=Paper=
{{Paper
|id=None
|storemode=property
|title=Towards a Semantic Alignment of the ArchiMate Motivation Extension and the Goal-Question-Metric Approach
|pdfUrl=https://ceur-ws.org/Vol-1041/ontobras-2013_paper24.pdf
|volume=Vol-1041
|dblpUrl=https://dblp.org/rec/conf/ontobras/CarvalhoNTGG13
}}
==Towards a Semantic Alignment of the ArchiMate Motivation Extension and the Goal-Question-Metric Approach==
https://ceur-ws.org/Vol-1041/ontobras-2013_paper24.pdf
Towards a Semantic Alignment of the ArchiMate Motivation
Extension and the Goal-Question-Metric Approach
Victorio Albani de Carvalho1,2, Julio Cesar Nardi1,2, Maria das Graças da
Silva Teixeira2, Renata Guizzardi2, Giancarlo Guizzardi2
1
Research Group in Applied Informatics. Federal Institute of Espírito Santo, Campus
Colatina, Colatina/ES, Brazil.
2
Ontology and Conceptual Modeling Research Group (NEMO). Federal University of
Espírito Santo, Vitória/ES, Brazil.
{victorio, julionardi}@ifes.edu.br, maria.teixeira@ufes.br,
{rguizzardi,gguizzardi}@inf.ufes.br
Abstract. Supporting Goal-Oriented Requirement Engineering (GORE) in a
systematic and comprehensive way may require the combination of distinct
goal-oriented approaches. However, due to lack of common semantics, to
combine these approaches can be challenging. In this work, we propose a
semantic alignment between two complementary goal-oriented approaches: the
ArchiMate Motivation Extension and the Goal-Question-Metric. The
approaches are semantically analyzed in light of the Unified Foundational
Ontology (UFO), which provides real-world semantics for both languages,
serving as a reference ontology to support the ontological analysis of concepts
and relationships of both approaches and the alignment between them.
1. Introduction
In the past decade, Goal-Oriented Requirement Engineering (GORE) became a de facto
standard approach in Requirements Engineering (RE) [Kavakli and Loucopoulos
2005][Yu et al. 2011]. In this context, many GORE approaches have been proposed, but
they use to focus on particular perspectives, and no single approach can address all the
needs of this engineering process. However, these single approaches can be put to work
together in order to compound a stronger and more complete GORE framework, which
could benefit of the strengths of each approach [Kavakli 2002].
Considering this, the current paper focuses on the alignment of ArchiMate
Motivation Extension and Goal-Question-Metric (GQM) approaches. ArchiMate is a
modeling-based framework that has gained visibility both in academia and in industry.
This framework provides a component called Motivation Extension (ME) that supports
modeling of the enterprise's actual motivations or intentions by adopting the concept of
“goal”, among others [The Open Group 2012]. GQM [Basili, Caldiera and Rombach
1994], in turn, is an approach for evaluating the fulfillment of enterprise’s goals. It is a
well established and widely used approach [Boyd 2005] [Kaneko et al 2011]. By being
complementary, ArchiMate ME and GQM can be adopted in tandem as a way to define
a more comprehensive goal-oriented framework.
59
Combining different languages, however, clearly characterizes a Semantic
Interoperability problem, since for combining languages one has to: (i) make clear the
meaning of the primitives that compose the languages (i.e., the real-world semantics of
the languages); (ii) establish the correct ontological relations between the alternative
semantic domains [Guizzardi 2005].
One of the defining aspects of ontologies is their use in making explicit shared
conceptualizations. Reference ontologies are kinds of ontologies used in an off-line
manner to assist humans in tasks such as meaning negotiation and consensus
establishment [Guizzardi 2005]. Foundational (or top-level) ontologies are domain-
independent reference ontologies that describe very general concepts independently of a
particular problem or domain (e.g., object, event, quality, action, etc.) [Guizzardi
2005][Guarino 1998]. These ontologies can provide real-world semantics for general
representation languages and constrain the possible interpretations of their modeling
primitives [Mika et al. 2004][Guizzardi 2005]. Also, by being adopted as a common
reference, foundational ontologies can be used to map different representation
languages and approaches [Cardoso et al. 2010].
