A truly personal machine, called a private machine and implemented as a Personal Digital Assistant (PDA), is fundamentally
different from traditional machines. It is
personal and
private in an unprecedented manner, and its modus operandi is such that network and power failures will not be rare. Designing distributed
systems where PDAs are treated as “first class citizens” is a challenge.
Furthermore, private assets (electronic money, keys for authentication and opening doors) will be stored in PDAs. Ownership
and control of these assets and the media that store and communicate them should remain with the user. This must be reflected
in the design of systems for private computing.
We introduce the “open-ended argument” to describe the design strategy we used for designing a system that is designed to
reveal information to the user (as opposed to hide it). We argue and show that when systems are designed this way, the user
(a human) is better able to control the system and his personal data, as he can make better decisions than the system itself
based on qualitative assessment of the provided information. The system we have designed and implemented under this design
guidelines is presented and discussed.
Acknowledgments Frode Fjeld, Åge Kvalnes and the anonymous referees gave us feedback that has improved the presentation. Arne Helme participated
in the work on offline delegation. Working in the “PASTA laboratory” is very stimulating.
Funded by the GDD project of the Research Council of Norway (project number 112577/431).