MS Project Pre-Proposal QUERYING THE DATAFLOW WAY Swathi Annamalai Chair: Professor Hans-Peter Bischof Department of Computer Science, Rochester Institute of Technology Visualization of a series constituting a problem, phases of operation towards acquiring a final output, can all be graphically perceived thus categorizing such a methodology as one of the most efficient programming practices in the scientific world of computing. [1] An attempt to give programming a different perspective focuses on the "data" being computed rather than the "manner" in which the data is obtained, which motivated the development of a new set of languages called "Dataflow Languages". The inherent objective for such a development was exploitation of parallelism, embedding concurrency into operations, thus enhancing performance for a system. The vital property for such a language is the gateway provided for users to visually view the sequence through which transformations are applied to their data producing a final graphical result. [2] Dataflow capabilities can be extended to relational databases incorporating new design principles and functionalities developing a branch of database languages called as visual query languages. My project will constitute work, which will be an effort towards making databases visually complete by incorporating dataflow principles in the existing database querying techniques. Several of the dataflow query languages developed today is an inception of programming queries for databases using dataflow practices for small solution sets [3], which needs to be extended for solving complex queries using diversified examples. I shall be concerned with experimental practices under such a study, which will include research for problems within the scope of the databases and subsequently develop solutions for the same using data retrieval and analysis operations as a dataflow graph [3]. My work will involve analyzing database domains concerning real world applications, produce dataflow solutions as a measure towards solving these examples with improved query performance, parallelizing operations and strategies for providing a graphical representation of the entire query execution process to the end user. The principle form of measure for my project will be considering dataflow properties, some of which are scheduling, template representation of queries, branch prediction using data and demand driven approach of dataflow language, efficient data tokens being reused within the system, reduction in processing of these data items, manipulation of input operands supplied to system modules and successive handling of dynamic data within pre-calculated subcomputations. My vision through this project is to develop principles and algorithms keeping dataflow as a prime aspect of consideration, thus contributing towards the computational society for enhancing databases in a graphical manner. [2] Such a development would facilitate a user execute queries without knowledge of the underlying data structure, database or querying language, but yet efficiently process their query to obtain a final result. REFERENCES: [1] Wesley M.Jognston, J.R. Paul Hanna and Richard J.Miller. Advances in Dataflow Programming. University of Ulster. [2] Bosco S.Tjan, Leonard Breslow, Sait Dogru, Vijay Rajan, Keith Rieck, James R.Slagle and Marius O. Poliac. A Data-Flow Graphical User Interface for Querying a Scientific Database. Computer Science Department, University of Minnesota, Minneapolis MN 55455. [3] Sait Dogru, Vijay Rajan, Keith Rieck, James R.Slagle, Bosco S.Tijan and Yuewei Wang. DFQL: A Graphical Data Flow Query Language for Retrieval, Analysis and Visualization of a Scientific Database. In July 1994, Computer Science Department, University of Minnesota, Minneapolis, Mn.55455