In this paper, we propose the use of the Unified Foundational Ontology (UFO)
[Guizzardi 2005] [Guizzardi et al. 2008] as a well-founded basis for defining
(ontological) real-world semantics for the concepts of ArchiMate ME and GQM
approaches. However, we have two basic semantic-based problems: (i) the ArchiMate
ME and the GQM do not share the same semantics; and (ii) some of their concepts’
semantics are not clearly defined. These problems, therefore, may lead different
designers to assume distinct meanings and uses for the supposed same concepts. Thus,
in order to properly use ArchiMate ME and GQM together, it is necessary to understand
the semantics of the concepts of each approach, and how to map the concepts between
the approaches. The choice for UFO as a foundational ontology here is motivated by a
number of successful cases of using this ontology in the analysis, re-design and
integration of different major modeling languages, including UML [Guizzardi 2005], i*
[Guizzardi et al. 2012] and ARIS [Cardoso et al. 2010], among others.
The remainder of this paper is organized as follows: Section 2 presents briefly
ArchiMate ME, GQM and the used alignment process; Section 3 describes the fragment
of UFO necessary for this work; Section 4 contains the proposed ontology-based
alignment between ArchiMate ME and GQM; Section 5 describes a running example
that illustrates how the semantic alignment may be put in practice; Section 6 presents
some related works; and, finally, Section 7 draws final considerations.
2. Background: ArchiMate Motivation Extension (ME) and GQM
ArchiMate ME addresses the way the enterprise architecture is aligned to its context by
using of motivational elements [The Open Group 2012]. It builds upon existing work on
GORE, such as KAOS and i* [Kavakli and Loucopoulos 2005], adopting interesting
findings of these previous initiatives. The concepts defined by ArchiMate ME are goal,
stakeholder, driver, assessment, requirement, principle and constraint.
To clarify the use of some concepts that are considered in this work, Figure 1
depicts a diagram developed by using of ArchiMate ME language. This diagram
presents two stakeholders, each one having at least one driver, which can be shared. For
each driver, a goal is defined. These goals are related to the stakeholder that commits
60
on pursuing it. Also, one can model that a fulfillment of a goal contributes positively (or
negatively) to the realization of other goals.
Figure 1. Illustrating some concepts of ArchiMate ME
GQM defines an approach for setting goals in a quality improvement paradigm.
It is based on the assumption that for an organization to work efficiently with a
measurement program, it should [Basili et al. 1994]: (i) specify the goals of the
organization itself and its projects; (ii) map those goals to data that operationally define
the goals (through questions, which direct the information that support the evaluation of
a goal; and metrics, which indicate the types of data to be collected in order to answer
the questions); and (iii) provide a framework for interpreting these data in relation to the
established goals. Figure 2 presents an example of a GQM model for illustrating the
relations between goals, questions and metrics.
Figure 2. An example of the relations between goals, questions and metrics in GQM
In GQM, a goal is defined for an object (i.e. a process, a product, or a resource)
based on a number of reasons (the purpose) with respect to models of quality (issue of
quality) from viewpoints in relation to a particular environment. Thus, a goal consists of
three coordinates (object, issue/focus and viewpoint) and a purpose [Basili et al. 1994].
Besides the fact that both approaches deal with goals, ArchiMate ME and GQM
present some distinctions. While GQM focuses on measurement and evaluation of the
fulfillment of organizational goals, ArchiMate ME focuses on specifying/representing
these goals. In the context of GORE, a GQM model aims at describing goals that should
have their fulfillment evaluated. ArchiMate, in turn, does not offer any approach for
goal measurement. The ArchiMate ME supports the representation of organizational
goals and their relations with other organizational elements. Some of these relations
cannot be clearly expressed in GQM. Moreover, providing guidelines on how to identify
61
goals is out of the scope of ArchiMate ME. The aforementioned distinctions indicate
that GQM and ArchiMate ME are complementary approaches. However, considering
that the semantics of these approaches are not clearly defined and harmonized, a
semantic alignment between them becomes necessary as a first step towards their use in
tandem. For example, despite being a common term in both approaches, a deeper
analysis reveals that the concept of goal in GQM specializes goal in ArchiMate ME,
since the definition of goal in GQM comprises a set of interrelated concepts in
ArchiMate (e.g., goal, driver, and stakeholder). We believe that understanding these
semantic aspects may be useful for establishing an integrated framework.
