Universal Enterprise Adaptive Object Model
David Aveiro
1,2,3
and Duarte Pinto
1
1
Exact Sciences and Engineering Centre, University of Madeira, Caminho da Penteada 9020-105 Funchal, Portugal
2
Madeira Interactive Technologies Institute, Caminho da Penteada, 9020-105 Funchal, Portugal
3
Center for Organizational Design and Engineering, INESC-INOV Rua Alves Redol 9, 1000-029 Lisboa, Portugal
Keywords: Enterprise Engineering, Model, Meta-Model, Abstract Syntax, Concrete Syntax, Adaptive Object Model,
DEMO.
Abstract: In this paper we present a novel conceptual model that systematizes the integrated management and
adaptation of: (1) enterprise models, (2) their representations, (3) their underlying meta-models, i.e., their
abstract syntax and (4) the representation rules, i.e., concrete syntax for the respective models. All this for
different modeling languages and also different versions of these languages. Thanks to our original use of
the adaptive object model and type square patterns – normally applied in the context of software
engineering, but here applied for enterprise engineering – we manage to provide a strong conceptual
foundation for the development of software tools that will allow a precise and coherent specification of
models and their evolution and also of meta-models and their evolution.
1 INTRODUCTION
A large amount of time is lost, in organizations, in
the handling of unknown exceptions causing
dysfunctions as exception handling can sometimes
take almost half of the total working time, and the
handling of, and recovering from, exceptions is
expensive (Saastamoinen and White, 1995). On
another hand, current Enterprise Engineering (EE)
approaches seem to lack in concepts and method for
a continuous update of organizational models, so
that they are always up to date and available as a
more useful input for the process of continuous
change of organizational reality and decision on
possible evolution choices. It seems that the root
problem is an absence of concepts and method for
explicit capture, and management of information of
exceptions and their handling, which includes the
design and operationalization of organization
artifacts (OA) – e.g., actor role pizza deliverer – that
solve caused dysfunctions. Not immediately
capturing this handling and the consequent resulting
changes in reality and the model of reality itself, will
result that, as time passes, the organization will be
less aware of itself than it should be, when facing the
need of future change due to other unexpected
exceptions. The lack of awareness of organizational
reality has been addressed with the coining of the
term “Organizational Self-Awareness” (OSA),
presented and refined in (Magalhaes et al., 2007)
and (Zacarias et al., 2007). OSA stresses the
importance and need of continuously available,
coherent, updated and updateable models of
organizational reality. With our research work we
aim to facilitate distributed awareness of
organizational reality and also coordinated
distributed change of models of the enterprise's
reality using adequate methods and software tools as
a support. In our tool development efforts a
necessity arose of allowing a precise way of
conceptualizing and implementing the separation of
3 concerns: reality; models of reality; and their
representations, while having adequate flexibility for
model and meta-model evolution. In section 2 –
Related Work & Problem – we present the basic
notions for a proper understanding of our work as
well as the problem itself and in section 3 – The
Universal Enterprise Adaptive Object Model – we
present our proposal contextualizing it and
explaining in detail its principles and applicability.
Finally, in section 4 – Conclusions – we briefly
present our conclusions and discuss what
contributions it brings.
89
Aveiro D. and Pinto D..
Universal Enterprise Adaptive Object Model.
DOI: 10.5220/0004550000890099
In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2013), pages 89-99
ISBN: 978-989-8565-81-5
Copyright
c
2013 SCITEPRESS (Science and Technology Publications, Lda.)
2 RELATED WORK & PROBLEM
We ground our research in a particular
Organizational Engineering approach, namely, the
Design & Engineering Methodology for
Organizations (DEMO) (Dietz, 2006) and its
underlying theory. Our research – presented in this
and the next sections – are heavily based in DEMO
so, while proceeding, the reader which is unfamiliar
with this methodology is advised to also consult
(Dietz, 2006) or (Dietz and Albani, 2005) or other
publications in: www.demo.nl. From several
approaches to support EE being proposed, DEMO
seems to be one of the most coherent,
comprehensive, consistent and concise (Dietz,
2006). It has shown to be useful in a number of
applications, from small to large scale organizations
– see, for example, (Dietz and Albani 2005) and
(Op’ t Land 2008) (p. 39). Nevertheless, DEMO
suffers from the shortcoming referred above.
