UFO, OntoUML and its associated Tools - Giancarlo Guizzardi
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
UFO, OntoUML and its associated Tools OntoCommons Workshop Giancarlo Guizzardi March, 2021
Acknowledgements • João Paulo Almeida • Fernanda Baião • Mattia Fumagalli • Nicola Guarino • Claudenir Moraes • Daniele Porello • Tiago Sales
Foundational Engineering Theories Tools Domain Domain Specific Specific Theories Applications
Foundational Theories
Foundational
Theories
↝
Types and Taxonomic Structures,
Dependence, Events, (Part-Whole)
Relations, Relations, Causality,
Multi-Level Modeling, Dependence,
etc…Foundational
Theories
↝
Types and Taxonomic Structures,
Dependence, Events, (Part-Whole)
Relations, Relations, Causality,
Multi-Level Modeling, Dependence,
etc…Foundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, LearnersFoundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, Learners1..*
1..*1..*
1..*1..*
1..*1..*
1..*1..*
1..*1..*
1..*1..*
1..*1..*
1..*1..*
1..*1..*
1..*Foundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, LearnersFoundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, Learnersg; α = e ×(some
mediation relations a, and of
ε= a × g.constraints are illustrated in Figure 3) [10].
these
Figure 8-10 Material ´mediationª ´relatorª ´mediationª
Relations and their c..d R e..f
founding relators (the η..λ
cardinality constraints of
the derived relation and axg bxh
the derivation relation
a..b ε..φ g..h
itself can be calculated
from the corresponding Everything else
F in the model is aG representation of a type that these ki
´materialª
mediation relations α..β /H γ..δ
involving the founding can instantiate contingently. fxb dxh
relators)
Two alternative versions of a concrete example of this situation are depicted
in figures
Fig. 3. Relator 8.11.a Relations
and Material and 8.11.bPattern.
below. However, due to the lack of expressivity
of the traditional UML association notation, these two models seem to
Since the formal
convey modeling primitives(from
the same information of this
the language
perspectivecan only
of the appear
material following these
relation
patterns, these patterns end
supervised-by), up beingthey
although the actual modeling
describe primitivesdifferent
completely of the language.
As a consequence, modelingAsindiscussed
conceptualizations. OntoUML is done
in section by the
6.3.3, the benefits
chainedofapplication
explicitly of these
ontologicalrepresenting
patterns [19]. relator
Thisuniversals instead of inmerely
idea is illustrated Figurerepresenting
4. We startmaterial
by modeling the
relations, becomes even more evident in n-ary relations with n >
type Customer. We first identify that a Customer is a RoleMixin: instances of Custom-2.
Figure 8-11
er can be different kinds (people and organizations);
´mediationª
1..*
´relatorª
Customer is an anti-rigid type (no
´mediationª ´mediationª
´relatorª
´mediationª
Customer
Examplification of how is necessity1..* a Customer);
Assignment 1..*
in order1..*for someone 1..*
to be a 1..*
Assignment
Customer, she has to
relators can
purchase
disambiguate two something from a Supplier. In applying the RoleMixin pattern of Figure 2.c,
1..*
1
we identify
conceptualizations that the´roleª
presence
1..*
of two phases
´materialª
1..*
´roleª
(Living Person 1
´roleª
and Active
´materialª Organization),
1
´kindª
a role
in the standard UML
(Supplier,
notation would have the which is assumed
GraduateStudent /supervised-by
1..* to1..*be played by entities of1..*the
Supervisor GraduateStudent
unique
/supervised-by
1..* kind Organization)
Supervisor
(b)
and a relation (purchases from).
same interpretation (a)
We then expand this model by applying to phases and
roles the patterns of Figure 2.a and 2.b, respectively. Finally, we apply the pattern of
Once more we should highlight that the relator individual is the actual
Figure 3 to instantiation
the materialofrelation purchases from.
the corresponding relational property (the objectified
This strategy of
relation). building models
Material relations standby the successive
merely instantiation
for the of
facts of these
derived from the patterns
Fig 1.
has been implemented
Representing
in the
relator individual andnew
the possibility
version entities.
its mediating of the OntoUML
change
editor.
