Towards ?Phenomenaria? in the Teaching of Distributed Systems Concepts

For Perkins [1], any learning environment can be parsed into five facets, not all of which are always present. These five facets are information banks, symbol pads, construction kits, ‘phenomenaria’ and task managers [1]. Information banks are sources of explicit information about a topic like textbooks. Symbol pads are surfaces for the construction and manipulation of symbols like word processors. Construction kits refer to a collection of building blocks that when moulded together form an entity. Lego Mindstorms is an example. ‘Phenomenaria’ are areas for the specific purpose of presenting phenomena and making them accessible to scrutiny and manipulation. SimCity, a program that models a city’s environment, is an example. Task managers are elements of the environment that sets tasks to be undertaken in the course of learning guide and provide feedback regarding process and/or product. The teacher is an example [1]. Taken together these five facets offer a perspective on the general structure and style of the learning environment and its underlying assumptions about the nature of learning and teaching. Typically today’s lecture theatre contains three of the five facets (information banks, symbol pads and task managers). Implicit in this profile are the premises that learning occurs through telling students about things, that students cannot manage their own learning and that solving problems rather than constructing entities is primary [1]. On the contrary, understanding is not something that comes free with full databanks; it is something won by the struggles of the student to learn, to conjecture, to probe and to puzzle out [1]. Thus, a shift in emphasis is needed away from the information banks and more towards construction kits or phenomenaria. The reason for this shift is because the latter facets place the learner directly and emphatically in the position of having something to make sense of or with, respectively. Moreover the role of task manager falls to that of the student albeit with scaffolding from the teacher. This shift in emphasis is what the author is proposing for the teaching of distributed systems concepts. But first what is a distributed system and why the need of phenonmenaria in this subject domain? A collection of processes, which are distinct, spatially separated, and which communicate by exchanging messages constitutes a distributed system [2]. A distributed algorithm defines the steps to be taken by each process within that system, including the transfer of messages. It must be able to deal with the failure of one or more of the processes involved in its computation and also the failure of one or more message transmissions. This makes the task of describing all states of the algorithm difficult [3]. A major problem in teaching the latter material is the ability to capture the dynamic movement of data. Typically, when demonstrating data movement on a white board, part of the existing data configuration must be erased in order to show the new configuration. Moreover, the lecturer must decide the appropriate pace at which to reveal the configurations. However, no matter what pace the lecturer chooses it will be the wrong pace for some students because different students learn at different speeds [4]. A phenomenarium offers a way forward for both lecturers and students in that it provides mechanisms not only to animate but to simulate dynamic movement in visual form. Moreover, it can be housed within an environment that permits ‘anytime anywhere’ learning, like the world wide web, and as such offers the possibility of direct manipulation of visual display in real time, something simply not possible with static media like the whiteboard. Simply stated, a phenomenarium permits a student to interact with elements of a given algorithm in their own time, and to perform operations appropriate to that algorithm thereby gaining knowledge as to the workings of that algorithm. This paper is an argument in favour of the use of the latter in the teaching of distributed algorithms.