Research And Implementation Of A Software Development Process Quantification System Based On Change Analysis

The continuous development of software development technologies has driven the development of software measurement technologies and software management technologies.The development of modern software is mainly based on the version control system and the team is the main body.Version control system can manage the version change history in the software development process,and save a lot of basic change data,which is convenient for later analysis and statistics and query.In previous software projects,the quantitative management of the development process was mostly carried out by the manager based on the actual progress of the project.Due to the strong subjective consciousness and lack of data support,the results can not accurately reflect the actual work of people or organizations.There are some management techniques and software measurement models,many of which are presented in combination with the development of the software industry at that time,and it is difficult to adapt to the continuous development of software development technology.Based on this,this paper studies the quantitative methods of software development process based on change analysis.The main research contents are as follows.(1)The daily development change data of SVN version management system was captured,and the quantitative management system was developed to manage the change data.The system can update,statistics and browse the change data.In the code statistics module,the data such as the output volume of the code and the type of the change file can be calculated according to different constraint conditions,such as personnel and time period,and displayed in a graphical form.In the change information management module,you can view the file's change record and the detailed change content.(2)Based on the change data,a software quantization model based on change analysis was proposed,including statistical model,output prediction model and development process management model.Through the statistics and analysis of the change content,using multiple linear regression method to establish add methods indices such as number,modify the bug number and code output,the relationship between the generated output prediction model.Based on the COCOMOII model,thesoftware scale estimation and schedule management method based on the change analysis are described.Finally,the Tomcat version library was selected for the case analysis,and the linear relationship between the output volume of the code was established through the statistics of various measurement parameters.Then the output quantity of the code in the next stage was predicted,and the prediction results were compared with the actual data captured,and the prediction model was verified.