Software Infrastructure and Design Challenges for Ubiquitous Computing Applications

Banavar, G. and Bernstein, A. 2002. Software infrastructure and design challenges for ubiquitous computing applications. Commun. ACM 45, 12 (Dec. 2002), 92-96.

This article identifies the important application design and software infrastructure challenges that must be address by the ubicomp research community. Like a paper I wrote last year on “getting real with the utopia around ubiquitous computing”, the authors mention that we are still far from Weiser’s vision.

Key characteristics of ubiquitous applications are:

Task Dynamism
Adaptation to the dynamism of the users’ environments and the resulting uncertainty. Sometimes, application won’t make the proper inferences. Therefore the user might actively reconfigure the system to adapt to the new task settings. Applications will have to be able to explain what they inferred and learned from their right and wrong inferences.

Device Heterogeneity and Resource Constraints
Hardware and software often lack of heterogeneity and devices have physical constraints. Those limitations influence the development of applications and their capabilities.

Computing in a Social Environment
Privacy issues

Research challenges are:

Semantic Modeling
Use of ontologies to describe users’ task environments, as well as their goals, to enable reasoning about a user’s needs and therefore to adapt to changes. It is a challenge to develop (and agree!) on a high-level model language to express the complex nature of ontologies.

Building the Software Infrastructure
The application must determine the user’s context, must provide reasonable functionalities with bad connectivity, must recover from failover, and must be scalable.

Developing and Configuring Applications
There is a need of a shift in the developers’ mindset while building pervasive application. There is a need to describe on a high-level the task a user needs to perform. The challenge is to be able to specify the interaction logic at an “intent-level” and the application’s requirements on data and computation.

Validating the User Experience
The utility of some computing advancements cannot be evaluated without performing significant user studies and in some cases, widely deploying it. Consequently, the development of effective methods for testing and evaluating the usage scenarios enabled by pervasive applications is an important area that needs more attention from researchers.

Relation to my thesis: The authors acknowledge that both applications and users must adapt in real-world ubiquitous environments. Because of inevitable sense of uncertainty, application should give a sense of situation awareness. In my thesis, I try to give the same pragmatic view on current challenges, being physical and “organizational”. The several “clouds of connectivity” over us, raises the bigger issues around heterogeneity of hardware and software (including agreeing on semantic modeling). Heterogeneity might be a bigger challenge than some of the physical issues of devices. To be widely and rapidly successful, ubiquitous environments should be based on the homogeneous Minitel standards and protocols (top-down government). Other huge challenge I see is, of course to grasp the user’s context and then scaling.

The authors do not cover other social impacts other than privacy. Very little about appropriation. The Relevance of Social Issues in Ubiquitous Computing Environments has more on that subject.

I never really thought on the way it changes the engineers perspection. But it seems that there is a need to be able to describe a system on its interaction with the user.

My thesis completely fits in the “validating the user experience” category and an output might to provide tools to describe tasks and system-user interaction integrating the limitations of the environment. Helping the change the engineers perception of ubicomp development might be a high-level outcome. Field-based quasi-experiments advised by the author to capture the rich nature of the usage environments make me feel that CatchBob! (and its children) as a good research platform.

Reference I should read:
Plans and Situated Actions: The Problem of Human-Machine Communication, Lucy Suchman. Cambridge University, 1987.