The Interpreter, Chapter 6
The continuing story of one man’s quest to write an interpreter.
Last time, our hero greatly improved the concept and implementation of the interpreter. This time, he took that even further. Soon, if statements and while loops were working. Previously, control structures had been managed by a single 500-line if statement. Not a pretty sight. The new and improved interpreter treated control structures as functions — even the statement separator (which in this case was a semicolon).
What about the interpreter’s standard output? The interpreter was printing a lot of stuff on the screen, so this was important. But the standard output was rather lacking in quality. If you tried to print out more than 1 MB of text, the program would crash.
Jackson spent over eight hours revising the output to be cleaner and catch more errors. The most useful development was a formatted output function, which caught virtually every error possible. All the other output functions could lean on this one core function. The interpreter expanded from having about three output functions to about ten — but each function was cleaner, simpler, and more versatile.
Almost everything from the old interpreter had been re-implemented, and was working better than ever. It was time to move forward. Jackson now had to implement a critical piece, something no language can do without: functions.
To be continued. . . .
