A software metric is a standard way of measuring a characteristic of software development. Measurement data are numerical data collected and recorded in the units of measure, defined for the metric. For example, if the metric is the number of hidden errors per thousand of lines of code ( KLOC), then the measurement data set might look like this: {Program A: 2.6 errors / KLOC; Program B: 12.1 errors / KLOC; Program C: 5.8 errors / KLOC}.
The main advantages of applying software metrics are as follows:
- The possibility of using the original comparative data in conjunction with other development projects.
- The ability to collect state data expressed in the same units throughout the organization to determine the success of the task at hand.
- The ability to predict the comparative effort.
- The possibility of estimating labor input or a work schedule, proceeding from the available development experience.
- The possibility to compile data reports, during formal reviews, characterizing the progress of a development process.
The measurement data of the program must be received or converted into numerical units of measure determined for the program metric. These data must characterize the attributes of the software process, its results or timings or resources of the development project. Outsourcing testing to third party partners, like Ukrainian service providers, you do not only save on costs but also build better digital products at the right schedule.
First of all, it should be understood that there are thousands of possible program metrics. Measuring most of them is a methodical and time-consuming process, and they may prove to be economically inefficient for one particular development group. In general, choosing the metrics from the list can be akin to walking across the minefield, unless you approach the process of determining and using these metrics very pragmatically. In this sense, testing metrics are not an exception. In the next article of ours you may get acquainted with a practical example that represents the point of view of one of the authors and illustrates his approach to determining metrics that meet the needs of the organization and does not become a burden for one of the projects within the organization. In this case, the term “organization” is used for several software development projects carried out under the same guidance (for example, in some subdivision of a larger company).
