Abstract

Software development process is a complex sequence of actions having source code of working system as a result. All project participants should track changes in source code during work process to know what’s happening. However to make «manual» code review everyone should have corresponding technical skills and a lot of time to spend. This work describes usage of automated source code changes classification aimed to control source code evolution. The method bases on statistical clusterization of change metrics. In this work we show usage of automatic classification of changes to optimize code review and code change control on final development stages. Development process report building is also shown.

Highlights

  • THE most important software development project’s asset is its source code

  • In this work it is suggested to use source code changes classification method, allows to automate separation of semantically different changes basing on values of metrics of source code

  • In this work we used set of change metrics, based on three source code metrics defined in table 2

Read more

Summary

INTRODUCTION

Almost all modern software projects keep all history of source code changes in special repository of source code versioning system. This information is available only for that project participants, who has been trained for source code analysis, i.e. mostly for developers. Managers and other specialists are work on project and are interested in information, retrieved from source code in the form of functionality lists for concrete version, different kinds of reports, etc. Automated changes classification gives testers an opportunity to get information about changes, in which new functionality was added, bugs fixed in the form of source code or comments, provided with changes by its developer. This value is on the border between significant and excellent agreement rate of expert and automated classification methods

Usage of changes classification by developer
Usage of changes classification by development team leader
Usage of changes classification by white-box testing team
Usage of changes classification by project manager
CODE CHANGES CLASSIFICATION
Change classification task formalization
Clusters interpretation
CONCLUSION
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call