A Framework for Effective Algorithm Visualization Using Animation-Embedded Hypermedia

Abstract : If a "picture is worth a thousand words," then why have attempts over the past decade to use pictures and animations to replace or supplement traditional instructional methods for teaching algorithms produced such disappointing results? Numerous studies and experiments have been conducted to show that pictures and animations can improve learning of challenging, abstract concepts like mathematical proofs and the algorithms used in computer science. While the pictures and animations seem to be enthusiastically received by the students, none of the studies have produced results that show consistently and conclusively that these visual tools actually improve learning. In fact, the accumulated empirical evidence is mixed at best, and could easily lead one to abandon the premise that animations are powerful vehicles for effectively conveying the dynamic behaviors of algorithms. However, this dissertation reports on research based on the premise that a rethinking of algorithm animation design is required in order to harness its power to enhance learning. Research reported here explores the integration of previous work in algorithm animation systems with recent developments in the cognitive and educational domains to produce a new model for using software visualizations to improve student comprehension. The model is based on focused learning objectives that drive a top-down design that carefully divides abstract concepts into discrete chunks for learning. The model takes a user-centered ("what do we need to show") view rather than a designer-centered ("what can we show") view, and employs hypermedia and multimodal presentation techniques to improve learning effectiveness. The key insights are that for algorithm animations to be effective, (1) they should be introduced using interactive analogies and real-world examples that serve a priming role for subsequent learning.