iStar 2.0 Language Guide

1 iStar Language GuideFabiano Dalpiaz1, Xavier Franch2, and Jennifer Horkoff3?1 Utrecht University, The Netherlands2 Universitat Polit`ecnica de Catalunya, Spain3 City University London, United Language was introduced to fill the gap inthe spectrum of conceptual modeling languages, focusing on the inten-tional (why?), social (who?), and strategic (how? how else?) *has been applied in many areas, , healthcare, security analysis,eCommerce. Althoughi*has seen much academic application, the di-versity of extensions and variations can make it difficult for novices tolearn and use it in a consistent way. This document introduces the core Language , evolving the basic concepts ofi*into a consistent andclear set of core concepts, upon which to build future work and to basegoal-oriented teaching materials. This document was built from a set ofdiscussions and input from various members of thei*community.

2 It isour intention to revisit, update and expand the document after collectingexamples and concrete experiences with iStar : Okhaide Akhigbe, Fatma Ba sak Aydemir, Juan Pablo Carvallo, JaelsonCastro, Luiz Marcio Cysneiros, Sepideh Ghanavati, Alicia Grubb, Giancarlo Guiz-zardi, Renata Guizzardi, Matthias Jarke, Alexei Lapouchnian, Tong Li, Lin Liu, LidiaL opez, Alejandro Mat e, John Mylopoulos, Soroosh Nalchigar, Elda Paja, Angelo Susi,Juan Carlos Trujillo Mond ejar, Eric Yu, Jelena [ ] 16 Jun 20162F. Dalpiaz, X. Franch, and J. HorkoffVersion HistoryVersion DateImplemented Changes3 June 17, 2016 New integrity rules: at most one actor link between a pair ofactors; contribution and qualification cannot connect the sametwo 3, 2016 Fixed typos in the original version1 May 26, 2016 - iStar Language Guide31 Motivation and OverviewThei* Language was presented in the mid-nineties [4] as a goal- and actor-orientedmodeling and reasoning framework.

3 It consists of a modeling Language along withreasoning techniques for analyzing created *was quickly adopted by theresearch community in fields such as requirements engineering and business mod-eling. Benefiting from its intentionally open nature, multiple extensions of thei* Language have been proposed (see [2,3] for useful reviews), either by slightlyredefining some existing constructs, by detailing some semantic issues not com-pletely defined in the seminal proposal, or by proposing new constructs for flexible use ofi*has been fruitfully employed by researchers, who wereable to benefit from a consolidated modeling and reasoning approach whilst tai-loring it to their needs. However, this use has also some drawbacks. The mostcritical is the difficulty to spread the framework outside the experts community: Newcomersfind it hard to learn the intricacies of the Language ; Educatorsdo not have a shared body of knowledge to teach; Practitionersare not provided with an established reference for usingi*intheir projects; Technology providerscannot easily determine which are the core constructsto be implemented and the techniques to apply on top of those a response to the need of balancing the framework s open nature anda possible solution to the aforementioned adoption problems, thei*researchcommunity started an initiative to identify a widely agreed upon set of coreconcepts in thei* Language .

4 The main goal is to keep open the ability to tailorthe framework while agreeing on the fundamental document summarizes the outcomes of the first iteration of the iStar1standardization process. To clearly distinguish this core Language from its prede-cessors, we name itiStar community discussed this Language in several meetings and discussionsstarting in a dedicated one-day meeting before the ER 14 conference in Atlanta(October 2014). At the subsequent community meeting at CAiSE 15 in Stock-holm (at the iStar teaching workshop, iStarT, June 2015), it was decided thata smaller group of researchers (the authors of this document) would Guide theprocess, making concrete proposals and processing the inputs of the rest of thecommunity. An initial draft of the core was discussed both at the iStar Workshopcolocated with RE 15 (August 2015) and in another dedicated one-day meetingbefore ER 15 in Stockholm (October 2015).

5 The obtained feedback has been in-corporated into the document. A preliminary version was distributed among theresearchers that participated in this process (December 2015), who provided alast round of comments, considered in this version (May 2016).Each of the following five sections (Section 2 6) addresses a particular cat-egory of Language constructs as presented in mosti*sources, for example , theiStar-wiki2. For each category, the document lists the concepts that are includedin iStar , with a definition, necessary comments, concise examples and the1 The Language is spelled iStar instead ofi* to allow better indexing through Dalpiaz, X. Franch, and J. Horkoffgraphical representation. The focus of this document is on concepts and relation-ships; methodological possibilities are mentioned only aspects are excluded from this first version of iStar but are plannedfor inclusion in the next versions.

