Analysis of Early Aspects in Requirements Goal Models: A Concept-Driven Approach - Transactions on Aspect-Oriented Software Development - page 66

data analysis will lack statistical significance and credibility. However, qualita-
tive analysis methods, such as categorizing and coding used in our evaluation,
can give an initial reaction to how such a systematic approach to early aspects
is considered by requirements analysts.
Based on the quality criteria provided by diffusion theory, five free-form
questions were designed. We collected data through interviews and question-
naires [21], and used coding (relating answer sections to proper quality attributes
under testing) and categorizing (classifying answers to be positive or nega-
tive) [37] to perform qualitative data analysis. Requirements analysts' quotes
are represented italic and cited in double quotation marks (“ ”) when analyzing
the following five questions.
Question 1: According to your experience, do you think that this approach
provides sucient constructs and guidelines to be tested on a limited basis before
adoption?
Quality Attribute: Triability.
The answer to this question was quite positive. Most requirements analysts
were familiar with the notions of softgoals and tasks, and supported the idea
of using contribution links to uncover and analyze goal models' crosscutting
properties. To quote from feedback provided by an analysts: “. . . (for one of our
projects), we're currently conducting interviews (with multiple stakeholder roles),
and would like to try out the (proposed) idea for separating concerns . . . ”
Question 2: Do you see preliminary observable results from the application of
the proposed approach to the analysis of crosscutting concerns in goal-oriented
requirements engineering?
Quality Attribute: Observability.
The response to this question was very positive. Analysts found the results in
the given media shop example “ non-trivial ”, “ insightful ”, and “ useful ”.
Question 3: Compared to relevant techniques you are aware of, do you think
that the adoption of the proposed approach can better help you improve the
quality of the goal-oriented requirements analysis?
Quality Attribute: Relative advantage.
Most analysts did not provide answers to this question. Instead, some men-
tioned that “ there's yet requirements method (that I'm aware of) that focuses on
crosscutting concerns (explicitly). ”
Question 4: Do you think that the proposed approach is overly complex to be
understood and used?
Quality Attribute: Complexity.
The answers to this question diverged. Some analysts regarded the approach
proposed was easy to apply since no new modeling notations were introduced. Oth-
ers admitted that even though the concept lattice offered interesting analyses, the
line diagram produced for any sizeable model “ was somewhat too cluttered and