Fostering the Adoption of i by Practitioners: Some Challenges and Research Directions - Intentional Perspectives on Information Systems Engineering - page 184

• Tool-support is essential. Fundamental capabilities for applying the operations

above, for managing a catalogue of modules and for providing views of the model
based in the modules are needed.

Visual representation. Being i ∗ a notation with a strong emphasis on the visual
dimension, an informed decision about the visual representation of modules
needs to be made. Work by Moody et al. [ 46] provides an excellent rationale
for making this decision.

•

5 Challenge 4: Use i ∗ Models in Later Development Phases

Context. As an intentional modelling vehicle, the i ∗ framework is used in early
stages of the system development. Some of the concepts that appear in i ∗ models
will pervade in later stages, e.g. some i ∗ actors will act as such in use cases, some
resources will appear also in conceptual data models, some tasks will become activ-
ities in a behaviour diagram, etc. This suggests for the need of having systematic
ways to transform i ∗ models into other formalisms.
Problem. The transformation of goal models into more elaborate artefacts that
appear later in the life-cycle has been tackled in several works (e.g., using the MAP
approach to derive data-flow diagrams from goal models [ 51] ). When trying to trans-
form an i ∗ model into some other kind of model some difficulties arise. Typically
the target of this transformation is a UML conceptual model [ 65] , composed at least
of a use case specification, a data conceptual model in the form of a class diagram,
and perhaps some behavioural model. Sometimes just one of these artefacts is the
target.

• Use cases are textual artefacts that reflect communication between actors in a

sequential form. The problems that arise are: (1) identifying the appropriate use
cases from the i ∗ model and also the relevant scenarios; (2) identifying the actors
that take a part in each use case; (3) inferring the interactions between these actors
and write them in the correct order; (4) generating the text itself.

• Data conceptual models are diagrams that include accurate and complete

information about classes or entities, their relationships and their attributes.
Discovering all of these elements from the i ∗ model is also a problem since the
information that it encloses is not as complete as in data conceptual models (due
to its intentional nature).

• Behavioural models like activity diagrams or sequence diagrams include interac-

tions among actors, or activities to be performed, with a flow of control that is
not expressed in i ∗ models.

Solving these problems can be considered a major challenge.
Challenge. There shall be techniques available to make easier (and automate up
to a given extent) the transition from i ∗ models into other types of models.