Intentional Requirements Engineering (ERi*c - Engenharia de Requisitos Intencional) is a method that produces i* models. It is a six steps interactive process that uses the notion of “Strategic Dependency Situations” a modularization strategy applied to situations of an organization. ERi*c uses an elicitation strategy that mines goals from the description of the context vocabulary. This paper reports on its use in the context of Investment Contracts - of a Telecom Company. The experience of using ERi*c served as basis for the dissertation of one student at UERJ (State University of Rio de Janeiro). A summary of the method application is provided as well as the final remarks on the effects of the modeling experience. Although not institutionalized in the company, the process used in the modeling experience did enact changes on important decision processes.
Model Actors’ Goals: driven by SDsituations, goals are modeled in i* SD models. Identifying actor’s rules, a Strategic Actors diagram is also built. Model Actors’ Goals Rationale: i* SR models and NFR models (SIGs – Softgoal Interdependency Graphs) are built following the “Intentionality Panel” and following the constructs provided by SDsituations.
Specify SDsituations: each SDsituation (component) is described by applying a scenario based description technique.
Analyse SD and SR Models: a technique called i* Diagnoses [7] provides verification and validation tasks to enhance the quality of the models.
ELICITATION - MODELLING - ANALYSIS
ELICITATION - MODELLING - ANALYSIS 1) Elicit Actors'Goals 2) Identify SDsituations 3) Model Actors'
Goals UofDInf-orSmoautricones of (people & documents)
BASELINE Requirements artifacts 6) Analyse SD and SR
Models 5) Specify SDsituations 4) Model Actors' Goals Rationale ELICITATION - MODELLING - ANALYSIS
ELICITATION - MODELLING - ANALYSIS
ELICITATION - MODELLING - ANALYSIS This paper is based on the final undergraduate project of a student from UERJ, who is a financial analyst with nine years of experience both in IT projects and in the overall financial area of the company. He recently graduated in Computer Science at UERJ and has used the ERi*c method as the central theme for his dissertation. The method was used to model the process of investment decision for the company [5]. The company uses SAP R/3 systems and the investment decision process was modeled using SAP’s business process modeling language (see Figure 2). This process description, and the Investment Contracts Pattern together with several interviews and meetings were used as the main sources of information [4] for the application of ERi*c, in the production of an intentional model for the referred process.
The investment decision process starts by a request from the responsible area. This request should be detailed with overall objectives and project costs and must pass several tests and evaluations, including analysis of economic and financial viability, budget forecasting and the nature of costs. This process includes feedback to the requester in order to improve the project. After these assessments, a sub-process is created to handle all purchase requisitions in the project for the procurement of materials and services with suppliers. After these steps, the solicitation of investment is still subject to the approval of the executive board, which may, depending on the value, involve the company CEO. Only after all these evaluations the project implementation can start (see Figure 3).
In next section the experience will be described and resultsexplored either by diagrams or key points after enforcing this perspective into representing the i* language.
This paper summarizes the experience of applying ERi*c to a real situation in a very large organization. The student who did the final undergraduate project described here, works for the organization as a financial analyst and had access to this investment approval process.
The first author was the creator of the ERi*c method and the second author was the the first author’s advisor to ERi*c. The student used Oliveira’s Ph.D. thesis [1] as the main source for the method, and consulted Oliveira during the process of creating the first draft. The second author provided guidance for the final version of the student’s dissertation.
The student spent 107 hours applying the method. The breakdown, considering Figure 1, is as such: Step 1: 45 hours; Step 2: 15 hours; Step 3: 10 hours; Step 4: 12 hours; Step 5: 11 hours; Step 6: 14 hours.
The first step (Elicit Goals) is very important for the performance of the method. The result in this work confirmed previous laboratory experimental experiences. The elicitation strategy used is Actors’ Goals from Lexicon – AGFL [6] which considers all kinds of actions revealed by LEL and performed inside the selected context. AGFL can improve the performance of the rest of ERi*c phases.
The method’s step “Elicit Actors’ Goals” [6] elaborated the LEL – Lexicon Extended Language (application vocabulary) of the application area, and detected 38 entries or symbols (10 subjects, 15 objects, 11 verbs, and 2 states). The behavior responses of lexicon symbols produced 144 goals in a preliminary list (concrete and flexible (softgoals). The step “Identify SDsituations” mapped 11 strategic dependency situations which are shared by 10 strategic actors. See Table I for overall data. Consequently, using i* language concepts engineers can design one particular approval process to meet specific goals.
Below we show one example of goal elicitation of the actor Contracts Analyst. The example shows the elicitation of one goal and one softgoal for the SDsituation. The softgoal’s correctness [purchase requisition] was elicited because the action “analyses” is a flexible action and this indicates a softgoal. The action “negotiates” results a concrete end “one value was established by negotiators” and consequently this action indicates a concrete goal.
---------------------------------------------------------------------------------------------------LEL´s behavioral responses: actor’s responsibilities, in order to achieve goals. • Action flexible/concrete (WHY?) Concrete Goal / Softgoal (quality attribute) ---------------------------------------------------------------------------------------------------Contracts Analyst - Contracts Analyst works over a contract negotiation by communicating with the Supply Director.
• Analyses “purchase requisition”. (was considered as flexible action) BECAUSE correctness [purchase requisition]
purchase requisition should BE approved BY Supply Director • Negotiates values (was considered as concrete action)
BECAUSE purchase requisition should BE approved BY Supply Director The investment approval process frequently encompasses several organizational units as actors within the enterprise context.
The step “Model Actors’ Goals” separated the problem into 11 parts (SDsituations) and consequently 11 diagrams of each type were produced. The SDsituation: Contract Negotiation is shown in Figure 4a and Figure 4b. During the modeling process a refinement of goals was conducted, see Table I and II. Concrete Softgoal elicited grouped refined 57 13 c)
The last ERi*c step “Analyse SD/SR Models” pointed potential problems in the diagnose product “Matrix: Goals x Problems”. These problems were important feedback both to the modeling and for the investment process.
4
Although the ERi*c has not been evaluated formally by the company, the student reported that the modeling exercise improved his understanding of the whole process and made him aware of improvements that could be made. He reported the possible improvements to the company managers, who agreed to change parts of the activities related to Investment Contracts. The student also reported that the breakdown of responsibilities, by looking at the SDsituations, made him more aware of the inner tasks of the whole area. We understand that even being experienced with investment analysis for the company, it was the method he followed that allowed him to better grasp the different situations and made him even more knowledgeable about the area.