Our research program is focused on the design of user interfaces that support and encourage cooperation. The paradigm of cooperative action had been proposed ear- lier for special cases of bimanual object manipulation [138], multi-user collabora- tion [46, 234, 269, 293], or multimodal interfaces [61]. Chapter 5 discussed several prior interaction systems and their support for cooperative action. This thesis pro- poses consideration of cooperation as an overarching interaction principle that can be consistently applied to different types of interactive computer systems. Interfaces that support cooperative action enable higher versatility, flexibility, and expressive- ness for single users. At the same time, they encourage the involvement of multiple users, through the redistribution of subtasks.
From a review of literature in the fields of human motor control, interpersonal coor- dination, and HCI, we deduced that the design of cooperative user interfaces should evolve around three main themes: workspace coherence, input complementarity, and emergent territoriality. We have discussed the relevance of these aspects for single-user and multi-user systems. However, such high-level design principles cannot guaran- tee successful results. One can easily come up with inappropriate designs despite adherence to these principles, e.g., if the application case or the user requirements are not fully understood or appreciated. In particular, the choice of interface functional- ities, their ergonomics, and their accessibility depend on the application content, the interaction environment, as well as the goals and capabilities of the involved users. More elaborate heuristics, like those of Baker et al. for groupware usability [19], offer more detailed guidance, but, they neither provide readily applicable solutions. The suggested principles of cooperative user interface are deliberate abstractions of core aspects that can be applied to interface design more generally.
Research Program and Working Hypotheses 99 The research presented in this thesis is based on the following working hypotheses: H1 Interface support for cooperative action is beneficial for single users and, at the
same time, facilitates the collaboration of groups.
H2 The combination of concurrent user actions enables improved workload balanc- ing, higher interaction fluency, more flexibility, and extended interface func- tionalities.
H3 The design of cooperative user interfaces can be successfully informed by the high-level design principles workspace coherence, emergent territoriality, and com- plementary capabilities.
These high-level hypotheses describe a fuzzy relationship between the design paradigm of cooperative user interfaces and potential usability benefits. They may not be proven to be completely true or false. As an alternative form of support, Gaver’s concept of an annotated portfolio suggests a compilation of representative examples. For these examples, a detailed account of the invention, its development process, relevance, and extensibility as requested by Zimmerman et al. can be consid- ered an extended annotation (see above). Where possible, also a formal evaluation of concrete goals and expectations can be added. Moreover, a qualitative analysis of co- ordination behavior can extend established evaluation measures of user performance and satisfaction. The experimental results can only account for a particular configu- ration of parameters, but if the configuration is a relevant one, they can increase the plausibility of the provided design annotations.
The chapters of Part II report in detail on three concrete developments of user inter- faces for cooperation. All three have been published as mentioned in the list of pub- lications on page xxv. The three reports comply to a certain degree with the above discussed evaluation criteria, but each with a different emphasis.
The first example takes a previously proposed 3D input device [98] and explores its benefits in a bimanual interaction task. This work emphasizes the relevance of ecological validity in the evaluation of interface designs. In contrast to common per- formance studies with single-handed aimed movement tasks, the expected benefits of certain interface characteristics are less pronounced when tested in a holistic se- quence of cooperative actions. Small but significant improvements of general control accuracy could be obtained and a post-hoc task decomposition revealed the actual effect size for the affected subtask only.
The second example describes a novel multitouch interaction technique for mode switching. The design of this technique was based on a theoretical model of biman- ual motor control and the formal specification was based on a detailed analysis of temporal behavioral patterns. Here, the focus is on the invention of the technique and the process of its development. The results of a formal comparison against an established mode switching technique revealed performance advantages that can be attributed to differences in cognitive processing.
The third report features a novel 3D projection system that provides up to six users individually with stereoscopic image pairs. We discuss the relevance of such a system for 3D visualization applications, describe its functional principles, and detail our design rationales. Moreover, we describe the challenge of group navigation in shared virtual environments and suggest novel interaction techniques as a resolution. A formal user study revealed benefits of the system for cooperative visual search as well as the usability of our novel group navigation techniques.
