SOFTWARE MEASUREMENT BY USING QVT
TRANSFORMATIONS IN AN MDA CONTEXT
Beatriz Mora, Félix García, Francisco Ruiz, Mario Piattini
Department of Information Technologies & Systems, University of Castilla-La Mancha, Ciudad Real, Spain
Artur Boronat, Abel Gómez, José Á. Carsí, Isidro Ramos
Department of Information Systems and Computation Polytechnic University of Valencia, Valencia, Spain
Keywords: Measurement, Model, FMESP, MDA, QVT, SMF.
Abstract: At present the objective of obtaining quality software products has led to the necessity of carrying out good
software processes management in which measurement is a fundamental factor. Due to the great diversity of
entities involved in software measurement, a consistent framework is necessary to integrate the different
entities in the measurement process. In this work a Software Measurement Framework (SMF) is presented
to measure any type of software entity. In this framework, any software entity in any domain could be
measured with a common Software Measurement metamodel and QVT transformations. This work explains
the three fundamental elements of the Software Measurement Framework (conceptual architecture,
technological aspects and method). These elements have all been adapted to the MDE paradigm and to
MDA technology, taking advantage of their benefits within the field of software measurement. Furthermore
an example which illustrates the framework’s application to a concrete domain is furthermore shown.
1 INTRODUCTION
The current necessity of the software industry to
improve its competitiveness forces continuous
process improvement. This must be obtained
through successful process management (Florac, et
al., 2000). Measurement is an important factor in the
process life cycle due to the fact that it controls
issues and lacks during software maintenance and
development. In fact, measurement has become a
fundamental aspect of Software Engineering
(Fenton, et al., 1997).
Software Processes constitute the work base in a
software organization. Companies therefore wish to
carry out an effective and consistent software
measurement to facilitate and promote continuous
process improvement. To do this, a discipline for data
analysis and measurement (Brown, et al., 2004), and
measure definition, compilation and analysis in the
process, projects and software products, is needed.
The great diversity in the kinds of entities which
are candidates for measurement in the context of the
software processes points to the importance of
providing the means through which to define
measurement models in companies in an integrated
and consistent way. This involves providing
companies with a suitable and consistent reference
for the definition of their software measurement
models along with the necessary technological
support to integrate the measurement of the different
kinds of entities.
With the objective of satisfying the exposed
necessities, it is highly interesting to consider the
MDE (Model-Driven Engineering) paradigm
(Bézivin, et al., 2005) in which software
measurement models (SMM) are the principal
elements of the measurement process. Its main goal
is to ensure that the core artifacts in software
engineering processes will be models rather than
code, so that designs are expressed and managed in
the manner of models with a much higher level of
abstraction than the code. MDA (Model-Driven
Architecture) is the OMG proposal by which to
carry out the MDE Paradigm. The core of MDA is a
set of standards (MOF, QVT, OCL and XMI).
According to the QVT standard, the software
development process is a set of model
117
Mora B., García F., Ruiz F., Piattini M., Boronat A., Gómez A., Á. Carsí J. and Ramos I. (2008).
SOFTWARE MEASUREMENT BY USING QVT TRANSFORMATIONS IN AN MDA CONTEXT.
In Proceedings of the Tenth International Conference on Enterprise Information Systems - DISI, pages 117-124
DOI: 10.5220/0001677901170124
Copyright
c
 SciTePress

transformations, from an abstract to a specific level.
The requirements are in the more abstract level and
the code is in the more specific level.
Software measurement can benefit from the MDE
paradigm, providing integration and support to carry
out an automatic software measurement of any
software type. This implies that: a) the definition of
measurement models conform to a Software
Measurement metamodel; b) the definition of
generic measurement methods are applicable to any
model-based software artifact; and c) support for
computing measures, for storing results and for
enhancing decision making.
