Distributed Language-based Environments

多年来，基于语言的环境一直是编程环境研究的首要主题之一，但是对于在大型软件系统上共同工作的专业程序员群体来说，使它们变得实用的工作很少。相反，基于语言的环境主要面向由单独工作的程序员开发相对较小的程序。这项研究的目标是完成一个新系统的开发，水星，改变这种情况。MERCURY从属性语法生成基于语言的环境。生成的环境可以由在同一台机器上工作的多个程序员同时使用，也可以在分布在网络上的不同机器上工作。MERCURY使用分布式属性计算算法提供从一个模块到另一个模块的更改的即时传播，立即通知程序员由于其他程序员对其他模块的更改而导致的模块中的错误。额外的算法支持高可用性和可靠性，因此程序员即使在部分网络损坏或某些机器停机时也可以继续工作，相信一旦网络或机器恢复，MERCURY将重新建立一致性。提出的工作包括开发新的算法来支持：在运行期间选择任意程序单元的分布；高效的聚合属性传播；以及多个版本和配置模块。

Language-based environments have been one of the foremost topics of programming environment research for many years, but little has been done to make them practical for groups of professional programmers working together on large software systems. Instead, language-based environments have been oriented primarily to the development of relatively small programs by individual programmers working alone. The goal of this research is to complete the development of a new system, MERCURY, that changes this situation. MERCURY generates language-based environments from attribute grammars. A generated environment can be used simultaneously by multiple programmers working on the same machine or on different machines distributed across a network. MERCURY uses a distributed attribute evaluation algorithm to provide immediate propagation of changes from one module to another, immediately informing programmers of errors in their modules caused by changes made by other programmers to other modules. Additional algorithms support high availability and reliability, so the programmers can continue working even when part of the network is broken or some machines are down, trusting that MERCURY will reestablish consistency as soon as the network or machine is restored. The work proposed involves developing new algorithms to support: distribution of arbitrary program units selected during operation; efficient propagation of aggregate attributes; and multiple versions and configuration of modules.