Metrics for testing

There are different types of metrics used in software testing to measure different parameters. So, for different purposes, different metrics are used by the developer. For example:

1. The type of metrics used to measure overall testing effort (how much time and effort are required in making software) – Function-Based Metrics.
2. Metric is used to measure several test cases used in a program – Bang Metric.
3. Metric used for basis path testing – Cyclomatic complexity

There is a different method of testing derived from Halstead's theory used to test the effort. Testing can be estimated with this method.

V= Program Volume

PL = Program Level

E = Software science effort (testing effort)

PL = 1/ (n1/2) * (N2/ n2)

E= V/PL

The breadth of testing = Number of requirement tested/ Total Number of requirement

Depth of testing = Independent Basis path covered by testing / total number of basis paths in the program * 100

We can measure the following terms with metrics:-

1. Field bugs
2. Bug arrival rate
3. Performance of software
4. Code churn
5. Counts (lines, files….)
6. Test failures
7. Static analysis findings

Software quality metrics are the results of test executions done on data and programs. Software quality metrics begins with three types of reports:

• Core
• Extra
• Pro tips

Core software reports

These are non- result types of reports resulting from exploratory testing ( focus less on actual/ pass or fail result). It shows the experience of the user. There are two types of core reports:-

• The test run summary
• Defect/ severity and status

Test run summary- A test run summary is a score report that everybody wants to see. It summarizes all testing activities (what is going on form project to project, cycle to cycle, and so many areas). It precisely tells what is going on in an application.

Defect reporting -It is the second report contains three things- status, priority, and severity. A defect report determines the importance of defects and finds out the solutions to the bugs immediately. It also answers about the status of the software.

Software reports also include some extra reports:

• Weekdays report
• Result received per requirement
• Defect density

Weekdays report - This report shows which days are active for testing and allows you to reallocate test resources on that need.

Result received per requirement- It helps specify under and over testing requirements for execution of test and adequately shows which percentage of testing is assigned to each requirement.

Defect density indicates the defect that makes application or functional areas identify risky applications or functional areas.

Pro- Tips for software quality include:

• Manual v/s animated testing
• Days since run
• Flapping

Manual v/s animated testing - Some manual test are moved to automated testing to understand which type of testing produce high value, Some manual test are moved to automated testing to understand which type of testing produce high value, Some manual test are moved to automated testing to understand which type of testing produce high value, Some manual test is moved to automated testing to understand which type of testing produce high value.

Days since run - Software quality is also measured by keeping a track record of the testing software from the starting day, and it will find out which test hasn’t run in an extended period and reduce the library of tests that have not been run recently.

Flapping - Flapping report shows the test pass result under one set of conditions that are always falling under another set. After the test has passed, recheck whether the test case has changed or written correctly or check if there’s a problem with the requirement.