Computational Methods in Finance, Fall 2000

      Comments on homework 1
      
READER FRIENDLY CODE. In the workplace, you will not be coding just for
yourself, but in collaboration with a group.  It is crucial that your
code be readable, that someone can read over your code and understand
it easily.  Your homeworks will be graded for readability.  It takes some
planning to write readable code.  Moreover, readable code is more likely
to be correct.  The person who misunderstands "spagetti code" is likely
to be it's author.

  Make good use of indentation, white space, and comments.  
  
  Avoid variable names like "temp1" or "aaa" (unless it's a bond rating).  
  
  Follow conventions, such as all caps for a #defined  variable.  
  
  Avoid very long names, such as "NumberOfWeeksRemainingForTheLoan",
  especially when they are involved in arithmetic expressions or procedure
  calls.  
  
  Avoid unnecessary pointer use and dynamic memory allocation.  One program
  had the expression "double ****f;", and without a comment.  That is,
  f points to a pointer that points to a pointer that points to a pointer 
  . . ..   Think of a simpler looking way to do this.
  
  Use intermediate quantities to break up large expressions.  If a single
  line of code is more than 2 lines long, you can almost certainly make
  it easier to understand by breaking it up.
  
  
BEWARE THE SUBLTIES OF COMPUTER ARITHMETIC.  Floating point arithmetic
is unlikely to be done exactly.  If you use a floating point computation
to get an integer (such as the number of possible interest rates), the
result is likely to be off by one, as in this example:

#include <stdio.h>

/* A test code to explore floating point arithmetic.  */

int main() {

   double x, y, z;
   int n, m;
   x = 1./5.;
   y = x*x;
   z = x/y;
   n = (int) x/y;  /* This computes (int) x then divides by y. */
   m = (int) z;    /* This computes (int) (x/y), as desired.   */
   printf("An illustration of floating point arithmetic");
   printf(" and truncation to integer.\n");
   printf("z = %8.4f , n = %2d , m = %2d\n", z, n, m);
   x = .99;
   n = (int) x;
   printf("The truncation of x = %8.4f is n = %2d.\n", x, n);
   return 0; /* A signal that nothing went wrong.  */
   }

Which gave output:

   An illustration of floating point arithmetic and truncation to integer.
   z =   5.0000 , n =  0 , m =  4
   The truncation of x =   0.9900 is n =  0.
   
I strongly recommend that you avoid serious scientific computing in 
VBA or Matlab precisely because the types of variables are uncertain
and you have no control over type conversions.  

BE CREATIVE AND SPEND SOME TIME FIGURING OUT HOW TO VALIDATE YOUR CODE
AND ITS ANSWERS.  This week's homework included many easily avoided
wrong answers.