| 摘要 |
A system of measuring software significance for reusing the software based on a certain objective metric is provided. In this system, in inter-file relationship extraction (S212), Java source code files 230 are analyzed, and inter-class inheritance is extracted as reuse relationships. In addition, in SMMT (S222), similarity among Java source code files 230 is calculated. In cluster analysis (S224), cluster analysis is carried out based on the similarity obtained by the SMMT (S222), and a set of files is categorized into M component groups. Relationships among component groups are extracted (S214) from the results of the cluster analysis (S224) and extraction of inter-file relationships (S212). Evaluation is carried out based on the values of relative significance for component groups using the inter-component group relationships (S216). Finally, the values of relative significance for component groups are converted to file (software component) evaluated values (S218).
|
