CAREER:Software Requirements Evolution in a Multi-Jurisdictional Socio-Technical Ecosystem

普及和分布式计算允许工程师在第三方组件、平台和基于云的服务中重用软件，从而减少了开发时间。因此，该软件受制于多个政策和法规，这些政策和法规对这些复杂系统的行为施加了法律要求。当开发人员推出新产品特性，进入跨越地缘政治边界的新市场，或者当现有法律发生变化或创建新法律时，法律需求会对系统设计产生进化压力。作为回应，软件工程师必须协调法律要求与他们的系统设计，以确保他们的软件符合政策和法律；当创新发生在没有现有法律的情况下，这个问题更具挑战性。本研究旨在通过分析管理软件的法规语料库来解决这个问题：(1)开发一套启发式和半正式的、特定于领域的语言，用于表达和推理法律需求，以确定需求覆盖范围；(2)实证地衡量不同司法管辖区的政策和法规之间的差距，这些差距表明了需求的权衡、趋势和由于需求变化而造成的潜在破坏；(3)使开发人员能够根据变化覆盖的模型来合理化和选择可选择的需求演进。其成果包括解释和预测跨司法管辖区需求演变的新理论，以及监管机构、法律专业人士和软件工程师可以使用的工具和技术，以减轻应对全球不断变化的监管格局的负担。这些结果将使用混合方法研究进行评估，该研究结合了正式方法、信息检索和人体受试者实验，旨在进一步了解专业人员如何表达和解释需求，以及他们如何在存在歧义和相互冲突的业务和监管目标的情况下协调相互冲突的需求。除了培训和教育之外，这项研究的更广泛的影响旨在协调软件系统的监管目标，参与软件工程和法律的专业，这些专业在历史上是分开工作的，并告知政策制定者和立法者关于法规对软件设计和开发的影响。

Pervasive and distributed computing decreases development time by allowing engineers to reuse software in third-party components, platforms and cloud-based services. Consequently, this software is subject to multiple policies and regulations that impose legal requirements on the behavior of these complex systems. Legal requirements create evolutionary pressure on system design as developers roll out new product features, enter new markets that cross geo-political boundaries, or when existing laws change or new laws are created. In response, software engineers must reconcile legal requirements with their system design to ensure their software complies with policy and law; a problem even more challenging when innovation occurs in the absence of existing law.This research aims to address this problem by analyzing corpora of regulations that govern software: (1) to develop a set of heuristics and semi-formal, domain-specific languages needed to express and reason about legal requirements for the purpose of determining requirements coverage; (2) to empirically measure gaps among policies and regulations from different jurisdictions that indicate requirements trade-offs, trends and potential disruptions due to changing requirements; and (3) to enable developers to rationalize and select alternate requirements evolutions based on models of changing coverage. The outcomes include new theory to explain and predict requirements evolution across jurisdictions, and tools and techniques that regulators, legal professionals and software engineers can use to reduce the burden of responding to a globally evolving regulatory landscape. These outcomes will be evaluated using mixed-methods research that combines formal methods, information retrieval and human subject experiments aimed at furthering our understanding of how professionals express and interpret requirements and how they reconcile conflicting requirements in the presence of ambiguity and conflicting business and regulatory goals. In addition to training and education, the broader impact of this research aims to harmonize regulatory goals with software systems, to engage the professions of software engineering and law that have historically worked separately, and to inform policy and lawmakers about the impact of regulations on software design and development.