C / C++ / MFC

#include <iostream>
template<class T> T TFun(T a, T b)
 {
   if (a > b)
     return(a);
   else
     return(b);
 }
float TFun(float, float);
//int TFun(int, int);
void main(void)
 {
 std::cout << "The TFunimum of 100 and 200 is " << 
     TFun(100, 200) << '\n';
 std::cout << "The TFunimum of 5.4321 and 1.2345 is " << 
     TFun(5.4321, 1.2345) << '\n';
 }

1. You are using iostream.h, this is a pre-standard header and should not be used. Instead use iostream, without a postfixed .h
2. You defined two functions manually, float max(float, float); and int max(int, int);. These two functions do not refer to the templated function and will be used before the match to the template function. Hence, you should remove these and the error will go away.
3. Lastly, have you considered what happens if you do something like this: short s = 37; int a = 2839; int b = max(s, a);? This will fail with an error message, because the compiler isn't able to deduce the type of T (your template parameter). Thus, you need to implement a more solid min/max function, as explained by Andrei Alexandrescu in his Generic< Programming >: Min and Max Redividus[^] article at http://www.cuj.com[^]
4. Also, this is just a small detail, but you might want to consider using '\n' rather than std::endl, because std::endl also explicitly flushes the stream - can slow down performance unnecessarily (albeit not a lot in most cases).
5. Lastly, the main signature is according to the standard either int main() or int main(int argc, char* argv[]). void main() is a Microsoft-specific extension, and unless necessary (I haven't yet seen any reason why) it should be avoided for the sake of portability.

#include <iostream>
<br />
template<typename T>
T max(T a, T b) {
  return a > b ? a : b;
}
<br />
int main() {
  std::cout << "The maximum of 100 and 200 is " 
            << max(100, 200) << '\n';

  std::cout << "The maximum of 5.4321 and 1.2345 is " 
            << max(5.4321, 1.2345) << '\n';

  return 0;
}