Analysis of Early Aspects in Requirements Goal Models: A Concept-Driven Approach - Transactions on Aspect-Oriented Software Development - page 53

Fig. 9. Cluster analysis for the repertory grid in Fig. 8
Comparison : After focused extraction of repertory grids from goal models and
structured exchange of tasks between stakeholders, stakeholders' softgoals can
now be compared. The set of all tasks offers a common ground to evaluate
softgoals in a particular context. We compare the softgoals according to the
extent to which they array the set of tasks.
The relationship between any two constructs can be examined by seeing to
what extent the ratings of all the elements on one of the constructs tend to
match, or differ from, the ratings on the other construct [39]. If two softgoals
orchestrate the tasks in the same or very similar way, “correspondence” (Fig. 4b)
is established between these constructs even though they may be labeled differ-
ently. If two softgoals that have been labeled using the same terms align the
tasks in a markedly dissimilar fashion, then “conflict” (Fig. 4b) is detected.
Figure 9 presents the cluster analysis result for the repertory grid shown in
Fig. 8. It is apparent from Fig. 9 that the terms used to express softgoals in-
terfered greatly. For example, what the customer meant by “Security” and the
concern “Confidentiality” in the developer's viewpoint were associated at the
96.2% level. In terms of the tasks shown in the grid, these two softgoals differ only
in the degrees, not in the poles' extremity, rated by three elements: one rated two