Namely, DEMO models have been mostly used to
devise blueprints to serve as instruments for
discussion of broader scale organizational change or
development/change of IT systems (Op’ t Land,
2008) (p. 58) and does not, yet, provide modeling
constructs and a method for a continuous update of
its models as reality changes. Current software tools
supporting DEMO also suffer from the same
shortcoming, the problem we address in this paper.
2.1 Basic Ontological Notions
We adopt the ontological system definition from
(Dietz 2008) (citing (Bunge, 1979)) which concerns
the construction and operation of a system. The
corresponding type of model is the white-box model,
which is a direct conceptualization of the ontological
system definition presented next. Something is a
system if and only if it has the next properties: (1)
composition: a set of elements of some category
(physical, biological, social, chemical etc.); (2)
environment: a set of elements of the same category,
where the composition and the environment are
disjoint; (3) structure: a set of influencing bonds
among the elements in the composition and between
these and the elements in the environment; (4)
production: the elements in the composition produce
services that are delivered to the elements in the
environment. From (Dietz, 2008) we find that in the
Ψ-theory based DEMO methodology, four aspect
models of the complete ontological model of an
organization are distinguished. The Construction
Model (CM) specifies the construction of the
organization: the actor roles in the composition and
Figure 1: The meaning triangle.
Figure 2: The ontological parallelogram.
Figure 3: The model triangle.
Figure 4: Model triangle applied to organizations.
Figure 5: Meaning triangle applied to a transaction OA.
Figure 6: Model triangle applied to the organization space.
KEOD2013-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment
90
the environment, as well as the transaction kinds in
which they are involved. The Process Model (PM)
specifies the state space and the transition space of
the coordination world. The State Model (SM)
specifies the state space and the transition space of
the production world. The Action Model (AM)
consists of the action rules that serve as guidelines
for the actor roles in the composition of the
organization.
In Figures 1 and 2, we find, respectively, the
meaning triangle and the ontological parallelogram,
taken from (Dietz, 2005) which explain how
(individual) concepts are created in the human mind.
We will also base our claims in the model triangle,
taken from (Dietz, 2006) and presented in Figure 3.
We find that the model triangle coherently overlaps
the meaning triangle. This happens because a set of
symbols – like a set of DEMO representations
(signs) that constitute a symbolic system – allows
the interpretation of a set of concepts – like a set of
DEMO aspect models, part of the ontological model,
constituting a conceptual system. This conceptual
system, in turn, consists in the conceptualization of
the “real” inter-subjective organizational self, i.e.,
the set of OAs constituting the concrete organization
system's composition structure and production.
Figure 4 is an adaptation from the model triangle of
Figure 3 and depicts our reasoning. We call the set
of all DEMO diagrams, tables and lists used to
formulate the ontological model as ontological
representation.
Now relating with the meaning triangle, we can
verify that a particular sign (e.g., a transaction
symbol with label membership fee payment), part of
an ontological representation (e.g., actor transaction
diagram, representing a library's construction model)
designates (i.e., allows the interpretation or is the
formulation) of the respective concept of the
particular transaction part of the respective
ontological model (e.g., construction model). This
subjective concept, in turn, refers to a concrete
object of the shared inter-subjective reality of the
organization's human agents (e.g., the particular OA
transaction T02). Figure 5, an adaptation from the
meaning triangle depicts this other reasoning.
Another example of an OA related with T02
would be the transaction initiation OA, relating T02
with actor role registrar (also designated by A02)
and formulated by a line connecting the transaction
and actor role symbols of T02 and A02. Actor role
registrar is, in turn, another OA of the construction
space of the library. Once such role is communicated
to all employees of a library, it becomes a “living”
abstract object part of the shared inter-subjective
reality of the library's human agents. Such object,
along with other OAs of the organizational inter-
subjective reality, give human agents a way to
conceptualize their organizational responsibilities –
in this case, requesting membership fee payments to
aspirant members. We name this set of all abstract
objects living in the inter-subjective reality of an
organization's members as the organizational self.
