Automation of composition of programs from components of significant size is one of the "Holy Grails" of programming. This lecture will present a conceptual framework called Compositional Objects for domain-specific composition of programs from components of significant size. This conceptual framework enables composition to take place at compile time, initialization time or run time. Performance models are the domain used to illustrate these concepts. Performance models span software, middleware and hardware semantic domains and multiple levels of abstraction. Evaluation of components requires multiple modes of execution. Execution may require translation across levels of abstraction and modes of evaluation. Compositional objects are an integration of associative interfaces with standard objects and use of hierarchical dynamic dependence graphs as a program structuring model. Associative interfaces extend the concept of interface from specification of services provided to include specification of behaviors of component instances and specification of the services a component instance invokes. This talk will define and describe an associative interface based specification system and a compiler for software systems implementing performance models.
Referent: Prof. Dr. James C. Browne,
University of Texas at Austin
USA
Zeitpunkt: Freitag, 19. Mai 2000, 14 Uhr c. t.
Ort: HS 3 der Universität Klagenfurt
James C. Browne is Regents Chair in Computer Sciences, Professor of Physics, and Professor of Electrical and Computer Engineering at the University of Texas at Austin, where he also received the PhD degree in 1960. His areas of interest include parallel computation with the major focus on parallel programming, high level specification languages and integration of computer science with application areas. Another current interest is performance evaluation of parallel and distributed systems.
Professor Browne has been involved in the CODE (Computationally Oriented Display Environment) project, where an abstract graphical environment for parallel programming has been developed. Ongoing research includes methods for optimization of parallel computation structures at high levels of abstraction, integration of parallel structuring through data partitioning into the generalized data flow model of CODE, debugging in the graphical/visual environment, compositional approaches to parallel programming and the addition of intelligence process control to parallel programs in computational fluid dynamics. Professor Browne is also working on design and development of narrow domain compilable high level specification languages including logic-based languages and robust methods for programming intelligent real-time decision systems.