For performing the ontological analysis and the alignment between ArchiMate
ME and GQM we have followed the iterative process illustrated by Figure 3. This
process is composed of three basic phases. In the “phase 1” and “phase 2”, respectively,
ArchiMate ME concepts and GQM concepts were analyzed and grounded in light of
UFO. In “phase 3”, such concepts were aligned taking as basis the ontological analyzes.
Figure 3. The iterative process used in the ontological alignment
3. The Unified Foundational Ontology (UFO)
UFO is a foundational ontology that has been developed with an interdisciplinary
approach comprising theoretical and empirical results from Formal Ontology,
Philosophical Logic, Linguistics, and Cognitive Psychology. UFO consists of three
main parts: UFO-A, UFO-B, and UFO-C. UFO-A is an ontology of endurants
[Guizzardi 2005], UFO-B is an ontology of events (perdurants) [Guizzardi et al.
2008][Guizzardi et al. 2013], and UFO-C is an ontology of social entities built on the
top of UFO-A and UFO-B [Guizzardi et al. 2008].
A fundamental distinction in UFO is between individuals and universals.
Universals are predicative terms that can be applied to a multitude of individuals,
capturing the general aspects of such individuals. Individuals are entities that exist in
reality possessing a unique identity and that can instantiate a multitude of universals
[Guizzardi, 2005]. Figure 4 presents a fragment of UFO.
In UFO-A, endurants are individuals that are wholly present whenever they
exist, and that can be further specialized into substantials and moments. Substantials are
existentially independent endurants (e.g., a person, a car). Moments are individuals that
can only exist in other individuals, being existentially dependent on their bearers (e.g., a
person’s headache, a covalent bond between atoms) [Guizzardi 2005]. Intrinsic
moments are kinds of moments that are dependent on one single individual (e.g., John´s
headache). Qualities are intrinsic moments associated with quality structures that inhere
in an individual. A quality structure can be either a quality domain or a quality
62
dimension; quality domains are composed of multiple quality dimensions. For example,
a color (quality) “c” of an apple (individual) “a” takes its value in a structure of three-
dimensional color domain constituted of the quality dimensions “hue”, “saturation” and
“brightness”. Modes are intrinsic moments that are not directly associated with a quality
structure (and, hence, are not directly measurable) (e.g., John’s intentions) [Guizzardi
2005]. Relators, in turn, are moments that existentially depend on two or more
endurants [Guizzardi 2005]. For example, consider that John and Mary are married. In
this case, the relator (their marriage) mediates the relation between John and Mary
aggregating all externally-dependent modes that they acquire by virtue of participating
in this relation (e.g, all commitments and claims towards each other). Thus, by virtue of
being connected by this particular marriage, John plays the role of “husband” and bears
responsibilities and rights towards Mary, who, by playing the role of “wife”, and also
bears the responsibilities and rights towards John.
Figure 4. A fragment of UFO
Universals in UFO-A are types instantiated by endurants (universals) and can be
substantial universal and moment universal, whose individuals are substantials and
moments [Guizzardi 2005], respectively. Sortal universals are substantial universals that
carry a principle of identity for their individuals (e.g., Apple, Person). The
specialization of sortal universal is based on a meta-property called rigidity. An
universal is rigid if it necessarily applies to its instances in every possible world (e.g.,
Apple, Person). In contrast to rigid universals, an universal is anti-rigid if it does not
apply necessarily to all its instances. For example, an individual ‘x’, which is instance
of the universal ‘Student’ in a world ‘w1’ can cease to instantiate this universal in
another world ‘w2’ without ceasing to exist as the same individual. Roles are anti-rigid
and relationally-dependent sortal universals (e.g., Student, Husband) [Guizzardi 2005],
which means that roles are played by a substantial whenever there is a relator
connecting it to other substantials. Role mixin represents an anti-rigid and externally-
dependent non-sortal universal, which aggregates properties that are common to
different roles (e.g., the role mixin ‘Customer’ aggregates properties from ‘Personal
Customer’ and ‘Corporate Customer’) [Guizzardi 2005].
