W6 HCI Issues in Proactive Computing

Antti Oulasvirta, Antti Salovaara & Peter Ljungstrand
Helsinki Institute for Information Technology; Interactive Institute &
Interaction Design, Chalmers University of Technology

Introduction

As computers are proliferating our everyday environments, technology-enthusiasts have envisioned that computers should work more together to leverage a greater net benefit for users. Advances in input and output technologies, low-power computing, intelligent middleware and operating systems, and sensor and wireless networks form the technological backbone of these visions. Even though much of the seminal research at places like Xerox PARC sprung out of user-oriented concerns, most advances in this area have not been driven by user needs. However, this is an increasingly important area in HCI. Issues like privacy and trust, multitasking and interruptions, personalization and configurability, and identity and roles are receiving more and more attention. To address these issues, different approaches to design have been put forward by prominent researchers.

Proactive computing believes in shadowing the user and inferring her interests and capabilities, and using this knowledge to act on behalf of (or pro) the user. The premise is that human behavior is regular enough to be learned and predicted. Even though it is not possible to infer everything correctly, we believe that it is possible to achieve a much higher degree of useful, accepted proactive services of end users than are available today.

The proposed proactive services range from resource preparation and optimisation to higher-level decision-making. It has been argued that without proactivity, pervasive computing leads to a situation where the user must monitor and control several computing processes simultaneously. As users' cognitive resources are severely limited, the multitasking situation easily results in disruption and loss of control. Yet another demand for proactivity comes from the contextually intelligent provision and preparation of resources for the user, for example notifying her of a friend passing by or alarming a nurse of a potentially dangerous situation in a patient care domain.

Workshop theme

Proactivity has, quite naturally, provoked critical voices from the HCI society. Depriving the user from control does indeed run counter to the heuristics of interaction design that aim for user empowerment and control. Several novel HCI questions have risen that are specific to the proactive computing endeavour and need to be solved. The workshop seeks high quality papers related to all aspects of proactivity and human-computer interaction. The topics to be contributed to at the workshop include, but are not limited to:

• User needs for proactive computing . What is proactivity good for, and what is it not suitable for?
• Trust . What is required for the user to trust in a computer working on behalf of her?
• Representing proactive logic . Representing and visualizing the proactive logic is essential for design goals like transparency, veridicality, and predictability.
• User control . What are the best means for controlling proactive computers and agents?
• UI design . What kind of manipulation and feedback mechanisms do users need, at what levels, how often, and how should feedback be manifested? How can we design user interfaces that take advantage of all the human senses, as well as our inherent skills in moving about in the real world and manipulating real things?
• User-centered design processes for proactive systems . Are there special requirements stemming from proactivity for user-centered design processes?

Web page: http://www.hiit.fi/uerg/nordichi04-proactive/