Interests of Nick Rossiter
I am generally interested in information systems, with particular attention to their formal basis and to their application. More specific areas of interest are:
Interoperability in Heterogeneous Information Systems
The way forward in this complex area appears to be the application of a multi-level architecture to information systems in order to relate a top layer of abstractions via a series of mappings to the actual application data. This direction of mapping identifies policies, organization and instantiation. The equally important dual direction identifies naming, meta and meta-meta mappings. Currently four levels seem sufficient and work is continuing on designing an approach, which is as general and powerful as possible.
Application of Category Theory to Information Systems
A significant interest is the application of category theory to real-world problems. The priority has been to formalize information systems where higher-order logic is of clear relevance. Areas covered are:
a. Object-relational database systems where there are interesting abstractions such as association, inheritance and aggregation.
b. Information resource dictionary systems, where reference models play an important part in handling heterogeneous systems.
c. Hypertext systems where awareness is an important concept to capture.
d. Transaction systems employing ACID principles for atomicity, consistency, isolation and durability.
Other areas examined are the application of category theory to image and multimedia systems, natural language and environmental systems.
The unifying theme in recent work is the intension/extension relationship. Such relationships prevail in metaphysics and informations systems and indeed everywhere where there is a logical-physical interface. For example for information systems in areas such as databases, data normalisation is an attempt to represent the physical world in a logical information system. By employing a 4-level catgeoreal architecture with the topos as the highest-order structure and adjointness within and between the levels it has been possible to understand more fundamentally the interaction of real-world systems.
Nick Rossiter Visiting Fellow home page