I would suggest to add an output algorithm for matrices and vectors to the file nr3.h.


const int os_precision_m = 7;
const int os_width_m = 5;

std::ostream& operator<<(std::ostream &os, const MatDoub &m){
os << std::setprecision(os_precision_m);
for (int i=0;i<m.nrows();i++){
for (int j=0;j<m.ncols();j++){
os << std::setw(os_width_m) << m[i][j];
}
std::cout << std::endl;
}
return os;
}


Than you can write a matrix to standard output with


#include "nr3.h"
int main(){
MatDoub a(2,2);

a[0][0] = 1; a[0][1] = -1;
a[1][0] = 0; a[1][1] = 2;

cout << "a" << endl << a << endl;
}


What do you think about that idea?

Yes, good idea. We were intentionally trying to keep nr3.h and its classes as minimal as possible, so as to be syntactically compatible with the maximum number of other vector-matrix class libraries out there. However, my sense now is that most NR3 users are using nr3.h, not something else. I wonder if it would be useful, now, to have a superset "nr3a.h" with a richer set of capabilities. For example, it would also be easy to overload the various arithmetic operators and provide "c=a+b" (etc.) capabilities for vectors and matrices.

Users, please give your views on this!

Cheers,
Bill P.