6 Among them, we mention the ontological defi-nition of constructs ( , what is a goal?), visual representation ( , what is aneffective graphical notation?), wording conventions ( , passive voices in goals),and methodological issues ( , when can a model be considered final?).Illustrative illustrate the concepts of iStar using a running ex-ample concerning University travel reimbursement. Students must organize theirtravel ( , to conferences) and have several goals to achieve, and options toachieve them. To achieve their goals, students rely on other parties such as aTravel Agency and the university s trip management information system. We willintroduce iStar concepts gradually, slowly building up the example . In Fig. 1 weshow a final view of the example in order to give readers an early idea of thecapabilities of iStar organizedAuthorization obtainedTrip bookedSupervisor authorizesHead-of-dept authorizesAuthorization signedRequest preparedFill in paper formFill in online formQuick bookingTickets bookedAccommodation bookedTrip parts bookedMinimal own payments Agency buys ticketsSelf-book ticketsBuy ticketsPay for ticketsBuy through through hotel websiteConference hotel bookedBudget hotel bookedNo errorsComfortCredit cardTrip bundle bookedBuy flight ticketsBook bundleStudentBook bundle via expediaTravel agencyUniv.

7 Trip mgmt ISOnline form processedProcess formRequest authorizationNotify applicantDetails validatedUniv. of Wonder-LandMike WhitePhD studentFig. preview of the Travel Reimbursement Scenario as captured in iStar Language 2 introduces the notion of actor and distinguishes betweenthree types of actors. Section 3 presents the links in iStar for relating 4 describes the intentional elements that characterize the actors. Section 5discusses the dependencies that socially relate the actors. Section 6 explains howthe intentional elements can be linked. Section 7 details how to create differentviews of an iStar model. Section 8 shows the metamodel of the , we conclude and present future directions in Section Actors and actor typesActors are central to the social modeling nature of the active,autonomous entities that aim at achieving their goals by exercising their know-how, in collaboration with other actors.

8 In the iStar Language , two types ofactors are distinguished: Role: an abstract characterization of the behavior of a social actor withinsome specialized context or domain of endeavor. Examples are: Student, PhDStudent. Agent: an actor with concrete, physical manifestations, such as a human indi-vidual, an organization, or a department. Examples are: Travel agency, PhDStudent, University of Wonderland, Mike distinguishing the type of actor is not relevant, either because ofthe scenario-at-hand or the modeling stage, the notion of genericactor withoutspecialization can be used in the model. For example , we can denote Travelagency as an actor to say that we do not know yet whether it is a specific agency(agent) or a characterization of the travel agency are represented graphically as circles. In the case of agent, a straightline is added in the top part of the actor circle.

9 For a role, a curved line is addedin the lower part. The graphical notation follows the mnemonic guidelines thatthe originali*adopts3. Fig. 2 illustrates the agencyPhD studentUniv. of Wonder-LandFig. of actor, role and agentActors intentionality is made explicit through theactor boundary, which isa graphical container for their intentional elements (see Section 4) together withtheir interrelationships (see Section 6). Fig. 3 shows the graphical representationof an actor boundary; elements and relationships will appear inside the grey symbols are stylized depictions of a person wearing a hat and viewed fromdifferent angles. The Agent symbol is the frontal view where the name of the agentappears on the face. The Role symbol is an overhead view so that the label on thehat is Dalpiaz, X. Franch, and J. HorkoffTravel agencyFig. of actor boundary3 Actors association linksActors are often interrelated.

10 In iStar , this is captured via actor links thatdefine/describe these relationships. Actor links are binary, linking a single actorto a single other actor. Two different types of actor links have been defined: is-a: represents the concept of generalization / specialization in iStar Onlyroles can be specialized into roles, or general actors into general actors. Forinstance, a PhD student (role) can be defined as a specialization of a Student(another role). Agents cannot be specialized via is-a, as they are concreteinstantiations ( , Mike White cannot be another agent). participates-in: represents any kind of association, other than generalization /specialization, between two actors. No restriction exists on the type of actorslinked by this association. Depending on the connected elements, this linktakes different meanings. Two typical situations are the following: When the source is an agent and the target is a role, this represents theplaysrelationship, , an agent plays a given role.