Testing Terms

Traceability Matrix

In a large testing project, it is easy to lose track of what has been tested and what should be tested. The question often arises, "Is the testing comprehensive enough?"

A simple way of determining what to test is to go through the source documents (Business Requirements, Functional Specifications, System Design Document, etc.) paragraph by paragraph and extract each requirement. A simple matrix is built, with the following format:

Source Document	Section	Requirement	Test Case
Func Spec	3.1	An account number must be entered, of 8 digits and a 1 digit checksum. Invalid account numbers must be diagnosed with an appropriate error message.	147
Func Spec	3.1	The account number must exist in the Customer Database; if not, an appropriate error message must be displayed.	148
Func Spec	3.2	Given a valid account number, the customer details must be retrieved from the Customer Database and displayed on the CD-102 screen form.	149

Basically, the Traceability Matrix relates each functional requirement to a specific test case. Thus if someone says, "Did we test that the account numbers must be valid?", the matrix indicates which test case does that test.

The Traceability Matrix is created before any test cases are written, because it is a complete list of what has to be tested.

Sometimes there is one test case for each requirement; other times, several requirements can be validated by one longer test case.

Frequently, there are NO business requirements available; although this is a major mistake (as no one has documented what the business needs), such is life. In this case there won't be a source reference document, so there is no "traceability"; the matrix is then simply called a "Test Matrix".