3D transformation widgets are commonly used in many 3D applications operated from mice and keyboards. These user interfaces allow independent control of translations, rotations, and scaling for manipulation of 3D objects. In this paper, we study how these widgets can be adapted to the tactile paradigm. We have explored an approach where users apply rotations by means of physically plausible gestures, and we have extended successful 2D tactile principles to the context of 3D interaction. These investigations led to the design of a new 3D transformation widget, tBox, that can been operated easily and efficiently from gestures on touch-screens.

Mobile interfaces are evolving towards touch-based approaches. This allows users to interact with their thumb directly on the screen. Such kind of direct approaches may be fussy for 3D interaction tasks, in particular because of thumb occlusions. In this paper, we introduce a new 3D user interface for the control of a planned trackball, where the users sketch horizontal or vertical movements to observe an object. A user study revealed no significant difference for error rate between this new approach and a standard trackball control. Despite a better completion time with the direct control, the study showed that the subjects preferred using the planned version of the trackball because it limits disorientation.

3D interactive applications are now appearing on mobile devices such as phones and PDAs. Compared to classical desktop or immersive configurations, mobile devices induce several constraints that intrinsically limit the user performance in interactive tasks. Consequently, a special effort has to be made in order to adapt the classical 3D user interfaces to mobile settings. In this paper, we propose a new key-based technique that favors navigation in 3D environments. Compared to a classical "go to" approach, our technique called Z-Goto directly operates in the 3D space. This accelerates the user's displacements by reducing the number of required keystrokes. Moreover, the construction of the cognitive maps is improved, as Z-Goto favors depth perception. An experiment shows that Z-Goto obtains better completion times compared to a standard "go to" technique for a primed search task on mobile device. It also shows that the user satisfaction for this new technique is good.