STO(i)INPUT(i)
RCL(i)VIEW(i)
STO +, –,⋅ ,⎟, (i)DSE(i)
RCL +, –,⋅ ,⎟, (i)ISG(i)
XEQ(i)SOLVE(i)
GTO(i)≥ FN d(i)
X<>(i)FN=(i)

Program Control with (i)

Since the contents of i can change each time a program runs — or even in different parts of the same program — a program instruction such as !1L2 can branch to a different label at different times. This maintains flexibility by leaving open (until the program runs) exactly which variable or program label will be needed. (See the first example below.)

Indirect addressing is very useful for counting and controlling loops. The variable i serves as an index, holding the address of the variable that contains the loop–control number for the functions DSE and ISG. (See the second example below.)

Example: Choosing Subroutines With (i).

The "Curve Fitting" program in chapter 16 uses indirect addressing to determine which model to use to compute estimated values for x and y. (Different subroutines compute x and y for the different models.) Notice that i is stored and then indirectly addressed in widely separated parts of the program.

The first four routines (S, L, E, P) of the program specify the curve–fitting model that will be used and assign a number (1, 2, 3, 4) to each of these models. This number is then stored during routine Z, the common entry point for all models:

' ! L

Routine Y uses i to call the appropriate subroutine (by model) to calculate the x– and y–estimates. Line Y0003 calls the subroutine to compute yˆ :

& %1L2

and line Y0008 calls a different subroutine to compute xˆ after i has been increased by 6:

13–22Programming Techniques