一句话:提出可塑界面以支持软件文档评审,让技术写作者用自然语言定制视图,并按角色呈现信息。
该论文关注软件文档评审中被忽视的技术写作者。通过对31名技术写作者的访谈,作者发现文档质量依赖多阶段评审:不同专长的参与者依据不同质量标准评估;但现有流程缺乏专用工具,团队常使用 GitHub 等面向开发者的平台,非技术评审者只能借助其他工具,导致反馈分散。论文提出面向文档评审的可塑界面(malleable interfaces):从业者用自然语言描述界面需求,系统生成适配不同专长并突出角色相关信息的视图。评估计划分两项研究:先考察个人如何表述并使用生成的界面,再考察不同角色从业者如何借助这些界面协作。工作旨在推动可由实践者塑造的软件工具愿景。
Research on documentation quality has examined developers and software artifacts but overlooked technical writers, the dedicated practitioners who synthesize distributed software knowledge into documentation. By interviewing 31 technical writers, we found that documentation quality is facilitated by a review process in which practitioners with different expertise evaluate against distinct quality criteria over multiple stages. This process lacks dedicated tooling, forcing documentation teams to adopt developer-focused platforms like GitHub and leaving non-technical reviewers to resort to alternative tools, resulting in feedback scattered across platforms. This thesis proposes malleable interfaces for documentation review, where practitioners reshape interfaces to fit their work rather than adapting to fixed tool interfaces. We propose to build a review environment where practitioners describe interface needs in natural language, and the system generates views to accommodate diverse expertise and surface information relevant to their roles. We plan two evaluation studies: examining how individuals articulate and use generated interfaces, then how practitioners with different roles collaborate using them. This work contributes toward a vision of software tooling that practitioners can shape to fit their needs.