These aspects constitute the main interest of this
paper, in which the application of MDA principles,
standards and tools are used in software
measurement. The goal of this proposal is to develop
a generic framework to define measurement models
which conform to a common measurement
metamodel, and to measure any software entity with
regard to a domain metamodel. In order to develop
this proposal, MOMENT environment has been
used, which supports the automatic model
management MDA compliant.
Publications of García, Bertoa, et al. (2006);
García, Piattini, et al. (2006); and García, et al. (2007)
were used as a starting point for this work. These
works present FMESP, which consists of a
framework based on MOF Architecture. This includes
a software measurement ontology and metamodel,
and the GenMETRIC tool which is used to define
software measurement models, and to calculate
defined measures for these models. The ontology
permits the identification of all the concepts,
proportions exact definitions for all the terms and
clarifies the relationship between them. This paper
presents an adaptation of FMESP to MDA, which is
described in detail in following sections.
The remainder of the paper is organized as
follows. Section 2 provides an overview of related
works and Section 3 describes the Software
Measurement Framework (SMF), including
conceptual architecture, technological aspects, and
method. In Section 4 the use of the framework is
illustrated with an example. Finally, conclusions and
future works are outlined in Section 5.
2 RELATED WORKS
We have found numerous publications which deal
with tools that have important success factors in
software measurement efforts (Komi-Sirviö, et al.,
2001), which supply work environments and general
approximations (Kempkens, et al., 2000), or which
give architectures more specific solutions
(Jokikyyny, et al., 1999). Brown, et al. (2004)
include a list of tools which support the creation,
control and analysis of software measurements.
Auer, et al. (2003) furthermore examine various
software measurement tools, such as MetricFlame,
MetricCenter, Estimate Professional, CostXPert and
ProjectConsole, in heterogenic environments.
It is also possible to find certain proposals
through which to tackle software measurement
which are more integrated and less specific than in
the aforementioned cases. Palza, et al. (2003)
propose the MMR tool which is based on the CMMI
model for the evolution of software processes, and it
is possible to consult similar tools of Harrison,
(2004); Lavazza, et al. (2005); and Scotto, et al.
(2004). These proposals are, however, restricted to
concrete domains or to evaluation models of specific
quality.
Vépa, et al. (2006) present a metamodel which
allows the storage of measurement data, and a set of
transformations through which to carry out the
measurement of models based on a metamodel is
presented. This paper focuses upon the technological
aspects needed to implement the software
measurement with ATL technology, by offering the
user a variety of graphic representations of the
measurement results obtained.
This final proposal and that which is presented
here are complementary as they both focus upon two
key support elements of generic measurement: the
conceptual base, which is the main contribution of
FMESP, and technological implementation. Some
differences from technological point of view exist.
The measurements which are applied in the work
of Vépa, et al. (2006) are previously defined in the
ATL transformation archives. The measurable entities
are typical of the metamodels presented in this work
(KM3 and UML2). For example, the measurable
entities for a model which is expressed in km3 might
be package, class, attribute, reference etc.
The measurements in the proposal presented here
are defined by the user, i.e. the model transformation
needed to carry out the measurement it is not a
model previously defined, but this model is defined
according to the users needs. The measurement
definition is possible thanks to the software
measurement model, which contains all that is
relative to the measurement to be carried out in each
case. Moreover, the measurable entities are those
which are defined in their corresponding domain and
measurement metamodel (expressed in ecore). A
further difference is that SMF uses QVT.
ICEIS 2008 - International Conference on Enterprise Information Systems
118

3 SOFTWARE MEASUREMENT
FRAMEWORK
In order to carry out this proposal it was considered
of interest to adapt FMESP to the MDE paradigm.
The objective of this was to exploit the benefits that
the paradigm could contribute to software
measurement by, on one hand adopting the software
measurement metamodel defined in FMESP, and on
the other by evolving GenMETRIC to an
environment which would allow the definition of
software measurement models and the computation
of the models defined. All this would take place
within the context of models and model
transformations of the MDA architecture. The
Software Measurement Framework (SMF) is the
evolution of the FMESP, but is adapted to the MDE
paradigm and uses MDA technology.
