Model-Driven Engineering of Integrated Digital Infrastructures

拟议的研究计划旨在为开发复杂的软件组件生态系统的理论和实践做出贡献，以支持从事基于知识的工作的个人团队。当今所有基于知识的工作都涉及使用软件、使用传感器和数字仪器、开发软件模拟以及交叉引用和分析不同类型的数字材料，以提取用于复杂问题解决和决策任务的信息。软件工程研究为这一挑战提供了三种主要的技术和方法解决方案：(A)面向服务和虚拟化，用于将复杂系统构建为可扩展的软件组件不断演变的聚合；(B)模型驱动工程(MDE)，用于通过模型转换链进行系统软件构建；以及(C)用于有效的分布式团队的工具和流程。认识到这种知识工作者需求和软件工程能力的协调中的巨大机遇，我计划了一个研究计划，旨在通过集成数字基础设施的模型驱动工程来支持知识工作者的跨学科团队。我使用术语“集成数字基础设施”是指由独立开发的软件组件和智能设备组成的不断发展的生态系统，这些组件和智能设备可以灵活地协调以支持专家团队的复杂活动。通过采用核心MDE方法，拟议的计划将回答有关如何支持MDE软件的演变和维护的基本研究问题，以便手动后生成代码改进可以传播回代码构造链中的相关模型。接下来，我们将通过将MDE方法与体现领域特定语义的不同类型的模型进行集成来应用MDE方法(并进行压力测试)，以开发(A)集成传感器和智能设备的软件系统的高级编程模型，以及(B)一种新型的具有社会性的、RDF中介的资产存储库。基本的假设是，模型转换过程应该依赖于被操纵的模型的特定于领域的语义，以支持构建的软件与其需求的一致性。为了使用户与底层系统的交互系统化，我们计划开发一个通用的监控和协调服务，通过底层数据模型通知并与个性化应用程序集成，使用户能够在任何时间通过任何设备与基础设施和彼此交互。最后，将分析监控服务捕获的基础设施的使用记录，以深入了解数据、团队实践及其演变。这些活动的潜在影响是巨大的。(1)提高我们维护模型驱动的已构建软件与其底层模型和转换规则保持同步的能力，即使它随着新的需求而发展，对于采用这种承诺改进代码质量和提高生产率的方法来说，是至关重要的。(2)同样重要的是，将数字基础设施的建设系统化，并提高其效用和可用性，有望大幅提高知识工作者团队的生产力，并推进我们的数字经济议程。(3)最后，选择Smart-CondoTM项目作为对拟议活动结果进行实证真实世界验证的试验台的动机是，希望将软件用于为患者和护理他们的临床医生服务，并开发一个以患者为中心的综合保健平台。

The proposed research program aims to make contributions to the theory and practice of developing complex ecosystems of software components to support teams of individuals engaged in knowledge-based work.All knowledge-based work today across disciplines involves working with software, using sensors and digital instruments, developing software simulations, and cross-referencing and analyzing different types of digital materials to extract information for complex problem-solving and decision-making tasks. Software-engineering research offers three major technical and methodological solutions to this challenge: (a) service orientation and virtualization for building complex systems as scalable evolving aggregations of software components; (b) model-driven engineering (MDE) for systematic software construction through model-transformation chains; and (c) tools and processes for effective distributed teamwork.Recognizing a great opportunity in this alignment of knowledge-workers’ needs and software-engineering competencies, I have planned a research program that aims to support interdisciplinary teams of knowledge workers through model-driven engineering of integrated digital infrastructures. I use the term “integrated digital infrastructures” to mean an evolving ecosystem of independently developed software components and smart devices, flexibly coordinated to support complex activities of expert teams. Adopting a core MDE methodology, the proposed program will answer basic research questions on how to support evolution and maintenance of MDE software, so that manual post-generation code refinements may be propagated back to the relevant models in the code-construction chain. Next, we will apply (and stress test) the MDE methodology by integrating it with different types of models that embody domain-specific semantics in order to develop (a) a high-level programming model for software systems integrating sensors and smart devices, and (b) a new breed of socially aware, RDF-mediated asset repositories. The underlying hypothesis is that model-transformation processes should rely on the domain-specific semantics of the manipulated models, to support the alignment of the constructed software with its requirements. Aiming at systematizing the users’ interaction with the underlying system, we plan to develop a general monitoring and coordination service, informed by the underlying data models and integrated with personalizable applications, to enable users to interact with the infrastructure and each other, any time through any device. Finally, the usage record of the infrastructure, as captured by the monitoring service, will be analyzed to gain insights on the data, the team practices and their evolution. The potential impact of these activities is substantial. (1) Improving our ability to maintain model-driven constructed software in sync with its underlying models and transformation rules, even as it evolves with new requirements, is essential for the adoption of this methodology, which promises improvements in code quality and increased productivity. (2) Equally importantly, systematizing the construction and improving the utility and usability of digital infrastructures promises to substantially improve the productivity of knowledge-worker teams, and advance our digital-economy agenda. (3) Finally, the choice of the Smart-CondoTM project as the testbed for the empirical real-world validation of the results of the proposed activities was motivated by the desire to put software in the service of patients and the clinicians who care for them and to develop a comprehensive platform for patient-centric healthcare.