In UFO-B, events are individuals composed of temporal parts. They happen in
time in the sense that they extend in time and accumulate temporal parts (e.g., a
63
conversation). Whenever an event is present, it is not the case that all its temporal parts
are present. Events universals are patterns of features that can be realized in a number of
different events [Guizzardi et al. 2008][Guizzardi et al. 2013].
In UFO-C, a basic distinction is the one between agents and (non-agentive)
objects. Agents are agentive substantial individuals that are classified as physical agents
(e.g., a person) or social agents (e.g., an organization). Objects are non-agentive
substantial individuals that are classified as physical objects (e.g., a book) and social
objects (e.g., a language) [Guizzardi et al. 2008]. Agents can bear special kinds of
modes named mental moments. Mental moments refer to situations in reality (also
called state-of-affairs, i.e., a portion of reality that can be comprehended as a whole) and
has propositional content (an abstract representation of a class of situations referred by
an intentional moment) [Guizzardi et al. 2008]. Mental moments are specialized in
intentions, beliefs, and desires. Belief can be justified by situations in reality (e.g., my
belief that the Moon orbits the Earth). Desires and intentions can be fulfilled or
frustrated. Whilst a desire (e.g., a desire that Brazil wins the 2014 World Cup) expresses
a will of an agent towards a state-of-affairs, intentions (e.g., go to the beach today) are
desired state-of-affairs for which the agent commits to pursuing (i.e., intentions are self-
commitments). Intentions cause the agent to perform actions and have propositional
contents that is termed goal [Guizzardi et al. 2008].
4. Ontology-based Alignment between ArchiMate ME and GQM
This section presents the proposed semantic alignment between GQM and ArchiMate
ME in light of UFO. Since goal is the common concept between these two approaches,
it was considered as the key concept for performing the alignment.
In GQM, a goal is characterized as having a purpose, which is associated with
three coordinates: object, issue and viewpoint [Basili et al. 1994]. According to GQM,
an object can be a process, a product, or a resource. In terms of UFO, a process can be
interpreted as an (complex) event [Guizzardi et al. 2008]. A resource is defined in UFO
as a non-agentive substantial (i.e., an object) participating in an event [Guizzardi et al.
2008]. However, in organizational contexts the term “resource” is also used to refer to
“human resource”. In light of UFO, a human resource is a person (i.e., an agent)
participating in an event playing a specific role. Therefore, the concept of resource in
GQM comprises the idea of a non-agentive substantial (object) as well as the idea of an
agent playing a role. In UFO, a product is a resource whose participation in events is
limited to two types [Guizzardi et al. 2008]: creation participation (i.e., a product can be
created) and changing participation (i.e., a product can be updated).
An issue, in GQM, refers to a quality aspect of an object, which can be
interpreted based on the concept of quality in UFO [Guizzardi 2005]. Thus, an issue is
as a quality that inheres in individuals (events and endurants). The individual that bears
such quality (in terms of UFO) can be interpreted as being the object in GQM. For
example, an issue can be the efficiency of a maintenance process (process as an object
in terms of GQM) in an organization.
According to GQM, the fulfillment of a goal must be measured from a viewpoint
(e.g., a manager's viewpoint, or a customer's viewpoint). Thus, the concept of viewpoint
is associated with a “who” question, which makes reference to the roles played by one
or more agents interested in a goal. Based on that, we may interpret that the viewpoint,
64
in terms of UFO, is associated with the role that an agent must be playing in the
organizational context in order to be a possible measurer of the goal’s fulfillment.
A purpose is related to a “why” question, in the sense that it refers to the
intended effect associated with a goal. This intended effect is associated with a quality
aspect (the issue coordinate) of an object (the object coordinate). For example,
considering the goal “decrease the rate of error of the manufacturing process” the
associated purpose is “decrease”, the issue is “the rate of error”, and the object is
“manufacturing process”. Thus, in light of UFO, we can interpret that the propositional
content of the intention defines an intended effect (purpose) associated with a quality
(issue) of an individual (object).
According to ArchiMate Specification 2.0, “a goal is defined as an end state that
a stakeholder intends to achieve” [The Open Group 2012]. Thus, we can say that a
stakeholder is committed (has an intention) to achieve a goal, and by achieving the
goal, certain effects in reality are brought about. Thus, in terms of UFO, a goal (in
ArchiMate ME) may be interpreted as “the propositional content of an agent’s
intention” [Azevedo et al. 2011]. In this context, we may state that GQM and
ArchiMate ME can be aligned w.r.t. the concept of goal, given that in both approaches a
goal may be interpreted, in terms of UFO, as a propositional content of an intention. As
follows, however, we will discuss the concept of goal in GQM as a specialization of the
concept of goal in ArchiMate ME.