The following subsections explain the conceptual,
technological and methodological elements which
are part of SMF.
3.1 Conceptual Architecture
Due to the necessity of having a generic and
homogeneous environment for software
measurement (García, Bertoa, et al., 2006; García,
Piattini, et al., 2006; García, et al., 2007), a
conceptual architecture and a tool with which to
integrate the software measurement are proposed. In
the following section, the main characteristics of this
proposal are described. In García, et al. (2007) a
more detailed description can be found.
MOF
Software
Measurement
Metamodel
Measurement
Models
Domain
Models
Data
Data
Data
Data
M3
M0
M1
M2
Integrated Measurement: Conceptual Framework
Model
Level
Meta-Model
Level
Meta-Meta-Model
Level
Domain
Metamodels
Domain
Metamodels
Data
Level
Figure 1: Conceptual framework with which to manage
software measurement.
The proposed software measurement described in
this paper is part of the FMESP framework (García,
Piattini, et al., 2006). The FMESP framework permits
representing and managing software processes from
the perspectives of modeling and measurement. We
focus on the measurement support of the framework
whose elements are detailed according to the three
layers of abstraction of metadata that they belong to,
according to the MOF standard. In Figure 1, the
conceptual architecture for integrated measurement is
represented.
As can be observed in Figure 1, the architecture
has been organized into the following conceptual
levels of metadata:
 Meta-MetaModel Level (M3). At this level, an
abstract language for the definition of
metamodels, is found. This is the MOF
language.
 Metamodel Level (M2). In the M2 level, two
generic metamodels which conform with this
framework are required. These are: the
Measurement Metamodel, to define specific
measurement models; and Domain
Metamodels, to represent the kinds of entities
which are candidates for measurement in the
context of the evaluation of the software
processes, such as, UML and Process
metamodels.
 Model Level (M1). Specific models are
included at this level. These models may be of
two types: Measurement Models, which are
examples of the measurement metamodel in
the M2 level and which are defined in such a
way as to satisfy some of the company’s
information needs; and Domain Models,
which are defined according to their
corresponding domain metamodels.
In order to establish and clarify the concepts and
relationships that are involved in the software
measurement domain before designing the
metamodel, an ontology for software measurement
was developed (García, Bertoa, et al., 2006). The
measurement metamodel was derived by using the
concepts and relationships stated in the ontology as a
base. The Software Measurement metamodel (which
is integrated in SMF) is organized around four main
packages (for greater detail see the work of García,
Bertoa, et al. (2006)):
 Software Measurement Characterization
and Objectives, which includes the
constructors required to establish the scope
and objectives of the software measurement
process.
 Software Measures, which aim at
establishing and clarifying the key elements in
the definition of a software measure.
 Measurement Approaches. This package
introduces the element of measurement
approach to generalize the different
SOFTWARE MEASUREMENT BY USING QVT TRANSFORMATIONS IN AN MDA CONTEXT
119

approaches used by the three kinds of
measures to obtain their respective
measurement results. A base measure applies
a measurement method. A derived measure
uses a measurement function. Finally, an
indicator uses an analysis model to obtain a
measurement result that satisfies an
information need.
 Measurement Action. This establishes the
constructs related to the act of measuring
software. A measurement (which is an action)
is a set of measurement results, for a given
attribute of an entity, using a measurement
approach. Measurement results are obtained as
the result of performing measurements
(actions).
3.2 Technological Aspects
In this section the technological aspects of SMF are
explained.
3.2.1 Adaptation to MDA
In Figure 2 the necessary elements for the FMESP
adaptation to MDA are presented according to MOF
levels.
Figure 2: Elements of the FMESP adaptation in a MDA
context.