From these notions we proposed a set of claims
presented in more detail in (Aveiro et al., 2010) and
summarized next. An organization – besides
producing a set of products or services for its
environment – also produces itself. That is, enclosed
in its day-to-day operation, there will be parts of its
operation which change the organization system
itself, i.e., change the set of OAs that constitute its
composition, structure and production. By formally
and explicitly specifying these change acts one
keeps a definite and updated record of produced
OAs. Such a record – the OAs base – constitutes the
means for one to always be able to conceptualize the
most current and updated ontological model of the
organizational self. Thus the continuous production
of the organizational self should include the
synchronized production of the collective and
subjective “picture” (awareness) of the
organizational self – the conceptualization that
constitutes its ontological model – thanks to the
synchronized production of the respective symbolic
system – an ontological representation that allows
the interpretation of the ontological model and the
conceptualization (awareness) of the organizational
self. To separate concerns, we propose that change
acts are performed by a (sub-)organization
considered to exist in every organization (O) that we
call: G.O.D. Organization (GO) – change acts lead
to the Generation, Operationalization and
Discontinuation of OAs. The GO's production world
will contain the current state of O's self as well as its
relevant state change history. The GO has the role of
continuously realizing and capturing changes of
organizational reality. Thus, by implementing the
GO pattern in a real organization, in an appropriate
manner, providing automatic generation of
ontological representations derived from the OAs
base, one can achieve OSA. This is possible because
one can implement clear rules that, based on the
arrangement of OAs of the organizational self,
automatically produce the appropriate ontological
representation which, in turn, allows the appropriate
interpretation of the ontological model, that is, the
correct conceptualization of the organizational self.
OAs constituting the organizational self are
arranged in a certain manner as to specify all the
UniversalEnterpriseAdaptiveObjectModel
91
spaces (state, process, action and structure) of an
organization's world, i.e., they have to obey certain
rules of arrangement between them. We call the
specification of these rules as the ontological meta
model. The ontological meta-model is the
conceptualization of the OA space. By OA space we
understand the set of allowed OAs. It is specified by
the OA base and OA laws. The OA base is the set of
OA kinds of which instances, called OAs, may occur
in the state base of the GO's world. The OA laws
determine the inclusion or exclusion of the
coexistence of OAs. The definition of the OA space
is quite similar to the definition of state space of an
organization's production world – specified in World
Ontology Specification Language (WOSL) (Dietz
2006) – and, thus, it is appropriate to use WOSL to
express the ontological meta-model in, what we
propose to call: the Organization Space Diagram
(OSD). DEMO's OSD is currently called as the
DEMO Meta Model (DMM), the chosen name for
the specification provided in (Dietz, 2009) and
consisting, in practice, in the OSDs corresponding to
the four DEMO aspect models: SM, CM, PM and
AM. These diagrams formulate, for each aspect
model, the OA kinds out of which instances – OAs –
can occur in the organizational self and coexistence
rules governing how to arrange these instances.
Another reason we propose to use the expression
Organization Space Diagram is because we're in fact
looking at a Space Diagram which, following the
model triangle (Dietz, 2006), is a symbolic system
which is a formulation of the conceptual system of
the ontological meta model. So, for coherency
reasons, one should not use terms “Meta” and
“Model” to name those figures but use, instead, the
term Organization Space Diagram. The OSD allows
the interpretation, in one's mind, of the ontological
meta model. The complete set of organization
artifact kinds and laws governing the arrangement of
their instances constitutes the organization space.
The conceptualization of the organization space
consists in the ontological meta-model which, in
turn, is formulated in what we call the Organization
Space Diagram. A depiction of this reasoning is
present in Figure 6, another adaptation from the
model triangle. The G.O.D. organization is
addressed in detail in (Aveiro et al., 2010). The
proposal presented in this current paper consists in
an evolution of the conceptual model proposed in
this other paper, taking in account state-of-the-art
related model theory and concepts described next.
