Hardware Reference
In-Depth Information
avoid different interpretations for the same model. Therefore, it is very important to
be able to evaluate and improve the quality of the conceptual models in order to im-
prove the quality of the software products generated by using MDD technologies.
One important quality issue to be evaluated is the amount of defects that the con-
ceptual models used in MDD environments can have. In many cases, the defect detec-
tion is performed by the MDD compilers, which presents disadvantages such as the
extra complexity included in the compiler, and also the identification of defects with
regard to specific technical platforms (i.e. Java, C#, etc.). To overcome the limitations
of the defect detection procedures embedded in MDD compilers, it is necessary a
defect detection procedure that can be applied directly in the conceptual models. Tak-
ing into account that in terms of the management of software projects: (1) it is widely
accepted that is essential to know the functional size of applications in order to suc-
cessfully apply estimation models, effort models, and budget models [26], and (2) the
measurement of the functional size in conceptual models allows the project leader to
generate indicators in early stages of the development cycle of a software product; we
advocate the use of a measurement procedure to detect defects in early stages of the
software product. Thus, the aim of this work is to present how a functional size meas-
urement procedure that allows the measurement of conceptual models can helps in the
detection of defects that can have the conceptual models used in MDD environments.
The rest of the paper is organized as follows: Section 2 presents a brief background
and a set of relevant related works. Section 3 presents the functional size measure-
ment procedure for the conceptual model of a specific MDD approach that is used to
apply the introduced ideas into practical settings. Section 4 presents how a measure-
ment procedure can be used to identify types of defects of the conceptual models.
Finally, Section 5 presents some conclusions and points out future work.
2 Background and Related Work
The ISO/IEC 14143-1 [11] standard defines
functional size
as the size of the software
derived by quantifying the functional user requirements. This standard also defines a
Functional Size Measurement (FSM) as the process of measuring the functional size.
In addition, this standard defines a FSM method as the implementation of a FSM that
is defined by a set of rules, which is defined in accordance with the mandatory fea-
tures defined in the ISO/IEC 14143-2 [12].
In order to measure the functional size of software applications, four measurement
methods have been recognized as standards: IFPUG FPA [18], MK II FPA [19],
NESMA FPA [20], and COSMIC FFP [17]. The first three methods are based on the
Function Point Analysis proposal [3]. These FPA-based methods have several limita-
tions for the correct measurement of systems: for instance, they only consider the
functionality of the system that the human user observes, they have units that are hard
to understand; they do not consider the functionality that allows communication be-
tween layers in systems with a layer-based architecture, etc. To overcome the limita-
tions of FPA-based measurement methods, the COSMIC measurement method was
defined as the second generation of functional size measurement methods. COSMIC
uses a mathematical function to aggregate the functional size of the functional proc-
esses specified in the conceptual models and is not limited by maximum values to
