Software Quality Grades for MI Software

The data provides a summary of the state of the practice for Medical Imaging (MI) software (as of August 2021). The summary is based on grading a set of 29 MI products using a template of 103 questions based on 9 software qualities. The raw data obtained by measuring each software product and the template used to grade the software are found in the SoftwareGrading-MI.xlsx file. We use the first section of the template to collect general information, such as the name, purpose, platform, programming language, publications about the software, the first release and the most recent change date, website, and source code repository of the product, etc. Information in this section helps us understand the projects better and may be helpful for further analysis, but it does not directly affect the grading scores. We designed the following nine sections in the template for the nine software qualities. For each quality, we ask several questions and the typical answers are among the collection of "yes'', "no'', "n/a'', "unclear'', a number, a string, a date, a set of strings, etc. Each quality needs an overall score between 1 and 10 based on all the previous questions. All the last three sections are about the empirical measurements. For some qualities, the empirical measurements also affect the score. We use tools to extract information from the source code repositories. For projects held on GitHub, we manually collect additional metrics, such as the stars of the GitHub repository, and the numbers of open and closed pull requests.本数据集汇总了截至2021年8月的医学影像（Medical Imaging, MI）软件行业实践现状。该汇总基于使用包含103个问题的评分模板，对29款医学影像软件产品开展评级所得。用于对每款软件进行评测的原始测量数据以及本次评分所使用的模板，均收录于SoftwareGrading-MI.xlsx文件中。 我们使用模板的第一部分收集软件的通用信息，包括产品名称、用途、运行平台、编程语言、相关学术文献、首次发布时间与最近更新日期、官方网站以及源代码仓库地址等。该部分信息有助于我们更深入地理解相关项目，可为后续分析提供支撑，但不会直接影响最终评分结果。 我们针对上述9项软件质量维度，在模板中设计了对应的9个评分章节。针对每个质量维度，我们设置了若干问题，典型回答类型包括“是”“否”“不适用”“不明确”、数值、字符串、日期、字符串集合等。需基于对应章节下的全部问题，为每项质量维度给出1至10分的综合评分。 本次模板的最后三个章节均围绕实证评测展开。对于部分质量维度，实证评测结果也会对最终评分产生影响。我们通过工具从源代码仓库中提取相关信息；对于托管于GitHub的项目，我们还会手动收集额外指标，例如GitHub仓库的星标数量、已开启与已关闭的拉取请求（Pull Request）数量。