Research On Software Traceability Method Based On Version Control

With the increasing complexity of software systems,the maintenance of software traceability becomes more and more important.Due to the increasing speed of iterations of software development processes and the increasing mobility of developers,it becomes harder for developers and maintenance staffs to judge whether the requirement has been implemented and whether the source code has been tested.Therefore,the traceability relationship between software artifacts is becoming increasingly important during software development and maintenance.Once the software artifacts change,developers can determine the scope and impact of these changes through these traceability.At present,the common problem in the field of software traceability is that the establishment and maintenance of the traceability relationship is difficult and the result of the traceability is poor.On account of the above analysis,this thesis combines dynamic traceability method with version control mechanism,and proposes a software traceability method,which include semantic analysis and domain ontology-related knowledge based on version control.Combined with the version control mechanism,this method can solve the problem of inconsistent language between software artifacts.And the introduction of semantic analysis and ontology-related knowledge can solve the problem of unexplored lexical implicit semantics.The main contents of this paper are as follows:(1)Aiming at the low accuracy of traceability methods between Chinese design documents and source codes,a traceability method based on version control is proposed.Firstly,a set of semantic analysis rules and heuristic rules between text and source code are defined,and the relevant data acquisition and preprocessing methods are given.The similarity score between the Chinese design document and the source code is calculated by the method which combines the similarity calculation method based on information retrieval and the custom rules.Then,the version update information in the version control software is processed to obtain the mapping relationship between the version description statement and the source code.Next,the mapping relationship is used to correct the former similarity score.Finally,the traceability relationship between the Chinese design document and the source code is determined based on the set threshold.(2)As for project stakeholders,in order to understand the reason of traceability links and identify the correctness of traceablity links better,a heuristic rule based method for explaining the traceability links is designed.Firstly,the generation process of the traceability links is analyzed in depth.Then,the relationship between domain concepts and the relationship between action frameworks are analyzed.Next,the heuristic rules for the interpretation of the traceability links are established according to these relationships.Finally,the interpretation statements of the traceability links are automatically generated according to these rules.(3)Aiming at the problems that the speed of the current software system is iteratively faster and the software integration test is not timely or accurate,a method for automatically recommending integrated test cases is designed.First of all,this thesis presents a traceability method between Chinese design documents and integrated test case document based on domain ontology and vector space model.Then,the traceability link between Chinese design documents and integrated test case document is drawed.Once the version of source code changes,the traceability link between design document and source code will be searched according to the changed source code,in order to obtain the corresponding design document statement.Then the design document and integrated test case document traceability links will be searched on the basis of this statement,so that the corresponding integrated test cases can be obtained.The experimental results show that the proposed method is feasible and effective,and the precision and the recall gain improvement compared with the conventional method based on vector space model.When the design document and the source code traceability links are formed,a corresponding traceability links interpretation sentence can be given.And when the version of source code changes,the corresponding integrated test case can be automatically recommended.