Pervasive computing applications often need to be context-aware, using various
contexts such as location and time to adapt to the evolving environment. Although the
use of context-awareness represents a promising approach to facilitating the
construction of intelligent pervasive computing applications, it also introduces various
challenges to context processing and management. This survey presents an analysis of
the challenges. We find that they are generally distributed over four aspects: the
acquisition of environmental contexts, the computation of derivative contexts, the
programming of context utilization, and the management of context services. For each
aspect, recent research work is systematically discussed. The discussion starts off
using a healthcare industry example.