First of all, I want to say a few common things about software metrics. These metrics can split into two category:
- Product metrics: These numbers are shows the attributes of the product, like number of lines, average number of live variables, cyclomatic complexity, etc. . Product metrics can split into two categories:
- External(qualitative) attributes: Like reliability, security, mobility, user friendly, etc. Not to easy to measure these metrics, because some of them cannot be defined with numbers.
- Internal(quantitative) attributes: Like number of lines, number of comments, etc. Easy to measure these.
- Process metrics: These describe the process of developing, like developing time, number of detected errors in one month, etc.
The usability of these two class of metrics are different. I will describe it later.
As I mentioned before, external attributes are hard to determine. A common method is to detect a connection between external and internal metrics, and by involving statistical technincian make a formula what is using internal metrics and gives one external.
