Test Evaluation
Test evaluation is the
quality control
task during which the
test
work products are evaluated.
The typical objectives of the test evaluation task are
to:
- Determine if the deliverable testing work products are:
- Correct.
- Complete.
- Consistent (internally, externally with other work
products, and externally with related conventions).
- Understandable.
- Identify defects in the deliverable testing work products
so that:
- The defects can be fixed.
- Defect trend analysis can be performed to improve the
process and staff training.
- Determine if the
testing tasks are:
- Completed.
- Effective.
- Efficient.
The test evaluation task can typically begin when the
following preconditions hold:
- The
construction phase has started.
- The
test inspection team is adequately:
- Staffed.
- Trained in the test evaluation task.
- Trained in its associated techniques (e.g.,
inspection).
- At least one testing work product is ready for
evaluation.
The test evaluation task is typically complete when the
following postconditions hold:
- The following testing work products have passed their
evaluations (e.g., inspection):
The test evaluation task typically involves the following
teams performing the following steps in an incremental,
iterative, parallel manner:
- The various testing teams informally evaluate their work
products in accordance with the associated technique
used.
- The
test inspection team formally evaluates the work products
in accordance with the associated technique(s) used.
The test evaluation task typically can be performed using
the following techniques:
- Quality Evaluation Techniques:
The test evaluation task typically results in the production
of the following work products:
- Inspecting tends to be the most effective technique for
identifying defects.
- Evaluations can be incremental and are typically
iterative.
- Because defects can occur in the test work products, it
is important to ensure that tests do not produce any false
positive or false negative results.
- The iterative and incremental nature of most development
cycles means that regression testing happens more often and
thus becomes more important. Test evaluation helps ensure
that the repeated tests do not contain defects that will
continually cause test failures.