As can be observed in Figure 2, two new
elements, namely the QVT Relations model and
metamodel, have been added to adapt the conceptual
architecture illustrated Figure 1 to MDA. The QVT
Relations Model (which is described in greater detail
in Section 3.2.2) is obtained automatically through a
transformation from a Measurement model. It
contains all the information necessary to carry out
the transformation of the SMF proposal. Ecore
language has been selected because it is a common
modeling language based on EMOF. EMOF is the
part of the MOF 2.0 specification that is used for
defining simple metamodels using UML-like
concepts.
3.2.2 QVT Relations Transformation
The QVT Relations model is the transformation
needed to perform the measurement. In this
transformation two source models are involved: a
Software Measurement model and a domain model;
the target model is the Software Measurement
Model with the measurement results (see Figure 2).
Due to the fact that the proposal is about generic
measurement, it is very important that the QVT
model is obtained in a generic way. The MDE
paradigm and MDA technology are applied for this
reason.
This transformation is obtained automatically
from the previous QVT transformation shown in
Figure 3. The QVT Relations model, called the
extended or final QVT Relations model, is obtained
from a QVT transformation, where there are two
source models: the basic or initial QVT Relations
model (which conforms to the QVT Relations
metamodel) and the Software Measurement model
(previously defined).
Figure 3: QVT Relations transformation model.
The extended QVT Relations model extends the
basic QVT Relations model with the following
aspects:
 Transformation Model: to obtain the
extended QVT Relations model, the source
M2
Software
Measurement
Metamodel
Domain
Metamodel
(1)
M3
M1
conforms to
conforms to
QVT
Relations
Metamodel
conforms to
conforms to
conforms to conforms to
Transformation
(
4
)
ECORE
Software Measurement
Model (target)
Software
Measurement
Model (2)
QVT
Relations
Model
Domain
Model (3)
Basic QVT-Relations Model
(.qvt)
Extended QVT-Relations Model
(.qvt)
QVT Transformatio
n
Software Measurement Model
(target)
Measurement Execution
Transformation
Software Measurement Model
(source)
ICEIS 2008 - International Conference on Enterprise Information Systems
120

model specification is needed. In this case,
there are two source models: the Software
Measurement model and the domain model.
Due to the fact that the Software Measurement
model is always the same, this model is
already defined in the basic QVT Relations
model. Therefore, only the domain model
needs to be defined. This information is taken
from the Software Measurement model which
contains all the measurement information.
 Relation Domain: in order to perform the
transformation, it is necessary to define the
checkonly domain rules. In this case there are
two, one for each source model: the domain
model and the Software Measurement model.
It is only necessary to define checkonly
domain of the domain model, because
checkonly domain of the measurement model
is already defined in the basic QVT Relations
model.
 Function: this contains the necessary OCL
queries to carry out the measurement. These
OCL queries are the implementations of the
“Measurement Action” package defined in the
Software Measurement Metamodel.
These elements are empty in the basic QVT
Relations model, and they are extended to obtain the
extended QVT Relations model, the transformation
model necessary to carry out the measurement. In
the Figure
4 all the Software Measurement process is
shown.
Figure 4. Software Measurement process.
3.2.3 Technological Environment
In this paper, the tool selected has been the model
management environment called MOMENT
(MOment manageMENT)(Boronat, et al., 2007).
This framework is integrated in the Eclipse platform.
It provides a set of generic operators to deal with
models through the Eclipse Modeling Framework
(EMF)("Eclipse Modelling Framework (EMF) Main
Page," 2007). The underlying formalism of the
model management approach is the algebraic
language Maude ("The Maude System," 2007).
From a functional point of view, MOMENT has
two components: OCL query execution (MOMENT-
OCL) and QVT Transformations (MOMENT-QVT).
MOMENT-OCL (Boronat, et al., 2006) has
implemented an editor integrated in the Eclipse
platform to check OCL invariants and to execute
OCL queries over instances of ecore models. It uses
Ecore models in the entire software development
process to store the OCL expressions by following
the Model Driven Engineering approach. One
advantage of this is the persistence mechanisms in
XMI that EMF provides automatically. In this work,
it has been used in order to check and validate the
OCL queries used in the QVT transformations. The
results have been shown by screen.