Therefore, we
for
claim This
Endurants
approach can
that the
representation
bring several benefits toofconceptual
the relators modeling.
of material relations must have
Firstly, since theseprimacy overare the rep-
patterns
the
of representation
resentationThis ontological of the material
theories, relations themselves.
the construction In otherbywords,
of models the
instantiating these modeling languag
model
representation of figure
of ´materialª 1
relations is represented in a conceptual
patterns preserves ontological consistency by can be omittedThis
construction. but can
whenever a
also facilitate the
process oftoUML [9]. This
model building, language
especially has been
to novice users. designis to
The hypothesis thatreflect
in each the ontological dis
step of theaxiomatization
modeling activity, theput forth
solution spacebythatthe UnifiedtheFoundational
characterizes possible choices Ontology (UFO) [
of modeling primitives to be adopted is reduced. This strategy, in turn, reduces the
cognitive ticular,
load of thethis modelerlanguage has as modeling
and, consequently, the complexityprimitives
of model buildingthose that represent on
using this tinctions
language [19]. between
Moreover,all thisthe aforementioned
strategy sorts oftotypes
also brings more uniformity the (e.g., kinds, pha
models (which become described in terms of known patterns) and provides for a natu-
mixins, relators). Figure 1 represents the possibility of change, i.e., how2.b.
Everything else in the model is a representation of a type t
(b)
can instantiate contingently.
ase Partition Pattern (b) and the RoleMixin Pattern (c).
y the variation of these ontological meta-properties can also be
1..*
s. One example is the notion of a RoleMixin. A RoleMixin is a
1..*
-rigid and relationally dependent. In other words, the RoleMix-
and, hence, is subject to many of the same constraints of the
r, unlike a role, a RoleMixin classify entities that instantiate
obey different principles of identity). Once more, the ontolog-
oleMixin cause it to manifest in OntoUML necessarily follow-
depicted in Figure 2.c. Like Roles, RoleMixins must be con-
ng relation with an opposite association having a minimum
ual to one (symbolizing the Figrelational dependence
1. Representing condition).
the possibility of change for Endurants
ixins classify entities of different kinds, they must be parti-
ecializing sortals (roles), each
This of which
model of classify
figure 1entities of a
is represented in a conceptual model
toUML [9]. This language has been design to reflect the ont(symbolizing the relational dependence condition). Likewise, the ontological axioms
defining phases cause the manifestation of its construct in OntoUML to obey neces-
sarily the pattern of fig. 2.b.
Everything else in the model is a representation of a type that these kinds of th
can instantiate contingently.
(a) (b)
Fig. 2. Role Pattern (a), Phase Partition Pattern (b) and the RoleMixin Pattern (c).
Distinctions generated by the variation of these 1..* ontological meta-properties can also be
found among non-sortals. One example is the 1..* notion of a RoleMixin. A RoleMixin is a
non-sortal, which is anti-rigid and relationally dependent. In other words, the RoleMix-
in category is similar to and, hence, is subject to many of the same constraints of the
Role category. However, unlike a role, a RoleMixin classify entities that instantiate
different kinds (and that obey different principles of identity). Once more, the ontolog-
Fig 1. Representing
ical axioms defining a RoleMixin cause it tothemanifest
possibilityinofOntoUML
change for Endurants
necessarily follow-
ing a particular pattern depicted in Figure 2.c. Like Roles, RoleMixins must be con-
nected to a characterizingThis modelwith
relation of figure 1 is represented
an opposite in a conceptual
association having a modeling
minimumlanguage termed
toUML [9]. This language has been design to reflect the ontological distinctions
cardinality higher or equal to one (symbolizing the relational dependence condition).
axiomatization put forth by the Unified Foundational Ontology (UFO) [9,13]. In
However, since RoleMixins classify entities of different kinds, they must be parti-
ticular, this language has as modeling primitives those that represent ontological
tioned in a series of specializing
tinctions sortals
between(roles), each of whichsorts
all the aforementioned classify entities
of types (e.g., of a phase, roles,
kinds,
particular kind [10]. mixins, relators). Figure 1 represents the possibility of change, i.e., how things co
Finally, in UFO, we have a fundamental
possibly distinction
be for the entities that are between
assumed theto so-called formal
exist in this domain (i.e., people
and material relations. A ganizations,
formal relation is acar
cars and relation
rentals).thatIn holds directly the
this approach, between
OntoUML its model of figu
relata and that is reduciblecan
to intrinsic properties
be automatically of these
translated to relata. Take,representation
knowledge for instance,languages
the such as OFoundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, LearnersFoundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, Learners1..*
1..*(Economic Agreement) (Economic Agreement)
Economi (Economic Agreement)
1..*
1..*ces of T1 are subject to different cardinality
ubtypes SBTj. An example in the model of
The Emerging Anti-Pattern: Relation
l Investigation can have at least two Detec-
Between Overlapping Types (RelOver)
in.
