Deriving My Current Context - A Framework to Unify Context Services
Daily users struggle with organsing and re-finding their stored documents, photos, emails, etc. Personal Cross-Media Information Management investigates the opportunities to provide the user with an alternative to the Desktop Metaphor where we step away from the file hierarchies used to store our information. Our cross-media PIM system is a solution which unifies all organisational structures such as email, bookmark and file hierarchies but also the physical classification structures such as bookshelves or piles of papers on the desk. This approach to PIM makes it possible to construct our own personal cross-media information space with less fragmentation of the information. By providing a tool for centralising all these possible ways of storing our information, we allow the user to re-find the information no matter where or how the information has been stored. Often, users only need information which is relevant in a certain context, such as the slides of a course when studying for that course. Therefore, we include a context-aware component to provide the user with the context-relevant information instead of the currently used static file system.
One of the main contributions of the PIM framework is that it provides the possibility to build applications which provide the user with information relevant to its current context. Though, the context-aware research field has seen a major evolution the last decade, there is still a lack of frameworks which may fuse several individual context identifiers. Context identifiers are individual components which may define when a certain event appears. They translate raw data input to a higher-level event. By combining multiple context identifiers a more complex higher-level event can be detected. Nevertheless, to process all these context inputs there is a need for an architecture and framework which manage the individual context identifiers. The focus lies on a framework which need to support the functionality that an individual context identifier may use knowledge of other identifiers to preprocess their raw data. For example, a use case might be that we have a speech component which makes use of an agenda component to determine whether the speech subject is related to an agenda topic. In addition, a higher-level meeting component may then use the speech and/or agenda component to identify if the user is in a meeting and in which meeting.
This thesis will develop such an architecture and framework to handle the individual context identifiers. The framework will be an extension of the Mudra framework specialised to context identification. A proof of concept will be done by using several context identifiers and integrating the output in the PIM framework.
- Declarative programming
- You will learn about context modelling frameworks.
- You will learn about Mudra framework.
- You will learn about our PIM framework.