Natural naming in software development and maintenance: Dissertation

The understandability of source programs and other types of software documents is important for several reasons. Software developers have to read documents written by their colleagues, and software maintainers often need to study old source programs about which they have no previous knowledge. Naming is one important factor that affects how understandable source programs are. In general, natural naming means avoiding abbreviations in software documentation. In the context of source programs, natural naming means that program elements such as variables, constants, tables, and functions should be named using whole natural words of a natural language with respect to the grammatical rules of the same natural language. This thesis introduces methods and tools to facilitate the use of natural naming in software development and maintenance. To support the use of natural naming in programming, source program elements are classified and specific naming rules provided for different program elements. An analytical name creation method is provided to make natural names in source programs consistent with written text in other types of software documents. Commonly used programming languages do not require any specific naming rules to be followed. For this reason, an experimental programming language is introduced in this thesis. The language is designed to support the use of natural naming. Existing source programs usually contain many abbreviated names. To make existing programs more maintainable, intelligent disabbreviation tools have been developed as part of this study. Disabbreviation means replacing abbreviated names with natural ones in existing source programs. The naming methods and tools have been evaluated by testing them in laboratory experiments and in practical software development and maintenance situations. The natural naming principles have been taught to software developers in different organizations. According to feedback received from users and the data collected in the experiments, natural naming is a promising approach to increase the understandability of software documents, and the methods and tools introduced in this thesis facilitate the use of this naming approach.