Static and dynamic visualisations of monadic programs

Abstract

iTasks is a shallowly embedded monadic domain-specific language written in the lazy, functional programming language Clean. It implements the Task-Oriented Programming (TOP) paradigm. In TOP one describes, on a high level of abstraction, the tasks that distributed collaborative systems and end users have to do. It results in a web application that is able to coordinate the work thus described. Even though iTasks is defined in the common notion of “tasks”, for stake holders without programming experience, textual source code remains too difficult to understand. In previous work, we introduced Tonic (Task-Oriented Notation Inferred from Code) to graphically represent iTasks programs using blueprints. Blueprints are designed to bridge the gap between domain-expert and programmer. In this paper, we add the capability to graphically trace the dynamic behaviour of an iTasks program at run-time. This enables domain experts, managers, end users and programmers to follow and inspect the work as it is being executed. Using dynamic blueprints we can show, in real-time, who is working on what, which tasks are finished, which tasks are active, and what their parameters and results are. Under certain conditions we can predict which future tasks are reachable and which not. In a way, we have created a graphical tracing and debugging system for the TOP domain and have created the foundation for a tracing and debugging system for monads in general. Tracing and debugging is known to be hard to realize for lazy functional languages. In monadic contexts, however, the order of evaluation is well-defined, reducing the challenges Tonic needs to overcome.