Software Metrics Enhancement
| Autor: | Pu-Lin Yeh, 葉普霖 |
|---|---|
| Rok vydání: | 2000 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 88 Software developers use measurements as a means to improve their software development and maintenance processes. There exist two categories of software metrics, i.e. static metrics and dynamic metrics. There have been many researches in defining static metrics for software attribute measurements. In static metrics, program source codes (without actually being executed) are manually or automatically analyzed. These static metrics involve in counting the number of operators, operands, lines of code, nodes, branches, def-use pairs, etc. In dynamic metrics, the program should be actually executed in a controlled and systematic way to demonstrate the presence of required attributes. The status data of a program can be collected while it is executed, and then we can figure out the attributes of the program with these data. These metrics are inadvertently affected by including undesirable entities, or by excluding desirable entities. Various inadvertence results of software measurement are identified in this research. The way to deal with reducing the undesirable structures to a standard model is called normalization. In this research, various software normalizations which allow those metrics for what to be measured and, at the same time, avoids the improper results are proposed. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|