2.2 Theoretical Foundations on Models
In a graphical modeling language, the vocabulary is
expressed in terms of pictorial signs. Those
graphical primitives form the concrete syntax i.e the
lexical layer of such language. The abstract syntax,
on the other hand, is usually defined in terms of an
abstract visual graph or a meta-model specification.
A meta-model specification of a language defines
the set of grammatically correct models that can be
constructed using that language, a vocabulary. The
concrete syntax provides a concrete representational
system for expressing the elements of that meta-
model (Guizzardi, 2005). In a communication
process, besides agreeing on a common vocabulary,
the participants need to also share the meaning for
the syntactical constructs being communicated so
they are able to interpret in a compatible manner the
expressions being used. To this end a language's
semantics can be constructed in two parts: a
semantic domain i.e. the real world entities to which
those semantics apply and a semantic mapping from
the syntactic vocabulary to such domain that tells us
the meaning of each of the language's expressions as
an element in that specific domain. In graphical
languages, vocabulary, syntax and semantics cannot
be clearly separable. A graphical vocabulary of a
modeling language may include shapes of differing
sizes and colors that often fall into a hierarchical
typing that constrains the syntax and informs about
the semantics of the system (Guizzardi, 2005). The
abstract syntax of a model manages the formal
structure of the model elements and the relationships
amongst them (La Rosa et al., 2011).
The MetaObject Facility (MOF) Specification is
the industry-standard environment where models can
be exported from one application, imported into
another, transported across a network, stored in a
repository and then retrieved, rendered into different
formats like XMI or XML, transformed, and used to
generate application code (OMG, 2012). The
Adaptive Object Model (AOM) is a pattern that
represents classes, attributes, and relationships as
meta-data. It is a model based on instances rather
than classes. Users change the meta-data (object
model) to reflect changes in the domain. These
changes modify the system’s behavior. In other
words, it stores its Object-Model in a database and
interprets it. Consequently, the object model is
active, when you change it, the system changes
immediately (Yoder et al., 2001).
KEOD2013-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment
92
3 THE UNIVERSAL
ENTERPRISE ADAPTIVE
OBJECT MODEL
Figure 7: Type Square.
The long term objective of our research is the
development of a wiki-based system that allows an
effective integrated enterprise modeling, while
allowing dynamic evolution of meta-models, models
and their representations, while providing intuitive
navigation through their elements and also their
semantics, allowing wide-spread model
interpretation and distributed model creation and
change, reflecting enterprise changes, thus
addressing our problem. An essential step in this
direction is what we call the Universal Enterprise
Adaptive Object Model (UEAOM), depicted in
Figure 8. We apply the AOM pattern referred in the
previous section so that each page or semantic
property of our semantic wiki-based system
corresponds to instances of classes of our AOM.
Wiki pages, that are instances of class
DIAGRAM, automatically generate SVG diagrams
based on shape and connector pages. These pages
also allow dynamic editing of diagrams and
underlying models. We also apply the type-square
pattern (Yoder et al., 2001) – depicted in Figure 7 –
4 times as to allow run-time dynamic change of: (1)
meta-model elements, (2) model elements, (3) shape
elements and (4) connector elements. Our UEAOM
is represented with the World Ontology
Specification Language (WOSL) (Dietz, 2005).
WOSL is based in Object Role Modeling language
(Halpin, 1998) which is also used as a base for the
specification of the anatomy of Archimate
(Lankhorst et al., 2010), a similar effort to ours. In
(Ferreira et al., 2008) a relation between Adaptive
Object Model pattern and the MOF standard is
presented, where run-time instances of the
operational level are equivalent to MOF's M0 and
knowledge level; classes, attributes, relations and
behavior is equivalent to M1, being M2 an
equivalent to the models used to define an AOM. As
in the work of Ferreira et al., in our UEAOM all
these MOF levels are projected as run-time
instances. In our prototype system, we have as
instances both organization artifacts – i.e., concrete
organization models – and organization artifact
kinds – i.e., the meta-model specification or, in other
words, the abstract syntax. So both M1 and M2
levels of the MOF framework exist and change at
run-time. But the MOF and Ferreira's initiative are
too software development oriented and too complex
for our needs. Our main contribution in this paper is
to apply these fundamental theoretical foundations
and adapt them to the field of enterprise ontology.
