International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882

59
Volume 6, Issue 1, January 2017

Comparative Study and Review on Object Oriented Design Metrics

Sonam Gupta Anish Lazrus

Shri Shankaracharya Group of Institution Shri Shankaracharya Group of Institution
Dept. of Information & Technology Dept. of Information & Technology
Bhilai, Chhattisgarh, India Bhilai, Chhattisgarh, India
[email protected] [email protected]

Abstract inconsequential to apply them to organized programming.

The best illustrations to programming advancement issues The above figure demonstrates the various leveled structure
are frequently touted as object oriented procedures. Software of the metrics.
complexity, computation time, project efforts are the essential
metrics which are measured by most popular technique, Object II. CK METRICS MODEL
Oriented Design. There are different methodologies through
which we can discover the product cost estimation and Chidamber and Kemerer characterize the purported CK
predicates on different sorts of deliverable things. Question metric suite [3]. CK metrics have produced a lot of intrigue
arranged measurements guarantees to diminish cost and the and are right now the most understood well suite of
upkeep exertion by serving as early indicators to gauge estimations for OO software [4]. Chidamber and Kemerer
programming flaws. Such an early evaluation enlarges the proposed six metrics; the accompanying exchange
nature of the final programming. This paper reviews some of demonstrates their metrics.
the metrics of Object Oriented metrics. A correlation table is
kept up through which we can break down the contrast A. Weighted Method per Class (WMC)
between all the object oriented measurements viably and
WMC measures the intricacy of a class. Many-sided
effectively.
quality of a class can for instance be figured by the
cyclomatic complexities of its techniques. High estimation of
Keywords Software Engineering Metrics, Object
WMC demonstrates the class is more mind boggling than that
Oriented Models, Inheritance Tree.
of low values.
I. INTRODUCTION
B. Depth of Inheritance Tree (DIT)
Various programming metrics identified with software
DIT metric is the length of the most extreme way from
quality confirmation have been proposed in the past are as yet
the hub to the base of the tree. So this metric figures how far
being proposed. A few books showing such metrics exist, for
down a class is announced in the legacy chain of importance.
example, Fenton's [1], Sheppard's [2] and others. The vast
The accompanying figure demonstrates the estimation of DIT
majority of these metrics are available to all programming
for a basic class progression. DIT speaks to the intricacy of
languages, a few metrics apply to a particular arrangement of
the conduct of a class, the multifaceted nature of plan of a
programming language. Among metrics of this kind, are
class and potential reuse.
those that have been proposed for objectoriented
Subsequently it can be difficult to comprehend a
programming language.
framework with numerous legacy layers. Then again, a huge
These days, a quality designer can look over a massive
DIT esteem demonstrates that numerous strategies may be
amount of objectoriented metrics. The question posed is not
reused.
the absence of metrics but rather the choice of those metrics
which meet the particular requirement of every software
C. Number of Children (NOC)
project. A quality architect needs to confront the issue of
selecting the fitting arrangement of metrics for his product
This metric measures what number of sub-classes will
estimations. Various objectoriented metrics exploits the
acquire the techniques for the parent class. As appeared in
information picked up from metrics utilized as a part of
above figure, class C1 has three kids, subclasses C11, C12,
organized programming and adjust such estimations in order
and C13. The measure of NOC around shows the level of
to fulfill the requirements of objectoriented programming.
reuse in an application. On the off chance that NOC develops
Then again, other objectoriented metrics have been created
it implies reuse increments. Then again, as NOC expands, the
particularly for objectoriented programming and it is

www.ijsret.org
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
60
Volume 6, Issue 1, January 2017

measure of testing will likewise increment since more found: I1 I2 (= {a, b, e}) and I3. Here, one sets of strategies
youngsters in a class demonstrate more duty. Thus, NOC who share no less than one example variable (I1 and I2). So
speaks to the exertion required to test the class and reuse. LCOM = 2-1 =1. [13] States Most of the strategies
characterized on a class ought to utilize a large portion of the
information individuals the vast majority of the time.
On the off chance that LCOM is high, techniques
might be coupled to each other through qualities and after
that class configuration will be intricate. Thus, planners ought
to keep union high, that is, keep LCOM low.

