have you seen www.fortran.com?

It was thus said that the Great Chris M once stated: Nope. There are certain optimizations you can do in Fortran that you can't do in C because of the language semantics. Given the hypothetical code: int y; int foo(int *px) { int a; int b; int c; a = y; *px = a * 2; c = y; ... } The compiler is hampered with this code. It would be much faster if the compiler could do: move.w (y),d0 move.w d0,a(sp) move.w d0,c(sp) add.w d0,d0 move.w d0,[px(sp)] (and if my syntax is a bit off, please forgive me, I don't have my 68k references handy right now) But the C compiler can't do that because it can't know if px points to y. So it has to be rather conservative in code generation: move.w (y),d0 move.w d0,a(sp) add.w d0,d0 move.w do,[px(sp)] move.w (y),d0 move.w d0,c(sp) Not much difference, but when you're writing code that takes a year to run, it does add up. Fortran also supports complex numbers. It may even support vector operations like (borrowing a C like-syntax): double A[MAXSIZE]; double B[MAXSIZE]; double C[MAXSIZE]; C = A + B; which does the equivilent of: for (i = 0 ; i < MAXSIZE ; i++) C[i] = A[i] + B[i]; or for (pa = A , pb = B , pc = C , i = 0 ; i < MAXSIZE ; i++) *pc++ = *pa++ + *pb++; or whatever is faster for the CPU architecture. I do know that the newer C standard (C99) has attempted to address some of these issues, but not all of them. -spc (Now? Where is my Numerical Recipies book again?)