Having the UEAOM contextualized, an
explanation of its content is now due. With the
UEAOM's classes we are not explicitly specifying
syntaxes of particular modeling languages. What we
can do, while instantiating these classes, is to specify
any syntax of any modeling language, along with
particular models of each language, and also their
evolution, all this in run-time. For a better
understanding and following the essential and
important validation by instantiation principle
(Dietz, 2009) we present, for all elements of our
AOM, example instances for the DEMO language,
namely a fragment of the EU-rent case's
Construction Model and its respective Actor
Transaction Diagram. Thus, we can find, in red color
expressions, instances of both our classes and fact
types of our UEAOM concerning the EU-rent case
which allow a better interpretation of our proposal.
3.1 Abstract Syntax
Relevant classes for the specification of the abstract
syntax of any version of any language are presented
in Figure 9. The main concepts of the abstract syntax
specification are expressed in the classes
LANGUAGE, MODEL KIND,
ORGANIZATIONAL ARTIFACT KIND (OAK)
and ORGANIZATIONAL ARTIFACT RELATION
KIND (OARK). They specify all allowed artifacts
(e.g. transaction kind OAK and transaction
execution relation OARK) for different types of
models that can exist for different languages. Class
ORGANIZATIONAL ARTIFACT RELATION
KIND has ten properties that can be divided in two
groups of five where each group specifies one of the
two sides of an allowed relation between two OAKs.
The ones named prefix, infix and suffix specify the
formulation that can be done around the names of
the two OAKs being related. Most times, only the
infix needs to be specified. With the unicity and
dependency properties we specify the cardinality of
UniversalEnterpriseAdaptiveObjectModel
93
Figure 8: Universal Enterprise Adaptive Object Model.
the relation and which OAKs are mandatory or not
to participate in the relation. Reference law fact
types specify which two OAKs are allowed to
participate in this relation. Practical example of the
first set of the referred 5 properties: F Transaction
Kind T is initiated by Elementary Actor Role
corresponds to a set of Dependency 1, Reference law
1, Unicitiy 1, Infix_1_2 and Reference law 2. F
Elementary Actor Role T is initiator of Transaction
Kind would be its corresponding Dependency 2,
Reference law 2, Unicitiy 2, Infix_2_1 and
Reference law 1. Thanks to this part of our UEAOM
specification we allow a precise and formal
formulation of the abstract syntax of models, already
giving considerable semantics thanks to the prefix,
infix, suffix and OAK names that can be composed
in formulations for each direction of the relation.
Instances of class ORGANIZATION ARTIFACT
PROPERTY specify intrinsic properties of OAKs,
like identifiers and names. The respective property
PROPERTY DOMAIN allows us to specify the
domain for each intrinsic property of an OAK (e.g.,
string, number, etc.). Examples of instances are
property transaction id with domain T<number> or
transaction name with domain <string>.
In Figure 10 we can see an excerpt of the current
DEMO ontological meta-model and the UEAOM
classes used to define it. Both these models are the
equivalent to the M2 MOF model that, as we have
seen, sets the rules for specifying concrete models.
All elements of this meta-model can be considered
instances of the classes we just have presented. The
binary fact type [elementary actor role] is an
initiator of [transaction kind] is, in our UEAOM, an
instance of ORGANIZATIONAL ARTIFACT
RELATION KIND class, with values for the infixes
being: initiates and initiated by. There are, however,
other classes: DIAGRAM KIND, SHAPE KIND,
CONNECTOR KIND, CONNECTOR and SHAPE
PROPERTY that are present in this Figure 10 and
KEOD2013-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment
94
Figure 9: UEAOM - Abstract Syntax classes.
Figure 10: DEMO Ontological Meta-Model and UAOM classes used to represent it.
