Taco - prototyping high-level object-oriented programming constructs by means of template based programming techniques

TACO (Topologies and Collections) is a template based object platform for cluster architectures, that provides the flavour of distributed data-parallel programming based on distributed object groups. This paper introduces TACO's basic concepts and discusses template based programming techniques to control the behaviour of (collective) method invocations. In particular, the problem of polymorphism. and concurrency control in the context of distributed, parallel computing is addressed