ating the (a) AC, (b) IA and (c) RWOR.Foundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, LearnersFoundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, LearnersConceptual Model Clustering: A Relator-Centric Approach 7
packages with the homonymous focal relator. As one can observe, the original model
can be broken down into four contexts, namely: the Car Rental, the Marriage, the
Car Ownership, and the Employment contexts. Each of these modules contains a
view of the original model with all the information required to understand each of the
contexts.
Fig. 2 An RCC for the model of Figure 1 organized as (Onto)UML packages.
The Car Rental RC shows the roles (and role mixin) directly mediated by the Car
Rental relator (Responsible Employee, Rental Car, Customer). The kinds involved
are made explicit: Person, Car and Organization (when playing the role of Corpo-Foundational ↝ Engineering
Theories Tools
Modeling Languages,
Methodologies, Computational
Tools, Patterns, Anti-Patterns, Code
Generators, Simulators, Complexity
Management Tools, Learnershttps://github.com/
OntoUML/ontouml-vp-
pluginSome References
• Guizzardi, G., Wagner, G., Almeida, J.P.A., Guizzardi, R.S.S., Towards Ontological Foundation for Conceptual Modeling: The Unified Foundational
Ontology (UFO) Story, Applied Ontology, IOS Press, 2015.
• Guizzardi, G., Ontological Patterns, Anti-Patterns and Pattern Languages for Next-Generation Conceptual Modeling, invited companion paper to the
Keynote Speech delivered at the 33rd International Conference on Conceptual Modeling (ER 2014), Atlanta, USA.
• Guizzardi, G. ,Ontological Foundations for Structural Conceptual Models, Telematica Instituut Fundamental Research Series No. 15, ISBN
90-75176-81-3, 2005.
• Ruy, F., Guizzardi, G., Falbo, R., Reginato, C.C., Santos, V.A., From Reference Ontologies to Ontology Patterns and Back, Data & Knowledge
Engineering, Elsevier, 2017.
• Sales. T.P., Guizzardi, G., Ontological Anti-Patterns: Empirically Uncovered Error-Prone Structures in Ontology-Driven Conceptual Models, Data &
Knowledge Engineering, Elsevier, 2015.
• Almeida, J.P.A., Guizzardi, G., Sales, T.P., Falbo, R.A., "gUFO: A Lightweight Implementation of the Unified Foundational Ontology (UFO)", Technical
Report, 2020.
• Guizzardi, G., Fonseca, C., Almeida, J.P., Sales, T.P., Benevides, A.B., Porello, D., Types and Taxonomic Structures in Conceptual Modeling: A Novel
Ontological Theory and Engineering Support, Data & Knowledge Engineering, Elsevier (accepted, forthcoming, 2021).
• Guizzardi, G., Sales, T.P., Almeida, J.P., Poels, G., Relational Contexts and Conceptual Model Clustering, 13th IFIP WG 8.1 working conference on
the Practice of Enterprise Modelling (PoEM 2020), Riga, 2020.
• Guizzardi, G., Figueredo, G., Hedblom, M., Poels, G., Ontology-Based Model Abstraction, IEEE 13th International Conference on Research
Challenges in Information Science (RCIS 2019), Brussels, Belgium, 2019.
• Guizzardi, G.; Wagner, G.; Guarino, N.; van Sinderen, M., An Ontologically Well-Founded Profile for UML Conceptual Models, 16th International
Conference on Advances in Information Systems Engineering (CAiSE), Latvia, 2004.Springer-Verlag, Berlin, Lecture Notes in Computer Sciencegguizzardi@unibz.it g.guizzardi@utwente.nl
You can also read