UniversalEnterpriseAdaptiveObjectModel
95
Figure 11: UEAOM - Concrete syntax.
are part of the meta-model level of the UEAOM but
are not part of the abstract syntax, these will be
explained in more detail in the next section.
3.2 Concrete Syntax
The UEAOM classes that allow the specification of
rules for the concrete representation of models, i.e.,
the concrete syntax, are presented next. These
classes, together with all their inter-relating fact
types are present in Figure 11. With the class
SHAPE KIND, instances of the types of shapes
allowed to be part of diagram kinds representing
certain model kinds are specified. These shape kinds
are also specifically connected to the OAKs whose
instances they will represent. For example, the
elementary actor role shape is allowed in diagram
kind Actor Transaction Diagram, that represents the
construction model of DEMO language. Instances of
this shape represent instances of OAK actor role.
With SHAPE PROPERTY, we specify the
properties for each shape, e.g., line color and actor
id label of actor role shape. Instances of
CONNECTOR KIND specify allowed
representations for OAKRs, e.g. transaction
initiation connector instances represent instances of
OAKR transaction initiation. With CONNECTOR
PROPERTY, the properties of each connector are
specified, e.g., for the just mentioned connector, line
color: black and line dashing: continuous.
Instances of REPRESENTATION RULES, class
are an informal textual based specification of rules
on how ORGANIZATIONAL ARTIFACT KINDS
and ORGANIZATIONAL ARTIFACT RELATION
KINDS should be represented. These rules are taken
in consideration in either SHAPES or
CONNECTORS that represent those OAKs and
OARKs. For example, a transaction is a black circle
with a black diamond inside. It is also according to
the REPRESENTATION RULES that we have a
definite answer if an OARK will give origin or not
to a connector or if instead it will be represented by
the connection of two shape kinds directly.
Revisiting the full example from Figure 8, an
elementary actor role shape would be an instance of
class SHAPE KIND, for the representation of
instances of the actor role OAK. Transaction shape
would also be an instance of SHAPE KIND for the
representation of instances of transaction OAK. So
an instance of class CONNECTOR KIND for the
representation of this OAKR would be transaction
initiator connector, with properties like line type:
dashed. Many of the SHAPE KINDs and
CONNECTOR KINDs are comprised by multiple
symbols that need to be considered individually as
having a set of properties. Although in most cases
KEOD2013-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment
96
Figure 12: DEMO concrete diagrams example and UEAOM classes used to represent them.
the aggregate of composing symbols are treated as
“one” in the diagram drafting, such as a circle and
diamond in an actor transaction diagram transaction,
that have a fixed size (height and width) and none of
them can be altered, there are also cases in which
symbols need to be treated and moved in the
diagrams in a separate and independent way having
their own set of SHAPE PROPERTIES or
CONNECTOR PROPERTIES like, for example, in
a process step diagram where the diamond inside the
transaction can be moved and re-sized according to
the needs. As a solution for this, we have classes
SYMBOL ELEMENT KIND that specify each
symbol element to be present in a shape kind or
connector kind and SYMBOL ELEMENT that are
instances of SYMBOL ELEMENT KIND and
specify concrete representations of SYMBOL
ELEMENTS of a specific kind. As an example of
this we can consider the actor transaction diagram
SHAPE KIND transaction as being composed by the
UniversalEnterpriseAdaptiveObjectModel
97
SYMBOL ELEMENT KINDS Transaction
Diamond and Transaction Circle.
In Figure 12 we have a partial example of a
concrete representation of the DEMO ontological
models of an Actor Transaction Diagram and Object
Fact Diagram and the corresponding part in the
UEAOM. DEMO Ontological models are the
equivalent to the M1 level of MOF and instances of
their OA's to the MOF's M0 level.