These three examples have been chosen because they reflect the diversity of research and design approaches that can be pursued towards cooperative user interfaces. The combination of multiple existing interface technologies enables multiple input streams, novel interaction techniques can increase the expressiveness of such parallel input actions, and novel technologies create a whole new context for the development of cooperative user interfaces. These three examples form the basis of a portfolio of cooperative user interfaces which is extended in Part III of this thesis with additional application examples of the derived design principles.
101
Chapter 8
Bimanual Cooperation with
Desktop 3D Input Devices
This chapter reports on joint work with Jan Hochstrate, André Kunert, and Bernd Fröhlich at Bauhaus-Universtität Weimar. It has been presented at IEEE 3D User Interfaces 2009 and was published in the conference proceedings under the title: “The influence of input device characteristics on spatial perception in desktop-based 3D applications ”
© 2009 IEEE. Reprinted, with permission, from Kulik et al. 2009 [186].
Abstract
In desktop applications 3D input devices are mostly operated by the non-dominant hand to control 3D viewpoint navigation, while selection and geometry manipula- tions are handled by the dominant hand using the regular 2D mouse. This asym- metric bi-manual interface is an alternative to commonly used keyboard and mouse input, where the non-dominant hand assists the dominant hand with keystroke in- put to toggle modes. Our first study compared the keyboard and mouse interface to bi-manual interfaces using the 3D input devices SpaceTraveller and Globefish in a coarse spatial orientation task requiring egocentric and exocentric viewpoint naviga- tion. The different interface configurations performed similarly with respect to task completion times, but the bi-manual techniques resulted in significantly less errors. This result is likely to be due to better workload balancing between the two hands
allowing the user to focus on a single task for each hand. Our second study focused on a bi-manual 3D point selection task, which required the selection of small targets and good depth perception. The Globefish interface employing position control for rotations performed significantly better than the SpaceTraveller interface for this task.
8.1
Introduction
The majority of 3D graphics applications are still desktop-based, which is largely based on ergonomic reasons. The physical support for the operating hand on the desktop surface efficiently reduces fatigue. While there is a broad variety of inter- faces for immersive virtual environments, desktop-based 3D applications are mostly managed with the familiar mouse and keyboard set-up. The operation of such 3D applications requires a lot of mode changes if only a 2D mouse and a keyboard are used. In this case, the non-dominant hand assists the dominant hand with keystroke input to toggle modes. This approach has two major drawbacks: The workload distri- bution between both hands is very unbalanced and integral 3D manipulations need to be separated into a sequence of 2D actions. Prior work of Jacob et al. [157] and Hinckley et al. [141] indicates that the second issue can affect performance.
Figure 8.1:The Globefish input device for CAD and DCC applications
The SpaceMouse™ (with its smaller descendants SpaceTraveller™ and
Related Work 103 respectable acceptance among users. The design enables the integral operation of 3D translation and rotation via elastic rate control. The 3D input device Globefish [98] separates translational and rotational input by its hardware design (Figure 8.1). Such 3D input devices are mostly operated by the non-dominant hand to control 3D viewpoint navigation, while selection and geometrical manipulations are handled by the dominant hand using the regular 2D mouse pointer. However, there is no scientific evidence if and why such an asymmetric bi-manual interface configuration is a good choice.
In a first step we analyzed the interaction requirements of desktop 3D applications. Based on our observations we implemented two scenarios and evaluated them by user studies. The first study compared regular keyboard and mouse input to two- handed input using a 3D input device and a mouse. The task focused on coarse spa- tial orientation in egocentric and exocentric viewpoint navigation. We constrained the required degrees of freedom such that they could be directly provided by the 2D mouse. Our results indicate that the input device configuration does not have much influence on the time efficiency in coarse spatial orientation tasks. However, bi-manual techniques resulted in significantly less errors. We argue that this obser- vation is due to better workload balancing between the two hands allowing the user to focus on a single task for each hand. Our second study compared the two 3D de- vice concepts SpaceTraveller and Globefish in a bi-manual 3D point selection task. The task required high selection accuracy and good depth perception, which had to be achieved through 6-DOF view point navigation. The results of this study re- vealed significant benefits for the position-controlled rotational input provided by the Globefish device.