Software evolution concern summarization and explanation

加拿大社会已经变得依赖于软件密集型系统，如手机，网上银行和超声波，仅举几例。为了跟上我们不断变化的世界，软件开发人员必须不断地改进这些系统。当进化任务执行不正确时，就会产生后果。例如，加拿大税务局（CCRA）的一项错误完成的演变任务导致2007年3月初暂停在线个人税务申报。由于这个错误，纳税人的退款被推迟，CCRA被迫承担额外的费用来解决这个问题。为了更高质量地完成演化任务，软件开发人员需要有效地沟通预期的变更，并且需要精确而完整地执行与变更相关的技术工作。目前，这些活动是不必要的复杂，因为开发人员正在处理与任务所需的各种信息相关的所有细节。该研究计划旨在通过使开发人员能够根据任务进行工作，而不是让开发人员陷入每个任务的细节中，来提高开发人员的生产力和工作质量。本研究计划将开发技术和工具，以支持表示的片段所需的软件项目信息的演变任务。我们把这种表示称为进化问题。该表示将包括自动生成的摘要和对包含进化关注的信息的解释，提供当前开发人员工作日中缺乏的抽象。我们将评估我们为支持进化问题而构建的技术和工具是否可以通过对开源项目的实验室实验和工业发展背景下的实地研究来缓解开发活动。这项研究将扩大加拿大1000亿美元的软件生产行业。

Canadian society has become dependent upon software intensive systems, such as cellphones, online banking, and ultrasounds to name just a few. To keep up with our changing world, software developers must continually evolve these systems. When evolution tasks are performed incorrectly, there are consequences. For example, an incorrectly completed evolution task at the Canada Revenue Agency (CCRA) resulted in the suspension of on-line personal tax filings in early March 2007. As a result of this error, taxpayer refunds were delayed and the CCRA was forced to incur additional costs to fix the problem. To complete evolution tasks with higher quality, software developers need to communicate efficiently and effectively about the intended change and they need to perform the technical work associated with the change precisely and completely. Currently, these activities are needlessly complicated because developers are working with all of the details associated with all the kinds of information needed for the task. This research program aims to improve developer productivity and the quality of a developer's work by enabling developers to work in terms of tasks rather than having the developers mired in the details of each task. This research program will develop techniques and tools to support the representation of the fragments of software project information needed for an evolution task. We refer to this representation as an evolution concern. The representation will include automatically produced summaries and explanations of the information comprising an evolution concern, providing the abstraction that is currently lacking in a developer's work day. We will evaluate whether the techniques and tools we built to support evolution concerns can ease development activities through laboratory experiments on open-source projects and field studies within the context of industrial developments. This research will expand Canada's $100 billion dollar software production industry.