Ontological Models and their representation are
covered in the UEAOM by the classes: DIAGRAM,
where concrete instances of a certain DIAGRAM
KIND are specified; SHAPE, where concrete
instances of SHAPE KIND are specified; SHAPE
PROPERTY VALUE, where concrete instances of
SHAPE PROPERTY are specified; CONNECTOR,
where concrete instances of CONNECTOR KIND
are specified; CONNECTOR PROPERTY VALUE,
where concrete properties of the CONNECTOR
PROPERTY are specified; ORGANIZATIONAL
ARTIFACT, where concrete instances of
ORGANIZATIONAL ARTIFACT KIND are
specified; ORGANIZATIONAL ARTIFACT
PROPERTY VALUE, where concrete OA
properties are specified and ORGANIZATIONAL
ARTIFACT RELATION, where concrete instances
of ORGANIZATIONAL ARTIFACT RELATION
KIND are specified and all their relating fact types.
In this way also allowing them to be changed in an
easy and consistent way in run-time environment.
Again using a concrete example from Figure 12,
we have the “CA-01 aspirant member shape”, this is
an instance of SHAPE (this SHAPE an instance
itself of the SHAPE KIND “Composite Actor Role”)
that represents the ORGANIZATIONAL
ARTIFACT “CA-01 aspirant member” (itself an
instance of the ORGANIZATIONAL ARTIFACT
KIND “Composite Actor Role”); the string “aspirant
member” is an instance of SHAPE PROPERTY
VALUE (that represents the instance of
ORGANIZATIONAL ARTIFACT PROPERTY
VALUE “aspirant member”) and so is “CA-01”.
These two strings are VALUES, instances of the
SHAPE PROPERTIES “Actor Name” and “Actor
ID” respectively (that again represent the instances
of ORGANIZATIONAL ARTIFACT PROPERTY
VALUE “Composite Actor Name” and “Composite
Actor ID”).
The DIAGRAM KIND and MODEL KIND
classes were not present in the original DEMO
Ontological meta-model as all models were
specified in this single meta-model. But in our
UEAOM, as we have generalized this definition to
accommodate any language for organizational
modeling, the DIAGRAM KIND and MODEL
KIND classes are vital so we can relate to each
specific Ontological Model. Actor Transaction
Diagram or Process Step Diagram would be
examples of instances of this DIAGRAM KIND
while the first would be a representation of the
MODEL KIND Construction Model and the second
a representation of the MODEL KIND Process
Model. The LANGUAGE class and its property
VERSION is used to define the modeling language
being modeled and the version of such language.
4 CONCLUSIONS
In this paper we proposed what we call the
Universal Enterprise Adaptive Object Model,
(UEAOM) that, through the multiple application of
the type square pattern and the adaptive object
model pattern, allows a robust and precise
conceptual solution to manage changes in
organization artifacts, their models and meta-
models, all in run time environment. It is possible,
through specification of instances of our UEAOM,
to manage the various aspects of modeling
languages: the semantics; the abstract and concrete
syntaxes and the pragmatics all in the same
implementation and object model. This gives
flexibility for a very important contribution of our
proposal: the possibility to completely change or
create a new behavior for the system by changing its
concrete and/or abstract syntax on the fly.
As we can see, important and essential
representation aspects like shapes, connectors and
their properties and representation rules, are not
specified formally and precisely (as we do in our
examples) in DEMO's original meta-model, nor in
the meta-model of other mainstream languages such
as BPMN or Archimate. The abstract and concrete
syntaxes as well as the language's pragmatics are
very important aspects of the modeling activity and
should be specified in the meta-model specification
itself, even if with semi-formal textual information
(in our case provided by instances of the class
REPRESETATION RULES) as to allow the least
ambiguity possible. Our UEAOM approach to
enterprise modeling differentiates itself from other
enterprise modeling languages like BPMN and
Archimate in the way that in expands the AOM
principles to also model the meta-model level in run
time environment taking as an advantage also of the
type square pattern in a coherent, consistent and
robust way. Our conceptual solution offers tool
makers a sound theoretical base for an extensive and
KEOD2013-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment
98
thorough management of knowledge of an
organization and also of the languages being used to
create models. It also considers the DEMO's Ψ-
theory principle that nothing ceases to exist, but
instead, artifacts have their state changed. Although
specified with the implementation of the DEMO
methodology in mind, the UEAOM is flexible
enough to allow the modeling of multiple languages
like Archimate or BPMN.
