This session measures how well subjects do applying the PBR2 technique
for finding faults in requirements documents. This information can
then be compared against the ad hoc baseline measured in session 1A to
detect any change in effectiveness due to PBR, and also used as a
baseline to gauge the effects of error abstraction applied in the next
Subjects were assigned to one of the three PBR2 perspectives as
described in session 1C.
We estimated this session to take about 2-3 hours, but in actuality
our subjects spent 3.7 hours on average (with a number of extreme
cases taking between 6 and 8 hours - though this is probably because
our subjects thought they would be graded on the number of faults they
Description and Training
- We provided an overview of the session. This handout also
contains some instructions to the subjects which you will need to
change for your own environment - we allowed our subjects to begin the
work in a controlled environment in which they could ask questions
about the techniques as needed, but allowed the work to be finished as
a homework exercise.
- We also gave a lecture in which we presented the technique for
Each lecture also contains a short application of each technique to an example
requirements document (the video store example).
- We gave a handout containing the detailed technique for each
perspective to the appropriate subjects.
Data Collection and Questionnaires
- Form 2.1 was provided as a template for submitting the fault list
and a questionnaire.
- We also gave appropriate forms to each subject to help them create
the model of the system that their technique demanded. These forms
are referenced by the detailed handout.