On the other hand, the MOMENT-QVT tool
(Queralt, et al., 2006) is a model transformation
engine that provides partial support for the QVT
Relations language. It implements the metamodel
definition QVT, given in the QVT standard, and
provides an editor for the QVT Relations language,
which permits the definition of model
transformations between EMF metamodels.
In order to carry out a QVT transformation in
MOMENT, a transformation textual specification
(coded by the Textual QVT Editor and stored in a
.qvtext) or, its equivalent QVT Relations model
(stored in a .qvt) can be used. This model conforms
to the QVT Relations metamodel and it is possible to
obtain it by parsing the textual specification.
3.3 Method
The necessary steps to carry out the software
measurement by using the SMF are explained below
(see Figure 2):
1. Incorporation of Domain Metamodel: the
measurement is made in a specific domain. This
domain must be defined according to its
metamodel (it is situated in the M2 level and it
conforms to the Ecore meta-metamodel).
2. Creation of Measurement Model: the
measurement model is created according to the
Software Measurement metamodel which is
integrated in SMF. This first model is the source
model, so the results are therefore still not
Basic transformation
Sofware
Measurement Model
source
Transformation
Extended
transformation
Domain Model
Software
Measurement Model
target
Input
Output
SOFTWARE MEASUREMENT BY USING QVT TRANSFORMATIONS IN AN MDA CONTEXT
121

defined, i.e. the “Measurement Action” package
from the Software Measurement metamodel is
still not instantiated.
3. Creation of Domain Model: which is defined
according to its corresponding domain metamodel
(created in the first step). The domain models are
the entities whose attributes are measured by
calculating the measurements defined in the
corresponding measurement models. Examples of
domain models are: the UML models (use cases,
class diagrams, etc.), or the E/R models.
4. Measurement Execution: the measurement
execution is carried out through QVT
transformation, in which, the measurement model
is obtained by starting from the two source
models (the measurement model and the domain
model) where the results are defined, i.e. the
“Measurement Action” package is instantiated.
The target measurement model is the extension of
the source measurement model. The measurement
results are calculated by running OCL queries on
the domain model.
An example of the method application is shown in
the following section.
4 EXAMPLE
To illustrate the benefits of the proposal, consider
the example of relational database measurement. For
greater simplicity, only the following elements are
shown in Figure 5: Measurement Method, Entity (to
which the measurement method is applied) and
Measurement result (the result is obtained by
executing the measurement method on the entity).
PrimaryKey
Foreign KeyTable
Attribute
Key
<<abstract>>
Model Element
<<abstract>>
RelationalSchema
Measuremet Method
(from Measurement Approaches)
Measurement Result
(from Meas urement Action)
Measurement
Entity Class
(from Characteriz ation and Objectives)
Entity
(from Characteriz ation and Objectives)
Figure 5: Relationship between Relational Database
(domain) Metamodel and SMM.
Furthermore, it is necessary for the domain
metamodel, in this case Relational Databases
domain, to have been previously chosen. Both
metamodels are independent (Figure 5), although
they are logically related. In Figure 5 the
measurement and domain metamodels have been
represented in a clear and a dark colour,
respectively.
In this example, the chosen measurement method
has been “COUNT elements of type TABLE”,
which is an instantiation of the abstract method
“COUNT elements of type X”.
In order to carry out the measurement, the
following steps (four steps) must take place:
1. Incorporation of Relational Databases metamodel
(represented in a dark colour in Figure 6).
2. Creation of measurement model conforms to
Software Measurement metamodel. For the
measurement method “COUNT elements of type
TABLE”, the values of Entity and Measurement
Method are Table and Count, respectively. The
Measurement Result is not still defined.
3. Creation of model conforms to the Relation
Database metamodel. In this case, the model
(relational schema) is a university domain
composed of five tables with their corresponding
primary keys (bold and shaded), foreign keys
(underlined and italic), and attributes (see Figure
6).
