Transcript Why Should We Care about Context Aware Computing

Why We Should Care About
Context-Aware Computing
Munindar P. Singh
www.csc.ncsu.edu/faculty/mpsingh
[email protected]
April 2002
Proposal
• Facts: Context-aware computing is
– Hot.
– Part of IDM.
• Choice: We can either
– Ignore it: so only platform and data questions will remain (with ad
hoc approaches for meaning).
– Acknowledge it: emphasize semantic questions; have another
motivation for our research.
• Motions: Add
– Context-aware computing, e.g., for mobile and pervasive
scenarios, as an application area.
– Suggested research questions.
Where can we help and why?
• State of the art: suck data from a bunch of sensors
– Stream it, store it, munge it, mine it.
– Call it user context and hope it relates to meaning.
• Great testbed for fancy modeling with subtle yet realistic
context “experience” that people can easily appreciate:
– Context  sensor data, but is related to it.
– Subtle relationships between value, quality of services, trust and
context appropriately modeled:
• Ontologies and process models to accommodate usage and user
context.
• Applying and mining sensor data in a meaningful manner.
Research Questions
• How can we model, fulfill services that are
– Suited for user interaction, i.e., intermittent, interruptible,
exceptional?
– Support precise notions of quality, relevance, …?
• How can we develop WS techniques and architectures that
apply context while enabling
– Enabling easy administration and portability across installations?
– Provisioning and personalization?
– Protecting privacy?
These questions have broader value, but we will address them
sooner and more clearly if we view them as independently
motivated for context-aware computing.