Requirements Management Through SDLC - Requirements Engineering and Management for Software Development Projects - page 167

objectives are met. The test strategy is normally captured in the software quality
assurance plan.
In test plans we determine the resources required to implement the test strategy,
the types of tests to be carried out and the schedule of testing.
Test cases implement the test plans and draw up details of how the product is
tested. A requirement normally results in multiple test cases and rarely, in only one
test case.
Testing consists of designing the test cases and then executing those test cases
on the software product. While designing test cases, we need to ensure that all
requirements of all stakeholders are covered. Often times, it is not possible to
cover all requirements in the designed test cases. The reason in most cases is either
the paucity of time or funds to test the product thoroughly. Therefore, we compute
a metric referred to as the ''Test Coverage'' metric. The formula for test coverage
metric is:
TCM = (Number of requirements covered by test cases 7 Total number
of requirements as established in the requirements traceability matrix or the
URS) 3 100
Where TCM = Test Coverage metric.
TCM is usually expressed as a percentage. While there is no standard per-
centage for TCM to determine the adequacy, 90 % is usually treated as good
coverage. By this statement, do not mistake me that I condone less than 100 %
coverage! Far from it! Test coverage of 100 % is the best practice and I strongly
advocate it.
In the testing phase, from the standpoint of requirements management, we
ensure that all requirements are properly implemented in the product and update
the traceability matrix with references of the test cases and test logs against each of
the requirements in the matrix.

12.7 Acceptance Testing

While acceptance testing is also part of testing, it is separately handled as it is
conducted by the customer. Acceptance testing is dedicated to proving that all
requirements are indeed implemented in the software product. So, we need to:

1. Ensure that the acceptance test plan and test cases do cover all the

requirements.
2. We need to conduct the test ensuring that all requirements are implemented in
the final software product.
3. The product works without defects when used positively.

First we need to verify and validate that the acceptance test plan and test cases
cover all the requirements comprehensively. Often, the acceptance testing
becomes a formality before accepting the delivery of the software product. True,
the acceptance testing is positive testing not intended to uncover defects but it is