrimes support
team asking them to delete it. 

Oops, somehow I didn't see the "int". 

Well, Maple can't do this integral in closed form, so the result has a t in it, and implicitplot doesn't understand about dummy variables.  You could try this:
 

> with(plots):
   V:= (e,T) -> evalf(Int(erf(  (T-e*t)/sqrt(2)  )*t^2*exp(-t/2),t=0..infinity));
   implicitplot(V, 0..1, 0..10);

The result will be a straight line (you do see why that is, I hope).

And similarly, for Joe's example:

> match(5 = n + 3,t,'s');

            true

> s;

{n = 2}

Not that match is without its problems.

> match(5*cos(2*t)+sin(2*t)=c*sin(w*t)+d*cos(w*t),t,'s');

    true

> s;

{w = 2, d = 0, c = 6}

1) Since this question is not related to the previous one in this thread, it's a good idea to start a new thread.

2) It might be helpful if the question were a bit more specific.  Can you give us an example of the type of inequality you want to solve?

1) Since this question is not related to the previous one in this thread, it's a good idea to start a new thread.

2) It might be helpful if the question were a bit more specific.  Can you give us an example of the type of inequality you want to solve?

Maple is complaining that you have an "extra unknown" t in your input expression, and indeed you do: your expression involves the three variables e, t and T.  Maple is case-sensitive, so t and T are different variables as far as it is concerned.