In this paper we introduce Prop, a multiparadigm extension of C++ with Standard ML-style algebraic datatypes and pattern matching, tree rewriting, DATALOG-style forward chaining inference, and constraint logical programming. Applications written in Prop can utilize various cooperating formalisms, integrated into the object oriented paradigm of the base language. We use efficient automata-based and semantics based algorithms to generate various pattern matching constructs into efficient and lightweight C++ programs. Interoperability with the base language is achieved transparently since all high level data structures in Prop are mapped into classes. Furthermore, we use conservative garbage collection schemes to minimize interaction with existing code by eliminating the need for manual storage management. Typical Prop program sources are 2--10 times more compact than equivalent programs written in C++. Our benchmarks also show that programs written in Prop's high level formalis...