Speech displaces the graphical crowd

Developers of visual Interface Design Environments (IDEs), like Microsoft Visual Studio and Java NetBeans, are competing in producing pretty crowded graphical interfaces in order to facilitate completion of interface-design tasks. Previous studies have shown that such interfaces cause the user to experience information overload as well as they create a fertile environment for usability problems. In this paper, we empirically investigate speech as an input and output means for solving usability problems with GUI interaction metaphors and enhancing usability of visual IDEs. The empirical investigation aimed at measuring usability of two experimental ID toolkits: typical visual-only (TVOID) and multimodal (MMID) in terms of efficiency, effectiveness and satisfaction. Usability of these two environments was tested twice, over two experimental phases, by four independent groups of users. The first experimental phase was for measuring usability by novel users in order to explore how learnable each of the two experimental toolkits would be. The second phase aimed at measuring usability by well-trained users to study experienced user performance when using a speech recognition ID.