How it works - an overview

Before describing the details of the Tcl/Tk library we will give an overview of how it works with the Prolog system.

The Tcl/Tk library provides a loosely coupled integration of Prolog and Tcl/Tk. By this we mean that the two systems, Prolog and Tcl/Tk, although joined through the library, are mostly separate; Prolog variables have nothing to do with Tcl variables, Prolog and Tcl program states are separate, and so on.

The Tcl/Tk library extends Prolog so that Prolog can create a number of independent Tcl interpreters with which it can interact. Basically, there is a predicate which when executed creates a Tcl interpreter and returns a handle with which Prolog can interact with the interpreter.

Prolog and a Tcl interpreter interact, and so communicate and cooperate, through two ways:

1. One system evaluates a code fragment in the other system and retrieves the result. For example, Prolog evaluates a Tcl code fragment in an attached Tcl interpreter and gets the result of the evaluation in a Prolog variable. Similarly, a Tcl interpreter can evaluate a Prolog goal and get the result back through a Tcl variable.

   This is synchronous communication in that the caller waits until the callee has finished their evaluation and reads the result.

2. One system passing a "message" to the other on an "event" queue.

   This is asynchronous communication in that the receiver of the message can read the message whenever it likes, and the sender can send the message without having to wait for a reply.

The Tk part of Tcl/Tk comes in because an attached Tcl interpreter may be extended with the Tk widget set and so be a Tcl/Tk interpreter. This makes it possible to add GUIs to a Prolog application: the application loads the Tcl/Tk Prolog library, creates a Tcl/Tk interpreter, and sends commands to the interpreter to create a Tk GUI. The user interacts with the GUI and therefore with the underlying Prolog system.

There are two main ways to partition the Tcl/Tk library functions: by function, i.e. the task they perform; or by package, i.e. whether they are Tcl, Tk, or Prolog functions. We will describe the library in terms of the former because it fits in with the tutorial style better, but at the end is a summary section that summarizes the library functions both ways.

Taking the functional approach, the library can be split into six function groups:

• basic functions
  • loading the library
  • creating and destroying Tcl and Tcl/Tk interpreters
• evaluation functions
  • evaluating Tcl expressions from Prolog
  • evaluating Prolog expressions from Tcl
• Prolog event functions
  • handling the Prolog/Tcl event queue
• Tk event handling
• passing control to Tk
• housekeeping functions

We go through each group in turn.