Analyzing the Evolution of the Source Code Vocabulary
Analyzing the Evolution of the Source Code Vocabulary
复制标题
DOI：
10.1109/csmr.2009.61
复制
发表时间：
2009-03
期刊：
影响因子：
--
通讯作者：
S. Abebe;S. Haiduc;Andrian Marcus;P. Tonella;G. Antoniol
中科院分区：
文献类型：
--
作者：
S. Abebe;S. Haiduc;Andrian Marcus;P. Tonella;G. Antoniol
源代码是一个混合的软件工件，包含编译器和开发人员的信息。虽然编程语言语法规定了源代码的编写方式，但开发人员在编写标识符和注释方面有很大的自由。这些本质上是有意为之的，并成为开发人员之间的沟通手段。本文的目的是通过对两个软件系统的探索性研究，分析源代码词汇表在进化过程中的变化。具体来说，我们收集数据来回答一系列关于词汇表演变的问题，例如：源代码词汇表的大小是如何随时间演变的？最常见的术语指的是什么？新的标识符是否引入了新的术语？不同类型的标识符和注释之间是否存在共享的术语？标识符类型中的新术语和已删除的术语是否镜像到其他类型的标识符或注释中？
Source code is a mixed software artifact, containing information for both the compiler and the developers. While programming language grammar dictates how the source code is written, developers have a lot of freedom in writing identifiers and comments. These are intentional in nature and become means of communication between developers.The goal of this paper is to analyze how the source code vocabulary changes during evolution, through an exploratory study of two software systems. Specifically, we collected data to answer a set of questions about the vocabulary evolution, such as: How does the size of the source code vocabulary evolve over time? What do most frequent terms refer to? Are new identifiers introducing new terms? Are there terms shared between different types of identifiers and comments? Are new and deleted terms in a type of identifiers mirrored in other types of identifiers or in comments?