SoD-HCER: Collaborative Research: Using Market Forces to Improve the Design of Software

可互换部件的市场给制造商带来了改进产品的压力。如果市场上有关部件质量的信息传播迅速，这一点尤其正确。但是现有的软件组件市场缺乏质量保证机制。在这个探索性项目中，pi将研究软件市场、质量保证和设计之间的关系。pi推测，如果软件组件市场配备了统一强制的质量控制机制，它将迫使软件生产者不断地检查他们的产品设计。为了验证这一假设，pi将在小型可控组件市场进行实验，其中产品带有明确的软件合同，底层软件平台监控这些合同并公开报告违规行为。生产者可以选择用技术贡献（测试用例、改进的合同等）对这些报告作出反应。在这种情况下，pi将通过执行代码检查和与制作人面谈来验证他们的猜测。这一探索性调查的结果将有助于其他国家构建类似的市场和机制；有些人甚至可以直接从作为项目一部分开发的软件中受益。更广泛的影响：这个项目有可能改变软件组件的生产方式和传授软件重用的方式。如果pi关于市场的猜想得到证实，它将为公共市场和组织内部市场的推动者提供论据。具有质量保证机制的市场的创建必然会使软件的生产更有效，并提高在市场上交易的组件的质量。不管实验结果如何，利用市场进行课程的想法可以帮助教师在课堂上进行软件重用的教学。它们可以利用pi的经验，或许还可以利用它们的软件，创建以课程为中心的市场；他们可以利用这些市场来展示软件重用的价值。

Markets for interchangeable components exert pressure on manufacturers to improve their products. This is especially true if the market comes with fast dissemination of information about the quality of the components. But existing markets for software components lack quality assurance mechanisms. In this exploratory project, the PIs will study the relationship between software markets, quality assurance, and design. The PIs conjecture that if a market for software components is equipped with a uniformly enforced quality control mechanism, it will force software producers to continuously overhaul their product designs. To test this hypothesis, the PIs will conduct experiments with small controllable component markets, in which products come with explicit software contracts and the underlying software platforms monitor these contracts and publicly report violations. Producers may choose to react to these reports with technical contributions (test cases, improved contracts, etc). In this context, the PIs will validate their conjecture by conducting code inspections and interviews with producers. The outcomes of this exploratory investigation will help others construct similar markets and mechanisms; some may even directly benefit from the software developed as part of the project.Broader Impacts: This project has the potential to change the way software components are produced and the way software reuse is taught. If the PIs' conjecture about markets is confirmed, it will provide arguments to the promoters of both public markets and markets that are internal to organizations. The very creation of a market with quality assurance mechanisms is bound to make the production of software more efficient, and to improve the quality of the components that are traded on the market. Independently of the outcome of the experiments, the idea of using a market for courses may help instructors deal with the teaching of software reuse in classes. They can use the PIs' experience and possibly their software as well to create course-focused markets; and they can use these markets to demonstrate the value of software reuse.