Cyclomatic complexity is a number used to express the complexity of source code (e.g. of a method). It is calculated based on the number of different possible paths through the source code. Low cyclomatic complexity is one factor to improve readability, maintainability, and testability of code.
The Cyclomatic Complexity (or conditional complexity) graphs theoretical software metric to measure the complexity of a program (or method, function, procedure ..). It was defined by Thomas J. McCabe in 1976.
Is calculated using the following formula:
V(G) = P + 1
where V(G) = Cyclomatic Complexity
P = number of decision predicate nodes (if, else, while, etc)
I.e. initially the cyclomatic complexity of each method is defined as 1. Each control flow instruction adds 1 to the complexity number of a method.
Since there is some correlation between good code quality and a low number of complexity, methods with a high Cyclomatic Complexity (e.g. 10 to 15) are considered hard to read and understand. Low number of Cyclomatic Complexity is a good indicator for readable, (re-)usable, reliabe and functional coding style. Because of this, it is recommended to split such methods with a high complexity into smaller ones.
For measuring the needed number of unit tests to get full code coverage it's better to use the Extended Cyclomatic Complexity or the NPath Complexity. Therefore McCabb’s Cyclomatic Complexity is not accurate enough, because it fails to distinguish between different conditional statements (control flow structures). It also does not consider nesting level of various control flow structures.
