Experimental Software Systems: Linear Naming: Experimental Software for Optimizing Communication Protocols

EIA-9806718 Mairson、Harry G. Bawden、艾伦布兰迪斯大学 ESS：线性命名：优化通信协议的实验软件 我们研究项目的目标是为本地和分布式通信设计和实现显着新形式的过程调用协议。该项目将应用技术的开发与编程语言设计的基础研究结合起来。该研究的重点是实现一种新的过程调用协议，该协议将按值调用的理想语义与按名称调用的效率结合起来。其实现依赖于共享值的技术以及共享计算线程的技术。密切相关的是用于以对称方式处理值和控制结构的语言原语的开发。 共享的实用性与分布式网络计算中“线性命名”的关键作用密切相关，特别是当命名计算是移动的时。过程调用协议及其与命名的交互形成了将复杂计算粘合在一起的粘合剂。在分布式环境中，不断发展的计算的各个部分可能会彼此分离。一个根本问题是在适当的时间到来时将这些部分重新组合在一起——这就是命名的目的。适当关注命名中的线性可以使跨网络引用变得便宜，可以支持高度移动的数据结构，并且可以促进启发式方法以在没有明确指导的情况下有效地按需迁移任务。

EIA-9806718 Mairson, Harry G. Bawden, Alan Brandeis University ESS: Linear Naming: Experimental Software for Optimizing Communication Protocols The goal of our research project is to design and implement significantly new forms of procedure calling protocols for local and distributed communication. This project combines development of applications technology with foundational research on programming language design. The research focuses on the implementation of a new procedure-calling protocol that incorporates the desirable semantics of call-by-value with the efficiency of call-by-name. The implementation depends on the technology of sharing values as well as the sharing of computation threads. Closely linked is the development of language primitives for handling values and control structures in a symmetric manner. The pragmatics of sharing is intimately connected to the crucial role of "linear naming" in distributed network computation, especially when named computations are mobile. Procedure-calling protocols and their interaction with naming form the glue that holds complex computations together. In a distributed environment, parts of an evolving computation can become separated from each other. A fundamental problem is bringing these parts back together again when the appropriate time arrives---this is what naming is for. Proper attention to linearity in naming can make cross-network references cheap, can support highly mobile data structures, and can facilitate heuristics to effectively migrate tasks on demand without explicit guidance.