University of Sindh Journal of Information and Communication Technology
(USJICT)
Volume 4, Issue 4, December 2020
ISSN-E: 2523-1235, ISSN-P: 2521-5582 © Published by University of Sindh, Jamshoro
Website: http://sujo.usindh.edu.pk/index.php/USJICT/
Why we need COCOMO III?
Sana Iqbal
1
, Dr. Muhammad Shahab Siddiqui
2
, Zazilah May
3
1, 2
Centre for Computing Research, Department of Computer Scienceand Software Engineering,
Jinnah University for Women, Karachi, Pakistan.
3
Electrical and Electronic Engineering Department, Universiti Teknologi PETRONAS,Malaysia
sanaiq2@gmail.com
1
; shahab.siddiqui@ieee.org
2
; zazilah@utp.edu.my
3
Abstract: In software development phase, projects are often delayed and cross the budget. This problem occurs due
to the inaccurate estimation of the project cost and effort for the project development. These estimations are
dependent on the project size that may use metrics, LOC and function point analysis. Variety of techniques exist for
project cost estimation, which includes algorithm-based models and expert opinion. Expert opinion is one of the
ways to estimate project effort. But it doesn’t support project re-estimation during its lifecycle, which helps in
tracking project. Cost estimation using algorithmic model plays important role in delivering quality product.
COCOMO is being used for cost estimation; but it needs to be updated based on new process models therefore
another version of COCOMO should be developed. This paper draws attention towards what should be in proposed
COCOMO III, to avoid project uncertainties, risks and deliver a quality product within budget and timeliness.
Keywords: COCOMO, Cost Estimation, Effort Estimation, Software Cost Drivers
I. INTRODUCTION
Cost Estimation problem often occurs in software projects
development, which results in a worst project
management. Software cost and effort estimation plays an
important role in SDLC (Software development life
cycle). It has been used in planning, monitoring and
budgeting the development activities. It helps in timely
and within budget delivery of a project. To solve cost and
effort estimation problems variety of models were
introduced. The one that is commonly used is COCOMO
(Cost Constructive Model) [1,2] developed by Barry W.
Boehm in 1981. Therefore, this version was referred as
COCOMO I or COCOMO 81[3, 4] and later one as
COCOMO II, which was introduced in 1997. COCOMO I
play an important role in estimating cost and effort at an
early phase while COCOMO II gives an effective
calibration process. But in case of COCOMO I estimation
at an early phase may leads to an estimation failure
whereas COCOMO II [5, 6, 7] requires calibration to a
specific context which can be difficult to adapt where
there is incomplete or limited data. Therefore, there is a
need to develop another improved version of COCOMO
as it’s been 22 years since the COCOMO II is updated
and calibrated new data to software engineering.
COCOMO II uses scaling factors, line of code and cost
drivers to estimate the project cost but it is still lacking
the accuracy. To overcome this now it’s the time that
COCOMO III should be developed.
This paper is organized as Section II covers literature
review and related work. Section III briefly describes
SWOT analysis for COCOMO III. In Section IV the need
of COCOMO III is discussed. COCOMO III parameters
are introduced in Section V. The paper is concluded in
Section VI. Section VII is highlighting future directions.
II. LITERATURE REVIEW AND RELATED WORK
COCOMO is a model for estimation of cost, effort and
schedule of software projects developed by Barry Boehm.
It uses regression basic formula having parameters (cost
drivers) derived from historical data and characteristics of
current projects. This model was first developed in 1981
and it’s another version named as COCOMO II was
developed in 1997 along with cost drivers. These drivers
are helpful in understanding various factors that can affect
the cost of the project. But now these cost drivers [8, 9,
10] are not applicable as they were in the past there is a
need to develop the improved version of the model
according to the latest trends.
COCOMO I help in understanding the cost consequences
of decisions in development and support of software
project. Whereas COCOMO II offers great amount
parameters along with the software cost estimation [11,
12]. COCOMO I is a chain of three sub model each of
them is more detailed than other. The first sub model is
known as Basic COCOMO which is responsible for
measuring the lines of code (LOC) for calculating the
development cost and effort of a software project. Basic
COCOMO is further divided into three types of project
modes that are Organic, semi-detached and embedded
projects [13, 14]. Organic projects are simple and
developed by good experienced small team, semi-