Procedures — named blocks of action

A shop program prints a receipt after a sale, after a refund, and after an exchange.

You could copy the same six OUTPUT statements into all three places. It works. Then the shop adds a VAT line.

Now you have three copies to find and edit, and three chances to miss one. The program still runs — it just prints the wrong receipt in two places, which is worse than a crash, because nobody notices.

A procedure removes the copies. You write the statements once, give them a name, and run them by calling that name.

PROCEDURE PrintDivider()
  OUTPUT "------------------------------"
ENDPROCEDURE

1. The header — PROCEDURE PrintDivider(). It names the procedure. Empty brackets mean it takes nothing in; they are still written.
2. The body — everything between the header and ENDPROCEDURE. These statements run when the procedure is called.
3. ENDPROCEDURE — closes the definition. Nothing after it is part of the procedure.

Notice that nothing has printed yet. A definition is a set of instructions in a drawer, waiting to be used — like a recipe card rather than the cooking.

To use it, write CALL:

OUTPUT "Receipt"
CALL PrintDivider()
OUTPUT "Total: ", Total
CALL PrintDivider()

CALL PrintDivider() means "run the statements with that name, then come back to the line after this one". The output is

Receipt
------------------------------
Total: 12.50
------------------------------

Two calls, one definition, three lines of output.