摘要: This thesis presents a new programming model called the symmetric model in which the representation of programs is identical to the representation of data: to specify a computation, one defines a data structure. This data structur... 展开 This thesis presents a new programming model called the symmetric model in which the representation of programs is identical to the representation of data: to specify a computation, one defines a data structure. This data structure possesses the semantics of a first-class naming environment-it defines a scope and can be used to affect the evaluation environment of other expressions. We present a new programming language called Symmetric Lisp based on the symmetric model. Program structures in Symmetric Lisp are considered non-strict: a program's components may be examined even as its other elements continue to evaluate. The first part of the thesis investigates the interaction of non-strictness with first-class naming environments. The second part of the thesis discusses the compilation and implementation of Symmetric Lisp. We present an extended type-inference system for first-class environments that can be used to infer the proper evaluation environment of identifiers found within the scope of environment-yielding expressions. We also present a translation of Symmetric Lisp into a high-level data flow language. (fr) 收起