The centrality of users in the design and development of complex systems, such as service-based applications, calls for new methodologies and techniques to extract and represent user needs and to translate them into real processes. In this short paper, we describe the integration of concepts and analysis techniques of di erent approaches, namely Goal-Oriented Requirements Engineering, User-Centred Design and Process-Oriented Modeling, that are being developed in the context of two projects related to Ambient Assisted Living and Internet of Services.
The central role of the users in the design and evolution of complex systems,
such as service-based applications for the Internet of Services (IOS1) or Ambient
Assisted Living systems, has been widely recognized in the last years [
Goal-oriented Requirements Engineering (GORE) plays a fundamental role in the development of this kind of systems, enabling reasoning about the domain features with the aim of identifying con icts and of checking for validity of functional and non-functional requirements. This technique has been exploited in the ACube project2, whose goal is to study technologies for monitoring complex environments that can be applied in areas such as assisted living homes to help personnel, as well as to support the independence and safety of users.
Moreover, the key for the operationalization of goal-oriented system requirements in terms of services is the de nition of a set of rules for associating process modeling concepts to goal-oriented ones. This enables designers to trace business processes of service-based systems back to intentional elements of the user model (e.g, goals, preferences, roles). The experience in this area comes from the Internet of Service project (IOS3), whose aim is to push the \Internet of Services" for real services, rather than for software services. Studies in di erent areas are conducted within this project, such as the area of service usage, representation, engineering, and delivery.
Nevertheless, we experienced that the two approaches, more than sharing the
language of goals, require an e ective way to center the design on the users of
the system. The aim of this short paper is to propose an integrated methodology
in which goal-oriented analysis [
The strength of the Goal-Oriented techniques in modelling the domains can be still enhanced by coupling the engineering perspective with a creative perspective typical of User-Centred Design approaches. The User Centered Design exploits a series of well-de ned methods and techniques coming from social sciences and psychology for analysis, design, and evaluation technologies. Contextual inquiries, personas and scenarios are widely employed for obtaining a rich picture of the context (organizational, social, physical), and easily communicating it to stakeholders in order to envision acceptable and innovative technological solutions. Additional values emerge from this collaboration at di erent phases of the process.
Our proposal is to integrate Tropos [
The result of this integration is a process that encapsulates activities from
both the two methodologies, promoting a very close collaboration between teams
and an easy exchange of data. The process begins with the investigation of the
domain in order to understand the organizational setting of the domain and
to derive possible needs and services that the system could provide to users.
The contextual inquiry produces a rich collection of data in a narrative format.
The data interpretation phase provides the necessary abstraction to create a
3 A joint Research project at FBK IRST-CIT. http://se.fbk.eu/en/node/15
believable model of the domain, but avoiding to loose important details typical
of a narrative analysis. Tropos early requirement plays a central role at this
step by providing the semi-formal language for describing the domain. The data
consolidation lters data to focus on relevant characteristics: activity scenarios
and personas allow designers to have insight the system, providing an anchor to
the real domain and the real users. Finally, the envisioning phase lets the analysis
team to reason on the system-to-be in order to expand designers' prospective,
to look at the problems from di erent points of view, to gure out how their
ideas can work in a real context, to identify design criticalities, and nally, to
generate requirements. Brainstorming and other creative techniques are used
with the result to shift from the Tropos early requirement phase toward the late
requirement phase, thus obtaining a list of requirements for the system. The
process terminates with the validation of requirements with customers, essential
for evaluating the value of the services the system will provide [
The key for the integration is the use of Scenarios and Personas along the
whole process. Whereas the use of Scenarios in RE is pretty established as an
instrument to describe instances of behavior of the system, Personas | from
social science | is still going to be consolidated in RE. Their conjunct usage may
increase the ability to envision the system [
In our integrated process, scenarios are stories about people (personas)
performing activities; they describe a context in which personas act with the aim of
summarizing, clarifying and reasoning about the collected information. The aim
is the validation with stakeholders and the technical team. They are presented as
narrative or visual stories easy to understand even for non technical people [
Among the possible limits of the approach, there is the need of keeping the huge amount of data, usually collected by contextual inquiry and generated by the scenarios-based design, always synchronized with the Tropos diagrams; this also represents a challenge from a theoretical and practical point of view. In the future work, we intend to re ne the approach and to investigate the requirements of a tool that support a multi-disciplinary team in this respect. 3
A real service is a combination of actual services and software services that
provide electronic access to and monitoring of the actual services [
In real services the user plays a key role: she expresses preferences (e.g., a cheap transportation mean); the service is adapted to her needs (e.g., if she has a meeting when her train should leave, a di erent actual transportation service is used); the real service is composed taking into account her perspective (e.g., the travel and the concert performance are part of the same real service, though in reality they are independent). Moreover, real services are context-aware, which may result in instant changes and timely responses of the user. Hence the user's preferences, needs and decisions guide the composition of real services, and they, in turn, impact and modify the user's assets (e.g., user's money, user's agenda). For example, buying the ticket for the concert, decreases the user's money.
Due to the \user intensive" nature of these systems, our high level purpose
is to stress the centrality of the user along their life cycle. Hence, in this phase,
our goal is to move from the user centred requirements (elicited as described in
the previous Section) to the system design and validation, while preserving the
central role of the user (and her assets). In detail, our work aims at de ning
a modeling framework that integrates the goal-oriented paradigm (speci cally
Tropos modeling methods [
More precisely, the framework rests on the following iterative steps for
designing a real service [