As future developments of this work, we will
provide a more detailed specification of how to
manage the versions of the language and the
possibility to migrate models and also a better
specification of the state changes that can occur with
all artifacts that are instances of our UEAOM
classes.
REFERENCES
Aveiro, D., Rito Silva, A. & M. Tribolet, J., 2010.
Extending the Design and Engineering Methodology
for Organizations with the Generation
Operationalization and Discontinuation Organization.
Em 5th International Conference, DESRIST 2010. St.
Gallen, Switzerland, June 4-5, 2010: Springer, pp
226–241.
Bunge, M. A., 1979. Treatise on basic philosophy, vol. 4,
a world of systems, Reidel Publishing Company.
Dietz, J. L. G., 2005. A World Ontology Specification
Language. Em S. B. / Heidelberg, ed. On the Move to
Meaningful Internet Systems 2005: OTM Workshops.
pp 688–699. Available at: http://dx.doi.org/10.1007/
11575863_88.
Dietz, J. L. G., 2009. Demo meta model specification
(forthcoming, in www.demo.nl).
Dietz, J.L.G., Enterprise ontology: theory and
methodology. Springer-Verlag New York, Inc.
Secaucus, NJ, USA. (2006).
Dietz, J. L. G., 2009. Is it PHI TAO PSI or Bullshit? Em
The enterprise engineering series. Methodologies for
Enterprise Engineering symposium. Delft: TU Delft,
Faculteit Elektrotechniek, Wiskunde en Informatica.
Dietz, J. L. G., 2008. On the Nature of Business Rules.
Advances in Enterprise Engineering I, pp.1–15.
Dietz, J. L. G. & Albani, A., 2005. Basic notions regarding
business processes and supporting information
systems. Requirements Engineering, 10(3),pp.175–183.
Ferreira, H. S., Correia, F. F. & Welicki, L., 2008. Patterns
for data and metadata evolution in adaptive object-
models. Em Proceedings of the 15th Conference on
Pattern Languages of Programs. PLoP ’08. New
York, NY, USA: ACM, pp 5:1–5:9. Available at:
http:// doi.acm.org/ 10.1145/ 1753196.1753203.
Guizzardi, G., 2005. Ontological foundations for structural
conceptual models. Available at: http://doc.utwente.nl/
50826/.
Halpin, T., 1998. Object-Role Modeling: an overview. Em
In http://www.orm.net/pdf/ORMwhitePaper.pdf.
Lankhorst, M. M., Proper, H. A. & Jonkers, H., 2010. The
Anatomy of the ArchiMate Language. International
Journal of Information System Modeling and Design,
1(1), pp.1–32.
Magalhaes, R., Zacarias, M. & Tribolet, J., 2007. Making
Sense of Enterprise Architectures as Tools of
Organizational Self-Awareness (OSA). Proceedings of
the Second Workshop on Trends in Enterprise
Architecture Research (TEAR 2007), June, 6,pp.61–70.
OMG, 2012. OMG’s MetaObject Facility (MOF) Home
Page. Available at: http://www.omg.org/mof/.
Op’ t Land, M., 2008. Applying Architecture and Ontology
to the Splitting and Allying of Enterprises. TU Delft.
La Rosa, M. et al., 2011. Managing Process Model
Complexity Via Abstract Syntax Modifications. IEEE
Transactions on Industrial Informatics, 7(4), pp.614–
629.
Saastamoinen, H. & White, G. M., 1995. On handling
exceptions. Proceedings of conference on
Organizational computing systems, pp.302–310.
Yoder, J. W., Balaguer, F. & Johnson, R., 2001.
Architecture and design of adaptive object-models.
SIGPLAN Not., 36(12), pp.50–60.
Zacarias, M. et al., 2007. Towards Organizational Self-
Awareness: An Initial Architecture and Ontology. Em
P. Rittgen, ed. Handbook of Ontologies for Business
Interaction. Information Science Reference, pp 101–
121.
UniversalEnterpriseAdaptiveObjectModel
99