Modality Preference – Learning from Users Rainer Wasinger 1 , Antonio Krüger 2 1 DFKI GmbH 66123 Saarbrücken Germany rainer.wasinger@dfki.de 2 University of Münster 48149 Münster, Germany antonio.krueger@uni-muenster.de Abstract. An important constituent for mobile and ubiquitous computing sys- tems is the interface and the associated human-computer interactions. Mobile contexts have different design requirements to stationary desktop contexts. Al- though previous work has concentrated on stationary domains and unimodal systems, and more recently on multimodal systems, user evaluation for the use of different modality combinations is limited. In this paper, we outline the qualitative results from a recent usability study. These results form a general guideline in determining which base modalities and modality combinations to use when designing for mobile and ubiquitous environments. 1 Introduction Stationary computing scenarios relating to document processing and web browsing are currently nourished by interactions based on keyboard and WIMP (Windows, Icon, Mouse and Pointer). Mobile scenarios in contrast quite often have restrictions placed on them that render such interaction as inadequate. A user may for example be standing or walking, and computing resources may be limited to the use of a mobile phone, Personal Digital Assistant (PDA), or the environment’s public infrastructure. Speech is one modality that has received a lot of attention in the recent past as processing power for mobile devices has increased. Similar to WIMP, speech input is however not always appropriate (e.g. in a noisy environment). Multimodal interaction is the current trend for modern mobile applications, and this is partly due to the flexi- bility and expressiveness [2] that such systems exhibit (e.g. Smartkom Mobile [3], QuickSet [1], and the Mobile ShopAssist [4]). User studies are however still limited, due to the large task of evaluating the very rich range of modality combinations that may be formed from even a limited number of base modalities such as speech (S), handwriting (H), and gesture (intra GI, and extra GE), as outlined in [4]. In this paper, we report on qualitative results that were obtained from two studies, the first conducted in a laboratory setting on 14 users, and the second conducted in a real-world electronics store of the Conrad 1 chain on 27 users. During the testing, our 1 Conrad Electronic, Saarbrücken.