Static and Dynamic High-Integrity Architectural Application Frameworks

通过组件组装构建模块化、柔性的软件系统是软件工程的核心问题之一。为了解决这个问题，许多不同的方法，如面向对象框架和产品线体系结构，已经被提出作为创建和扩展软件系统的范例。这些范例支持依赖于所有应用程序共享的核心系统的体系结构应用程序家族（AAFs），其中每个应用程序因其变量扩展而不同。******然而，虽然这些体系结构应用程序家族是通用的，但是有很多机制支持它们的多样性。实际上，这些方法和其他方法所支持的扩展可以基于许多不同的基于关注点的组合机制，例如方面、面向对象视图、模式、高阶多态性、连接器和包装器。这些扩展导致了不同的模块概念，但是这些方法之间的关系目前还没有得到很好的理解。组合和集成静态和动态的基于关注的扩展（分别在设计和运行时）是一项复杂的任务，内部违反或不期望的交互导致的失败可能最终只会显示为代价高昂的事故。此外，基于关注的扩展需要由涉及约束和契约的保证来管理，并且需要验证方法来确保基于关注的扩展按照预期的方式运行。******这个研究项目的总体目标是定义和评估一种新的灵活和模块化的方法，用于在高完整性架构应用程序家族的设计和进化中实现基于静态和动态关注的扩展。研究的重点是：(1)基于关注的模块化和扩展的基本问题；（ii）基于面向过程的方法、约束和契约的基于静态关注的AAF扩展；（iii）基于约束、契约、上下文和事件的动态关注的AAF扩展，包括依赖于上下文感知的发布-订阅样式的支持；（iv）基于工作流的高保证技术，用于验证基于多个静态和动态分析（例如，模型检查，逻辑编程）的体系结构应用程序框架设计、集成和进化。******这项研究将以基于关注的视角增强当前以模型为中心的架构关注，并为模块化和灵活性提供原则性支持，这是开发AAFs的两个基石。它还将通过对加拿大具有战略意义的国家和国际合作提供优秀的培训，支持软件设计和发展方面的软件研究人员、工程师和开发人员，并帮助提供更熟练的、具有全球竞争力的软件专家，提高加拿大软件产业的经济竞争力。

The construction of modular and flexible software systems by assembling components is one of the central problems in software engineering. To address this problem, many different approaches such as object-oriented frameworks and product line architectures have been proposed as paradigms in which to create and extend software systems. These paradigms support architectural applications families (AAFs) that rely on a core system shared by all applications where each application differs by their variable extensions.******However, while these architectural application families are generic, there is a wide range of mechanisms that support their diverse variability. Indeed, the extensions supported by these and other approaches can be based on many different concern-based composition mechanisms such as aspects, object-oriented views, patterns, higher-order polymorphism, connectors and wrappers. These extensions lead to different module notions, but the relationship among these approaches is currently not well understood. Composing and integrating static and dynamic concern-based extensions (i.e., at design and runtime, respectively) is a complex task for which failures from internal violations or undesirable interactions may be at last revealed only as costly accidents. Further, the concern-based extensions need to be governed by guarantees involving constraints and contracts, and verification methods are needed to ensure that the concern-based extensions behave as expected.******The general goal of this research project is to define and evaluate a new flexible and modular approach to static and dynamic concern-based extensions in the design and evolution of high-integrity architectural application families. The research focuses on (i) fundamental problems in concern-based modularity and extensions; (ii) static concern-based AAF extensions based on process-oriented methods, constraints and contracts; (iii) dynamic concern-based AAF extensions based on constraints, contracts, context and events, including support relying on the context-aware publish-subscribe style; and (iv) high-assurance techniques based on workflows for verifying architectural application framework design, integration and evolution based on multiple static and dynamic analyses (e.g., model checking, logic programming).******This research will augment the current model-centric focus on architectures with a concern-based perspective and provide principled support for modularity and flexibility, two cornerstones of developing AAFs. It will also offer excellent training through national and international collaborations in an area that is strategic to Canada, support software researchers, engineers and developers in software design and evolution, and help to provide more skilled, globally competitive software experts and increase the economic competitiveness of the Canadian software industry.