As aforementioned, the GQM viewpoint coordinate can be interpreted as the
roles played by agents that judge the fulfillment of (organizational) goals. These agents
bear the intention of performing the evaluation of the (organizational) goals’ fulfillment.
On the other hand, there may be agents that bear the intention of pursuing the
(organizational) goals’ fulfillment, i.e., they are committed to perform actions in order
to fulfill these goals. Thus, there may be agents that are committed at pursuing
(organizational) goals and agents that are committed to judge if these goals were
achieved or not. For example, in goal “Analyzing the customer relationship information
system (object) for the purpose of improving its usability (issue) from the viewpoint of the
customer” the “customers” judge the fulfillment of the goal, but they do not necessarily
have a commitment at pursuing the goal. “Having a system with a great usability” may
be only a desire for the “customers”, whereas other stakeholder, possibly the “customer
relationship manager”, has the commitment at pursuing it. It is important to highlight,
however, that the same agent may be committed to both, pursuing the goals’ fulfillment
and judging their fulfillment, although this is not desirable in organizational quality
programs. This distinction between these two kinds of agents becomes clearer by the
ontological analysis in light of the intentional concepts of UFO.
Interpreting the GQM's goal concept, we have realized that the three coordinates
(object, issue and viewpoint) seem to characterize the problem being addressed by the
goal, i.e., the “source” of the goal. Thus, in order to provide a semantic alignment
between GQM and ArchiMate ME, the interpretation of the ArchiMate’s goal concept
is not enough. Besides that, it is also necessary to analyze two other ArchiMate ME's
concepts used to model the “source” of the intentions: the driver and the stakeholder.
In Archimate ME, a driver is defined as "something that creates, motivates, and
fuels the change in an organization". This definition is too vague and allows many
interpretations. We consider that a driver may be interpreted as an event (external or
65
internal to the organization) that leads to a change of situation, which generates a
concern about a key interest (e.g., process, products, resources, costs, etc.) of an
organization. On the other hand, we can consider that some changes in organizations
may also be motivated by some stakeholder’s concerns despite explicitly considering
the event that has generated the concern. In this case, a driver may also be interpreted as
representing a stakeholder´s concern. According to [Azevedo et al. 2011], a concern is
something that a stakeholder believes to be important. Therefore, a concern can be
interpreted as “the propositional content of an agent’s belief” [Azevedo et al. 2011]. The
propositional content of the belief refers especially to properties believed to be
important (the object of the concern) in a specific context. These two possible
interpretations lead to a construct overload in Archimate ME language, since a driver
may represent an event that generates a concern or the concern itself (without
considering the event). In this context, it seems that the driver concept from ArchiMate
ME is, somehow, related to the object and issue coordinates from GQM, in the sense
that all of them refers to something a stakeholder is interested in, i.e., something the
stakeholder believes to be important. In terms of UFO, the driver concept as well as the
object and issue coordinates are related to the propositional content of a belief that
refers especially to properties or characteristics believed to be important. For example,
considering that the manager of a company has a goal of increasing the employees’
productivity we can infer, in terms of UFO, that the manager (an agent) has a belief that
the employees’ productivity is important (the propositional content of the belief). Thus,
in terms of ArchiMate, “the employees’ productivity” may be represented as a driver. In
terms of GQM “employees” may be seen as the value of the object coordinate (a
resource) while the “productivity” refers to the issue coordinate. However, we remark
that the object (as a product, a process, or as a resource) and issue coordinates together
comprise only a subset of the possible drivers (in ArchiMate ME).
Finally, a stakeholder is defined by ArchiMate Specification 2.0 as “the role of
an individual, team, or organization (or classes thereof) that represents their interests in,
or concerns relative to, the outcome of the architecture” [The Open Group 2012].
