An Evalution Approach Of Identifier Quality Based On Lexical Rules

The rapid development of the software industry makes the identifier naming rules has been paid more and more attention. Many scholars are making rules from different perspectives. Because different types of identifiers play different roles in software program, the meaning of them are different,too. Thinking about how to combine roles and meaning of identifiers,this paper has come out. It is a in an exploratory study between identifier roles and lexical rules.Firstly, this paper summarizes the natural language part of speech. And then the identifier roles and meanings have been analysised based on their types. Firstly, this paper summarizes the natural language part of speech.And then the identifier roles and meanings have been analysised based on their types.After that, part of speech rules has been draft. In this paper, identifiers are separated into two categories: property set identifier and the method set identifier. While the part of speech rules are divided into three categories: nominal rule, verbal rules and adjective rules.As regards to the rule evaluation, with the advantages of open source software, 55 projects of Java/C/C++ are tested as samples. It can not only verify the practicality of rule set, but also can promote an order of the three languages according to their scores. Using lexical analysis tools, identifiers can be extracted, segmentated and counted. Through the result, vocabulary distribution, the matching degree between two types of identifiers and three of rules,and the scores of each rules can be counted. Finally, the matching order among Java/C/C++ will be given.