Collaborative Research: Formal Methods for Behavioral Subclassing and Callbacks

建议：合作研究：行为子类和回调的形式化方法david A. Naumann和Gary T. leavens为了可演化性、可扩展性和生产力，软件系统必须由可扩展组件组成。面向对象编程语言的特性，如继承和动态分派，以及回调和下调等技术，都是至关重要的，但它们颠倒了通常的抽象层。对象间的混叠对于提高效率至关重要，但可能会破坏封装边界。然而，要分别验证各个组件，抽象和封装是必要的。该项目将推进面向对象软件的规范、开发和验证理论，重点关注行为子类化、别名限制和回调。将研究用于行为接口规范的java建模语言（JML）的核心特性，使用约束纪律来控制混叠和模型程序来指定回调。这些想法将以一种有说服力的、简单的和健壮的理论来呈现，以促进工具的应用，比较规范符号和证明规则的替代建议，以及教学。该理论将被编码到定理证明器中，关键结果将由机器检查。这个项目将为编程和规范语言的设计者提供理论指导。研究结果还将有助于澄清和改进实践中使用的技术。直接应用预计在基于JML的项目和在我们的巴西合作者的工作。

Proposals 0429894/Naumann 0429567/LeavensCollaborative Research:Formal Methods for Behavioral Subclassing and CallbacksDavid A. Naumann and Gary T. LeavensFor evolvability, scalability, and productivity, software systems must be composed of extensible components. Features of object-oriented programming languages such as inheritance and dynamic dispatch, andtechniques like callbacks and downcalls, are crucial but they invert the usual layering of abstractions. Aliasing among objects is crucial for efficiency but can breach encapsulation boundaries. Yet abstractionand encapsulation are necessary to separately validate individual components.This project will advance the theory of specification, development, and verification for object-oriented software, focusing on behavioral subclassing, alias confinement, and callbacks. Core features of theJava Modeling Language (JML) for behavioral interface specification will be studied, using a confinement discipline to control aliasing and model programs to specify callbacks. The ideas will be presentedin a cogent, simple, and robust theory, to facilitate applications by tools, comparison between alternative proposals for specification notations and proof rules, and teaching. The theory will be encodedin a theorem prover and key results machine-checked.This project will provide theoretical guidance for the designers of programming and specification languages. The results will also help clarify and improve techniques used in practice. Direct application is expected in projects based on JML and in work by our Brazilian collaborators.