According to UFO, a stakeholder can be interpreted as a role played by an agent (e.g.,
human individual, team or organization) able to refer to reality (in this case, “the
enterprise architecture”). Thus, the agent instantiates a role, and, as consequence, the
agent bears all the moments that characterizes that role, which include intrinsic
moments (e.g., skills and capabilities that a person should have in order to play the role)
as well as externally dependent modes associated with relators (e.g., as the rights and
obligations that a person bears by participating on an employment contract). In this
context, we may state that the concept of stakeholder in ArchiMate ME is aligned with
the viewpoint coordinate of GQM since both can be interpreted, in light of UFO, as a
role played by an agent. Table 1 summarizes the ontological analysis aforementioned.
Table 1. Summary of the ontological analysis of GQM and Archimate ME concepts
GQM
Concept Interpretation in light of UFO
Object An object can be a process, a resource, or a product.
Process: an (complex) event.
Resource: a non-agentive substantial or an agent playing a role (human resource).
Product: resource whose participation in events is limited to creation participation
and changing participation
Issue A quality that inheres in individuals (events and endurants)
66
Viewpoint The roles played by agents that judge the fulfillment of (organizational) goals
Purpose The intended effect expressed by the propositional content of an intention
Goal The propositional content of an intention which defines an intended effect
(purpose) associated with a quality (issue) of an individual (object).
Archimate ME
Concept Interpretation in light of UFO
Driver An event that generates a concern or the concern itself. A concern can be
interpreted as the propositional content of an agent’s belief. The propositional
content of the belief refers to the importance that the agent ascribes to something.
Stakeholder A role played by an agent (e.g., human individual or organization) able to refer to
reality.
Goal The propositional content of an agent’s intention
5. A Running Example of the Alignment between ArchiMate ME and GQM
This section presents an example that illustrates how the proposed semantic alignment
between GQM and ArchiMate ME may allow an organization to use the strengths of
each approach maintaining the traceability between the generated artifacts. In the
example, ArchiMate ME is used for specifying the organizational goals and GQM for
implementing an evaluation program regarding the fulfillment of these goals.
GQM establishes a structure for defining goals. By other hand, the ArchiMate
ME does not define any specific structure for goal. Thus, we believe that the more
detailed definition proposed by GQM can assist the designer in eliciting the
organization’s goals, i.e., the detailed structure provided by GQM can guide the
designer in asking about and conceiving goals. Thus, we start this example by
describing two organizational goals by using a GQM template, as follows: (G1)
Analyzing the customer relationship process (object) for the purpose of decreasing its
costs (issue) from the viewpoint of the CEO; and (G2) Analyzing the customer
relationship information system (object) for the purpose of improving its usability
(issue) from the viewpoint of the customer.
The proposed semantic alignment has shown that the concept of goal in GQM is
a specialization of the concept of goal in Archimate ME. Thus, each goal described in
GQM can be represented in an ArchiMate ME model. Also, it is possible to derive
drivers in ArchiMate ME from objects and issues in GQM. For example, by the goals
“G1” and “G2”, we can infer, respectively, two drivers - (i) “customer relationship
process costs”, and (ii) “customer relationship information system usability” – which
are represented with their respective goals, as Figure 5a.
Moreover, the alignment shows that it is possible to derive stakeholders in
ArchiMate from viewpoints in GQM (i.e., the agents that judge the goal’s fulfillment).
For example, the “G1” goal is said to be measured by the viewpoint of the “CEO”,
which indicates that the “CEO” may be defined as a stakeholder in ArchiMate.
Similarly, by “G2” goal we can derive another stakeholder: the “customer”. Figure 5a
represents these two stakeholders, associated with the respective drivers. Thus, an initial
ArchiMate ME diagram (as presented by Figure 5a) is directly derived from goals
described in GQM using the proposed alignment.
Moreover, although GQM addresses the agent committed to judge the goals’
fulfillment, it does not offer a coordinate that is directly associated with the agent
committed to pursue the goal’s fulfillment. ArchiMate Specification 2.0, in turn, offers
67
the association relationship as the only way to link stakeholders to goals. However, this
relationship does not differentiate the stakeholders committed to pursue the goals’
fulfillment and the stakeholders committed to judge it. By understanding these
particularities, we suggest that the designer names each association for differentiating
the “pursue” relations from the “judge” relations. For example, in Figure 5b, there are
two stakeholders associated with the goal “improve the customer relationship
information system usability”: (i) the “customer”, who is associated with that goal to
represent the commitment at judging the goal’s fulfillment; (ii) and the “CEO”, who is
associated with that goal for representing the commitment at pursuing the goal. The
diagram in Figure 5b also presents the stakeholder “system analyst”, who is committed
at pursing the “Improve the Customer Relationship IS Usability” goal.
