| 主权项 |
1. A method performed by a computing system, comprising:
receiving, by a code analyzer, program code for code analysis, wherein the received program code has no annotations for the code analysis; generating, by the code analyzer and as a result of the code analysis, a list of specifications, and a list of defects without requiring modification of the program code; computing, by the code analyzer, multiple metrics for the program code based on the list of defects and specifications, wherein the metrics include an evidence metric, a counter-evidence metric, and an importance metric, the computing including computing at least a first metric pair and a second metric pair for each of multiple components of the program code, wherein at least one of the metrics is common between the first metric pair and the second metric pair; generating multiple quality maps using the metrics to denote a quality of the program code, wherein the generating includes classifying each of the multiple components to one of multiple portions of:
a first quality map of the multiple quality maps based on the first metric pair, anda second quality map of the multiple quality maps based on the second metric pair, wherein a first portion of the multiple portions of the first quality map corresponds to components of the program code whose quality is below a first specified threshold and a second portion of the multiple portions of the second quality map corresponds to components of the program code whose quality is below a second specified threshold; and identifying a set of the components that is classified to the first portion or the second portion for improving the quality of the program code. |
