Abstract

Software visualization and algorithm animation have been tackled almost exclusively from the visual point of view; this means representation and control occur through the visual channel. This approach has its limitations. To achieve better comprehension, we deal with multimodal interfaces that include the extended facilities of interaction together with those of the standard systems for data visualization and algorithm animation. The notion of specific concept keyboards is introduced. As a consequence, modern information and learning systems for algorithm animation are enhanced in such a way that control and interaction take place through appropriate interfaces designed and semi-automatically generated for this special purpose. In this paper, we provide some examples and report on a thorough evaluation to show the relevance of this new approach. © 2007 Elsevier Ltd. All rights reserved.