CAREER: Expanding Developers' Usage of Software Tools by Enabling Social Learning

创建和维护软件对软件开发人员来说是一个越来越大的挑战。软件开发工具可以帮助应对这一挑战，但大多数开发人员仅使用可用工具的一小部分，因为他们不知道相关工具的存在。要了解相关工具，先前的研究表明，最有效的方法是社交学习，即开发人员从同伴那里学习工具，但研究也表明社交工具学习很少见。这项研究的目的是提高社交工具学习的频率和有效性，进而从根本上增进我们对技术如何在软件开发及其他领域介导工具学习的理解。为了实现这一目标，我们将帮助改进人们日益依赖的软件。研究人员将创建一个测试平台系统，该系统将记录开发人员工作的连续截屏视频，并使用工具使用情况的时间戳数据进行索引。通过将开发人员的工具使用数据与同事的工具使用数据进行比较，系统会选择开发人员不使用但同事使用的候选工具列表。然后，该系统鼓励同伴通过分享从彼此的截屏视频中选择的工具使用剪辑来互相教学和学习。该方法旨在使分布式和异步开发人员能够高效、有效且频繁地共享工具知识。该测试平台将允许研究人员尝试各种类型的工具学习，以确定不同类型的工具学习如何、为何以及何时有效或无效。

Creating and maintaining software is an increasing challenge for software developers. Software development tools can help meet this challenge, but most developers use only a small subset of the available tools because they are unaware of the existence of relevant tools. To learn about relevant tools, prior work suggests that most effective approach is social learning, where a developer learns about a tool from a peer, yet research also suggests that social tool learning is rare. The aim of this research is to increase the frequency and effectiveness of social tool learning, and in turn, fundamentally advance our understanding of how technology can mediate tool learning in software development and beyond. In reaching this aim, we will help improve the software on which people increasingly rely.The researchers will create a testbed system that will record a continuous screencast of the developer's work, indexed with time-stamped data on tool usage. By comparing the developer's tool usage data with her peers', the system selects a list of candidate tools that the developer does not use, but that her peers do use. The system then encourages peers to teach and learn from one another by sharing tool-usage clips chosen from each others' screencasts. The approach aims to enable distributed and asynchronous developers to share tool knowledge efficiently, effectively, and frequently. The testbed will allow the researchers to experiment with various types of tool learning to determine how, why, and when different types of tool learning are effective or ineffective.