30 May 2010   8:50:19.111 AM
 
CONSTANT_TYPE
  FORTRAN90 version.
 
  Demonstrate that constants have a type.
 
  EXAMPLE 1:
 
  Print the value of
    1.0203040506070809 - 1.020304050607
 
     1.0203040506070809 - 1.020304050607 =    0.0000000    
 
  If single precision is the default, this will be 0.
 
  EXAMPLE 2:
  A constant has a type.
  If you do not specify one, one will be provided.
  If we assign a 32 decimal constant to a single,
  double, and quadruple precision variable,
  IN EVERY CASE, only the first 8 decimals will be used,
  because we did not set the type of the constant:
 
    a = 1.020304050607080910111213141516
    b = 1.020304050607080910111213141516
    c = 1.020304050607080910111213141516
 
   A =    1.0203041    
   B =    1.0203040838241577     
   C =    1.0203040838241577148437500000000000      
 
  EXAMPLE 3:
  Use an exponent marker of "E", "D" or "Q" to indicate
  that your constant is single, double or quadruple
  precision.
 
  If the compiler does not like our specification,
  the constant is treated as single precision, and
  some of the information in the constant is discarded,
  with NO WARNING.
 
    a = 1.020304050607080910111213141516E+00
    b = 1.020304050607080910111213141516D+00
    c = 1.020304050607080910111213141516Q+00
 
   A =    1.0203041    
   B =    1.0203040506070808     
   C =    1.0203040838241577148437500000000000      
 
  EXAMPLE 4:
  Use an kind marker of "_4", "_8" or "_16" to indicate
  that your constant is single, double or quadruple
  precision.  Might work, might not.
 
    a = 1.020304050607080910111213141516_4
    b = 1.020304050607080910111213141516_8
    c = 1.020304050607080910111213141516_16
 
   A =    1.0203041    
   B =    1.0203040506070808     
   C =    1.0203040506070809101112131415159898      
 
CONSTANT_TYPE
  Normal end of execution.
 
30 May 2010   8:50:19.113 AM