作为代码分析和理解的重要内容，代码标识符及其归一化是国际学术界的前沿热点研究领域。标识符归一化旨在将标识符解析成自然语言词汇，以提高代码的可理解性和可维护性。现有的标识符归一化算法的性能有待提升，标识符及其归一化的演化规律尚未经过系统性归纳，严重影响了标识符归一化提升相关软件开发任务的效果。本项目针对现有工作的不足，围绕标识符归一化的算法设计、数据分析、应用，凝练出三个层面的研究内容：1）在算法设计层面，利用标识符的通用性和独特性提升标识符归一化的算法效果；2）在数据分析层面，归纳标识符及其归一化在软件代码库的不同版本和不同粒度的演化规律；3）在应用方面，利用标识符归一化提升典型的软件开发任务的效果。针对这些挑战性的研究内容，本项目结合现有的研究基础，制定了切实可行的研究方案。本项目有望显著提升标识符归一化的算法性能，归纳标识符归一化的演化规律，拓展标识符归一化的研究范畴。

As an important research content of source code analysis and comprehension, identifier normalization is the leading field of the current research of software engineering. Identifier normalization aims to parse identifiers into natural language terms so as to improve the understandability and maintainability of source code. The performance of existing identifier normalization algorithms needs to be improved. In addition, existing studies do not systematically analyze the evolution patterns of identifiers with their normalization, which seriously affects the results of some software development tasks by leveraging identifier normalization. Targeted towards tackling the shortcomings of existing studies, this project issues a series of research contents to resolve, analyze, and apply identifier normalization. These research contents can be divided into three levels, namely 1) the level of algorithm design – how to boost the performance of identifier normalization algorithms by leveraging the generality and uniqueness of identifiers, 2) the level of data analysis – how to analyze the evolution patterns of identifiers with their normalization along with different versions and granularities of source code repositories, 3) the level of application – how to improve the results of typical software development tasks by leveraging identifiers with their normalization. To resolve these challenging research contents, this project develops a feasible research program based on existing work. The project can effectively enhance the current work on identifier normalization, summarize the evaluation patterns of identifier normalization, and extend the research area of identifier normalization.