Figure 6: Relational Database model (relational schema).
The extended QVT Relations model was needed to
carry out the fourth step. This transformation is
obtained automatically (see section 3.2.2). The
extended elements are detailed below:
 Transformation Model: the target model is the
relational databases domain model.
 Relation Domain: the checkonly domain of
the relational schema domain is indicated (see
Figure 7).
Teacher
id
name
office
course
Course
course
name
ke
y
Student
id
name
degree
course
Department
key
name
URL
course
hl
School
id
name
URL
ICEIS 2008 - International Conference on Enterprise Information Systems
122

Figure 7: “Relation Domain” elements from extended
QVT Relations model.
 Function: this contains the OCL queries with
which to perform the measurement, in this
case, the queries necessary to implement the
“count element of type X” measurement
method where X is Table (see Figure 8).
Figure 8: “Function” elements from extended QVT
Relations model.
4. The source models used to carry out the
measurement are: the measurement model (2
nd
step), the domain model (3
rd
step) and the
extended QVT Relations model. The target model
obtained is the measurement model with defined
Measurement Result (see Figure 9). In this
example the value of Measurement Result is 5
(number of tables).
Figure 9: Measurement result.
In the same way as is illustrated with Relational
Databases, the method can be applied to any other
domains, such as for example, UML models, Project
Management or Business Processes, etc.
5 CONCLUSIONS AND FUTURE
WORK
In this paper, a generic framework for the definition
of measurement models based on a common
metamodel has been presented. The framework
allows the integrated management and measurement
of a great diversity of entities.
Following the MDA approach and starting from a
(universal) measurement metamodel, it is possible to
carry out the measurement of any domain by means
of QVT transformation, and this process is
completely transparent to the user.
With SMF, it is possible to measure any software
entity. The user task consists in selecting the
domain metamodel (the domain to be measured) and
defining the source models. The software metamodel
is integrated in the framework.
At the present time, a Software Measurement
Modeling Language (SMML) is being developed to
supply measurement engineers with the definition of
software measurement models according to the
proposed metamodel; this language will be
integrated in SMF.
Among related future works, one important work is
the realization of a plug-in based on Eclipse which
will supply the user with the data introduction and
the measurement process. This plug-in will enable
users to instantiate measurement models in an easy
and intuitive way. Other future work will be to align
our metamodel with the Software Metrics Meta-
Model (SMM) OMG proposal (OMG, 2007), which
is at present in its development phase. Finally, we
shall apply SMF to real environments to obtain
further refinements and validation.
SOFTWARE MEASUREMENT BY USING QVT TRANSFORMATIONS IN AN MDA CONTEXT
123

ACKNOWLEDGEMENTS
This work has been partially financed by the
following projects: ENIGMAS (Junta de
Comunidades de Castilla-La Mancha, PBI-05-058),
ESFINGE (Ministerio de Educación y Ciencia,
TIN2006-15175-C05-05) and META (Ministerio de
Educación y Ciencia, TIN2006-15175-C05-01).
REFERENCES
Auer, M., Graser, B., & Biffl, S. (2003). A Survey on the
Fitness of Commercial Software Metric Tools for
Service in Heterogeneous Environments: Common
Pitfalls Paper presented at the Ninth International
Software Metrics Symposium. (Metrics '03).
Bézivin, J., Jouault, F., & Touzet, D. (2005). Principles,
standards and tools for model engineering. Paper
presented at the 10th IEEE International Conference
on Engineering of Complex Computer Systems
(ICECCS’2005).
Boronat, A., & Meseguer, J. (2007). Algebraic Semantics
of EMOF/OCL Metamodels (No. Technical Report
UIUCDCS-R-2007-2904): CS Dept., University of
Illinois at Urbana-Champaign.
Boronat, A., Ramos, I., & Carsí, J. Á. (2006, Marzo 2003).
