Practical and Theoretical Aspects of
Forth Software Development

Peter J. Knaggs

Abstract

This is an investigation into the use of the Forth programming environment. The main areas of enquiry were: interfacing Forth to other languages; interfacing Forth and local area networks; and the use of RISC processors with stack based architecture such as the NC4000 and Harris RTX series.

We describe how to interface Forth and C. We also provide a system with a multi-tasking interrupt driven interface to the IBM NetBios networking software and a simple, generic, method of task activation through message passing.

Many aspects of the investigation proved to be dependent on a more thorough theoretical underpinning for the Forth language. The use of a typeless parameter stack means that a programmer must concern himself with the intellectual burden of managing the parameter stack. The mismatching of stack elements can be the cause of subtle logic errors. We therefore investigated the possibility of developing a "type algebra" that would allow us to develop a typed version of Forth. This thesis includes a theory for a "type signature algebra" for the stack based argument passing method used by Forth.

To support the use of multi-tasking we provide a simple, but formal, theory of concurrent tasks based on state machines that synchronise on events. This has a graphical notation for people who are not familiar with formal notations.

We also looked at how formalisms might be used to define a semantic model for the Forth language and how formalisms can help to define the relationship between Forth's stack based virtual machine and register based target processors.

[ Peter's Home page | Table of Contents | Forth Research page ]
C makes you think that C is the world's greatest language.
Forth makes you think that you're the world's greatest programmer.