III. LITERATURE SURVEY

Kumar Rajnish et al. [7]. The inheritance metrics give us

information about the inheritance tree of the system.
Inheritance is a key feature of the Object-Oriented (OO)
paradigm. This mechanism supports the class hierarchy design
and captures the IS-A relationship between a super class and
its subclass. Several OO inheritance metrics have been
Fig.1. value of DIT in class hierarchy
proposed and their reviews are available in the literature. In
doing so, an attempt has been made to define empirical
D. Coupling between articles (CBO)
relationship between the proposed inheritance metric suites
with considered existing inheritance metrics and the focus was
The possibility of this metrics is that a question is coupled
on which how the inheritance metric suites were correlated
to another protest if two question follow up on each other. A
with the existing ones. Data for several C++ classes has been
class is combined with another if the techniques for one class
collected from various sources.
utilize the strategies or characteristics of alternate class. An
expansion of CBO shows the reusability of a class will
Kailash Patidar et al. [8]. Software engineering aims at
diminish. Consequently, the CBO values for every class
development of high-quality software and tools to promote
ought to be kept as low as could be allowed.
quality software that is stable and easy to maintain and use. In
order to assess and improve software quality during the
E. Response for a Class
development process, developers and managers use, among
RFC is the quantity of strategies that can be summoned in other means, ways to automatically measure the software
light of a message in a class. Pressman [5] States, since RFC design of object oriented programming. Cohesion, coupling,
expands, the exertion required for testing likewise increments and complexity are common types of such metrics. The
in light of the fact that the test grouping develops. On the off cohesion of a module indicates the extent to which the
chance that RFC builds, the general plan multifaceted nature components of the module are related. A highly cohesive
of the class increments and turns out to be difficult to get it. module performs a set of closely related actions and cannot be
Then again bring down qualities demonstrate more split into separate modules.
noteworthy polymorphism. The estimation of RFC can be
from 0 to 50 for a class12, a few cases the higher esteem can K.K.Aggarwal et al. [9]. The increasing importance of
be 100-it relies on upon venture to extend. software measurement has led to development of new software
measures. Many metrics have been proposed related to various
F. Lack of Cohesion in Methods (LCOM) constructs like class, coupling, cohesion, inheritance,
information hiding and polymorphism. But there is a little
This metric uses the thought of level of closeness of understanding of the empirical hypotheses and application of
strategies. LCOM measures the measure of cohesiveness many of these measures. It is often difficult to determine
present, how well a framework has been planned and how which metric is more useful in which area. As a consequence,
complex a class is [6]. LCOM is a check of the quantity of it is very difficult for project managers and practitioners to
strategy combines whose comparability is zero, short the tally select measures for object-oriented systems. A key element of
of technique matches whose similitude is not zero. any engineering process is measurement. Measures are used to
Raymond [6] examined for instance, a class C with 3 better understand the attributes of the model that we create.
techniques M1, M2, and M3. Let I1= {a, b, c, d, e}, I2= {a, b, But, most important, we use measurements to assess the
e}, and I3= {x, y, z}, where I1 is the arrangement of example quality of the engineered product or the process used to build
factors utilized by technique M1. So two disjoint set can be it.

www.ijsret.org
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
61
Volume 6, Issue 1, January 2017

Gopal Goyal et al. [10]. A large numbers of metrics have Size metrics based on the count of various attributes.
been proposed for measuring properties of object-oriented Whereas Inheritance based on operation to be reused.
software such as size, inheritance, cohesion and coupling. The External metrics are used to examine and reuse.
Various categories of metrics are:
coupling metrics presented in this paper exploring the
Class Size
difference between inheritance and interface programming.
Number of Operation
Object-oriented design and programming is the dominant Average Operation Size
development paradigm for software systems today. Recently Operation Complexity
so many languages are object-oriented (OO) programming
languages. In object oriented programming we provide
abstraction by classes and interfaces. TABLE I. shows various source construct and its metrics

