Database systems are valued today on account of their ability to manage complex information and efficiently process queries on large data sets. However, they are difficult to use, requiring careful design of database structure, and precise specification of queries to match this structure. In consequence, there is a large barrier to adoption. This project seeks to remove these burdens from the user through the notion of organic, rather than engineered, creation of both database and queries.
This project is developing databases that can be used even before the complete structure is specified. In the beginning, before much information is added, there may even be no structure at all. Over time, the user will be able to grow and evolve the structure organically as data is added and needs change. This project is also developing querying techniques for databases that reduce the burden on the user to specify a query exactly, and to know both the structure of the database being queried and the desired structure of the query result. This project, in contrast, will allow users to state an information need incrementally. Rather than structuring the result at the time of query specification, the user will be able to manipulate and structure the result set. These benefits are accomplished through the use of a presentation data model, which is implemented in the database system as a full-fledged layer above the logical and physical data model layers.
Biomedical database applications are employed to test and refine the developed system. The research results are expected to greatly improve the way databases are used by both technical and non-technical users and to influence a generation of research in data management. Additional information about the project is available at www.eecs.umich.edu/db/usable/.
