“Ubiquitous computing” has been fueled by strong visions of Mark Weiser’s disappearing computers” or the need for “calm” technologies, but although the ever-increasing numbers of smart houses, intelligent assistants or mobile location-based applications have found niches, this has not yet led to their adoption by quotidian users.

Learning from others’ mistakes

If one does not want to reinvent the wheel, interrogating past failures and drawing parallels with other domains is relevant. What do the failures of ubicomp mean for grid computing? Simply put, the whole point of focusing on failures is the assumption that any human-made artifacts can benefit from past failures in other fields.

Although grid technologies are more remote from users, they may suffer from similar issues regarding, for instance, bad integration of utility computing in commercial practices or lack of trust in volunteer computing.

Getting back to failures

At the LIFT 2008 conference, recently held in Geneva, Switzerland, 80 practitioners from different fields discussed the failures of ubiquitous computing prototypes and products, aiming to criticize assumptions and existing models of interaction.

A large part of the discussion focused on determining what counts as a “failure”: eBooks, smart furniture, disappearing technologies, vending machines and mobile social software all received a mention. Further discussion covered the reasons for these “failures,” which were thought to include:
- Bad design and non-usable interfaces
- User frustrations: painful experiences; practices that become more complex with a product than without
- Contextual issues: conflicting context and practice idioms; flawed user models or scenarios
- Marketing problems: no market, no match with consumer expectations, no integration in the value chain
- Technoptimism: poor product vision and expectations; perpetuation of wrong ideas
- Technological shortcomings

While this list of failures stemmed from discussion of ubiquitous computing, it is not a list exclusively restricted to the ubicomp arena. Perhaps grid computing can avoid falling in to some of these traps as it too evolves towards a greater market and acceptance.

- Nicolas Nova, Fabien Girardin and Julian Bleecker