Dr. K.P.Yadav et al. [11]. The increasing importance of Object

SNO Source Construct Metrics Oriented
software measurement has led to development of new software
Structure
measures. Many metrics have been proposed related to various Cyclomatic
constructs like class, coupling, cohesion, inheritance, Function
Complexity
information hiding and polymorphism. The central role that Lines of
1 Traditional Metrics function
software development plays in the delivery and application of Codes
information technology, managers are increasingly focusing Comment
function
on process improvement in the software development area. It Percentage
is very difficult for project managers and practitioners to select Weight
function
Method per
measures for object-oriented system. This demand has spurred and class
class
the provision of a number of new and/or improved approaches Response for class and
to software development, with perhaps the most prominent New Object Oriented a class message
being object-orientation (OO). 2
Metrics Coupling
between coupling
IV. REVIEW OF METRICS objects
Depth of
Inheritance
Inheritance
Two important of the OO metrics for object oriented
software development are.
V. CONCLUSION
a. Chen Metrics
b. Lorenz and Kidd Metrics This review paper presented some of the metrics used for
evaluation of object oriented system. Metrics are the indicators
of the performance of the system. Paper has analyzed and
A. Chen Metrics
compare various software metrics. As a number of software
Chen et al. [12] proposed some major metrics for processes are increasing, the complicity also increasing. There
measurement performance, throgh which it can define "What will always be need of software metrics to test system
is the behavior of the metrics in object-oriented design". performance. In this paper, we have demonstrated all the
They describe the behavoirs like: software metrics for object oriented systems. They are basics
for measuring various characteristics like size, time,
Class Coupling Metric complexity, performance etc. Paper intended to provide
attribute Complexity Metric reviews and survey of software metrics to help researchers and
Cohesion Metric
practitioners for better understanding of Object Oriented
Reuse Metric
metrics in short and selective way.
B. Lorenz and Kidd Metrics

Lorenz and Kidd et al. [13] proposed set of metrics REFERENCES

that can be grouped in 4 category.
[1] N. Fenton et al, Software metrices: a rigorous and
Size practical approach, International Thomson Computer
Inheritance Press 1996
Internal [2] M.J. Sheppard & D. Ince, Derivation and Validation of
External. Software Metrics, Clarendon Press, Oxford, UK, 1993

www.ijsret.org
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
62
Volume 6, Issue 1, January 2017

[3] Chidamber, Shyam, Kemerer, Chris F. "A Metrics Suite

for Object- Oriented Design" M.I.T. Sloan School of
Management E53-315, 1993
[4] C. Jones, Estimating Software Costs: Bringing Realism
to Estimating, 2nd Edition, Mc Graw Hill, New York,
2007
[5] Roger S. Pressman: Software Engineering, Fifth
edition, ISBN 0077096770.
[6] Raymond, J. A, Alex, D.L: A data model for object
oriented design metrics, Technical Report 1997, ISBN
0836 0227.
[7] Kumar Rajnish , Arbind Kumar Choudhary, Anand
Mohan Agrawal. INHERITANCE METRICS FOR
OBJECT-ORIENTED DESIGN,International Journal of
Computer Science & Information Technology (IJCSIT),
Vol 2, No 6, December 2010
[8] Kailash Patidar, RavindraKumar Gupta,Gajendra Singh
Chandel , International Journal of Advanced Research in
Computer Science and Software Engineering 3(3), March
- 2013, pp. 517-521
[9] K.K.Aggarwal, Yogesh Singh, Arvinder Kaur, Ruchika
Malhotra,: Empirical Study of Object-Oriented Metrics,
in Journal of Object Technology, vol. 5. no. 8, Novmeber-
December 2006, pp. 149-173
[10] Gopal Goyal, Sachin Patel: Importance of Inheritance and
Interface in OOP Paradigm Measure through Coupling
Metrics, International Journal of Applied Information
Systems (IJAIS) ISSN : 2249-0868, Foundation of
Computer Science FCS, New York, USA, Volume 4
No.9, December 2012
[11] Dr. K.P. Yadav, Ashwini Kumar, Sanjeev Kumar: bject
Oriented Metrics Measurement Paradigm, IJMIE Volume
2, Issue 6 ISSN: 2249-0558
[12] Chen, J-Y., Lum, J-F.: "A New Metrics for Object-
Oriented Design." Information of Software Technology
35,4(April 1993):232-240.
[13] M. Lorenz, J. Kidd, Object Oriented Software
Metrics, Prentice Hall, NJ, (1994)

www.ijsret.org