Figure 5 – An ArchiMate ME diagram derived from GQM (a) and an improved version (b)
As a result, the organization would have an ArchiMate ME diagram specifying
the organizational goals, their sources and the relationships between them, aligned with
a GQM model that could be carried out to evaluate the fulfillment of such goals.
6. Related Work
We are unaware of competing approaches which have attempted either an ontological
analysis of GQM or semantic alignments between complementary goal-oriented design
approaches. There are, nonetheless, in the literature a growing number of reports on the
use (foundational) ontologies for performing analysis of (goal) modeling languages.
For instance, in [Cardoso et al. 2010], Cardoso and colleagues propose a
semantic alignment between the ARIS framework (for business process modeling) and
the TROPOS (for modeling and analysis of goals). Their proposal (which is also based
on UFO) contributes to the establishment of a more comprehensive goal-oriented
modeling approach by connecting a goal-modeling perspective with the modeling of
business processes which are supposed to achieve these goals.
In [Azevedo et al. 2011], the authors perform a semantic analysis of the
ArchiMate ME in light of UFO. This analysis was conducted by considering the
whitepaper of the Motivation Extension. The ontological analysis performed in our
work, however, has considered the ArchiMate ME Specification 2.0, which presents
some differences of the initial version presented in the whitepaper, such as: (i) in the
ArchiMate ME Specification 2.0 there is no longer the concept of “concern”, which was
replaced by the concept of “driver”; and (ii) the definition of the “role” concept in the
standard suffered some changes, which were actually driven by the ontological analysis
in [Azevedo et al. 2011]. So, in one sense, the analysis of ArchiMate ME performed
here benefits directly from the previous work of Azevedo and colleagues. The two
efforts, however, also differ also in focus: their work focuses exclusively on the real-
68
worlds semantics of this fragment of ArchiMate; our focus instead is in leveraging this
semantics for integrating it with a complementary approach.
In [Soffer and Wand 2005], the authors employs some of the concepts of the
BWW ontology to analyze the notion of goals in the context of Business Process
Modeling (BPM). The account provided there is a language independent one and, in this
sense, it is comparable to the analysis of hardgoals and softgoals present in UFO
[Guizzardi et al. 2012]. In contrast with UFO, the view of goal provided by these
authors take them as “sets of states of the domain”. Under this view, goals are sets of
elements closer to what is termed a situation (or a state of affairs) in UFO and, as such,
are independent of intentions and, hence, independent of Agents (e.g., people,
organizations). Such a view seems to fail to capture the requirements engineering and
enterprise modeling intuition that goals are “desired state of affairs” [Yu et al. 2011].
Moreover, from an ontological standpoint, the UFO view of goals as propositional
contents of intentions which can possibly satisfied by sets of situations allows even for
goals which are unsatisfiable (an important analysis notion since the satisfiability of
goals cannot always be defined a priori), as well as for distinct goals but which happens
to be satisfied by exactly the same set of situations.
7. Final Considerations
Specification of organizational goals and the evaluation of the fulfillment of these goals
are two complementary and essential activities. For supporting these activities in a
systematic way, it may be necessary to combine distinct goal-oriented approaches. Due
to lack of a common semantics between different approaches, their combination can be
challenging. In this work, we propose a semantic alignment between the Archimate ME
and GQM, which are two complementary approaches that can be applied in tandem to
support the aforementioned activities. These approaches were aligned in light of UFO,
which was used as a domain-independent reference ontology.
The ontology-based analysis conducted in this work contributed to clarify the
meaning of some concepts of GQM and ArchiMate ME and to identify how these
concepts can be aligned. For future works, we plan to continue the ontological analysis
of the ArchiMate ME and the GQM, addressing other concepts, such as, the assessment
concept. After that, we plan to define a cyclic process that comprises goal’s
specification and fulfillment evaluation, by adopting of ArchiMate ME (along with the
GQM goal structure) to specify goals and GQM to assess the goals’ satisfaction. We
also plan to apply the proposed approach in real organizations in order to evaluate it.