Definition of OCL 2.0 Operational Semantics by
means of a Parameterized Algebraic Specification.
Paper presented at the WAFOCA'06. First
International Workshop on algebraic foundations for
OCL and Applications.
Brown, M., & Dennis, G. (2004). Measurement and
Analysis: What Can and Does Go Wrong? 10th IEEE
International Symposium on Software Metrics
(METRICS'04), 131-138.
. Eclipse Modelling Framework (EMF) Main Page (2007),
from http://www.eclipse.org/emf
Fenton, N., & Pfleeger, S. L. (1997). Software Metrics: A
Rigorous & Practical Approach, Second Edition: PWS
Publishing Company.
Florac, W. A., Carleton, A. D., & Barnard, J. (2000).
Statistical Process Control: Analyzing a Space Shuttle
Onboard Software Process. IEEE Software, 17 (4)(4).
García, F., Bertoa, M. F., Calero, C., Vallecillo, A., Ruíz,
F., Piattini, M., et al. (2006). Towards a consistent
terminology for software measurement. Information
and Software Technology 48(8), 631-644
García, F., Piattini, M., Ruiz, F., Canfora, G., & Visaggio,
C. A. (2006). FMESP: Framework for the modeling
and evaluation of software processes. Journal of
Systems Architecture - Agile Methodologies for
Software Production 52(11), 627-639
García, F., Serrano, M., Cruz-Lemus, J., Ruiz, F., &
Piattini, M. (2007). Managing Software Process
Measurement: A Metamodel-Based Approach.
Information Sciences.
Harrison, W. (2004). A flexible method for maintaining
software metrics data: a universal metrics repository.
Journal of Systems and Software 72(2), 225-234
Jokikyyny, T., & Lassenius, C. (1999). Using the internet
to comunicate software metrics in a large
organization. Paper presented at the Proceedings of
GlobeCom'99.
Kempkens, R., Rösch, P., Scott, L., & Zettel, J. (2000).
Instrumenting Measurement Programs with Tools.
Paper presented at the PROFES 2000, Oulu, Finland.
Komi-Sirviö, S., Parviainen, P., & Ronkainen, J. (2001).
Measurement Automation: Methodological
Background and Practical Solutions-A Multiple Case
Study. Paper presented at the Seventh International
Software Metrics Symposium (METRICS'01),
London.
Lavazza, L., & Agostini, A. (2005). Automated
Measurement of UML Models: an open toolset
approach. Object Technology, 4(4), 115-134.
. The Maude System (2007), from
http://maude.cs.uiuc.edu/
OMG (2007). Architecture-Driven Modernization (ADM):
Software Metrics Meta-Model (SMM). OMG
Document: dmtf/2007-03-02: Object Management
Group.
Palza, E., Fuhrman, C., & Abran, A. (2003). Establishing
a Generic and Multidimensional Measurement
Repository in CMMI context Paper presented at the
28th Annual NASA Goddard Software Engineering
Workshop (SEW'03), Greenbelt (Maryland, USA).
Queralt, P., Hoyos, L., Boronat, A., Carsí, J. Á., & Ramos,
I. (2006, October 2006). Un motor de transformación
de modelos con soporte para el lenguaje QVT
relations. Paper presented at the Desarrollo de
Software Dirigido por Modelos - DSDM'06 (Junto a
JISBD'06), Sitges, Spain.
Scotto, M., Sillitti, A., Succi, G., & Vernazza, T. (2004). A
relational approach to software metrics. Paper
presented at the Proceedings of the 2004 ACM
symposium on Applied computing (SAC'2004),
Nicosia, Cyprus.
Vépa, É., Bézivin, J., Brunelière, H., & Jouault, F. (2006).
Measuring Model Repositories. Paper presented at the
Model Size Metrics Workshop at the MoDELS/UML
2006 conference, Genoava, Italy.
ICEIS 2008 - International Conference on Enterprise Information Systems
124