Acknowledgments
This research is funded by the Brazilian Research Agencies CAPES/CNPq
(402991/2012-5), FAPES/CNPq (PRONEX 52272362/11), FAPES (59971509/12) and
CNPq (483383/2010-4, 310634/2011-3, and 311578/2011-0).
References
Azevedo, C., Almeida, J. P. A., Van Sinderen, M. J., Quartel, D. and Guizzzardi, G. (2011) "An
Ontology-Based Semantics for the Motivation Extension to ArchiMate", In: 15th IEEE
International Conference on Enterprise Computing (EDOC), Helsinki, Finland, p.25–34.
69
Basili, V. R., Caldiera, G. and Rombach, H. D. (1994) "The goal question metric approach".
Encyclopedia of Software Engineering, v. 1, p. 528–532.
Boyd, A. J. (2005) "The evolution of goal-based information modelling: literature review".
Aslib Proceedings - New Information perspectives, v. 57, n. 6, United Kingdon, Emerald
Group Publishing Limited, p. 523–538.
Cardoso, E. C. S., Santos Jr., P. S., Almeida, J. P. A., Guizzardi, R. S. S. and Guizzardi, G.
(2010) "Semantic Integration of Goal and Business Process Modeling", In: IFIP International
Conference on Research and Practical Issues of Enterprise Information Systems
(CONFENIS), Natal/RN, Brazil.
Guarino, N. (1998) "Formal Ontology in Information Systems", In: Formal Ontology in
Information Systems (FOIS), Trento, Italy, IOS Press, p. 3-15.
Guizzardi, G.(2005) "Ontological Foundations for Structural Conceptual Models". Centre for
Telematics and Information Technology, University of Twente, The Netherlands.
Guizzardi, G., Falbo, R. A. and Guizzardi, R. S. S. (2008) "Grounding software domain
ontologies in the Unified Foundational Ontology (UFO): the case of the ODE software
process ontology", In: Proceedings of the XI Iberoamerican Workshop on Requirements
Engineering and Software Environments. Recife, Brazil, p.244–251.
Guizzardi, G., Wagner, G. and Falbo, R. D. A. (2013) "Towards Ontological Foundations for
the Conceptual Modeling of Events", In: 32nd International Conference on Conceptual
Modeling (ER 2013), Hong Kong.
Guizzardi, R., Franch, X. and Guizzard, G. (2012) "Applying a Foundational Ontology to
Analyze Means-end Links in the i* Framework", In: IEEE International Conference on
Research Challenges in Information Sciences, p.1–11.
Kaneko, T., Katahira, M., Miyamoto, Y. and Kowalczyk, M. (2011) "Application of
GQM+Strategies@ in the Japanese Space Industry", In: Joint Conference of the International
Workshop on Software Measurement and the International Conference on Software Process
and Product Measurement, v. 0, Los Alamitos, CA, USA, IEEE Computer Society, p. 221–
226.
Kavakli, E. (2002) "Goal-Oriented Requirements Engineering: A Unifying Framework".
Requirements Engineering, v. 6, n. 4, p. 237–251.
Kavakli, E. and Loucopoulos, P. (2005) "Goal Modelling in Requirements Engineering:
Analysis and Critique of Current Methods", K. S. John Krogstie, Terry Halpin (Ed.); In:
Information Modeling Methods and Methodologies: Advanced Topics in Database Research.
p.102–124.
Mika, P., Oberle, D., Gangemi, A. and Sabou, M. (2004) "Foundations for Service Ontologies:
Aligning OWL-S to DOLCE", In: The Thirteenth International World Wide Web
Conference Proceedings. New Your, USA, p.563–572.
Soffer, P. and Wand, Y. (2005) "On the Notion of Softgoals in Business Process Modeling".
Business Process Management Journal, v. 11, n. 6, p. 663–679.
The Open Group (2012). "Archimate 2.0 Specification". Berkshire, United Kingdon.
Yu, E. S. K., Giorgini, P., Maiden, N. and Mylopoulos, J. (EDS.) (2011) "Social Modeling for
Requirements Engineering", Cambridge, Massachusetts, MIT Press.
70