International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 23 Issue 9 – DECEMBER 2016.

A Comparative Analysis of Function Point

Analysis for the Deployment of Software
Projects

I. Priyanka#1
#
M.Phil. Scholar, Dept. of Computer Science, S.vellaichammy Nagar college, Madurai, India

management information system (MIS), database intensive.

Abstract— The appearance of the Function Point technique is Function Points are indirect quantitative measures of
practice of software measurement, with respect to the use of the application software functionality and size. This procedure is
traditional “Lines of Code approach”. A FP count, however, composed of three logical divisions, finding the unadjusted
requires a complete Functional Specifications of the software
system, to be performed. There are two situations in which
function point count, value adjustment factor, and Function
having an estimation method. The first case occurs when the Points. Finding the unadjusted function point count consists
development or enhancement project is in such an early phase of counting the number of external inputs, external outputs,
that it is simply not possible to perform a FP count. an external inquiries, internal logical files, and external interface
evaluation of the existing software asset is needed for required files. Finding the value adjustment factor consists of rating
time and resources to perform a detailed FP calculation are not system, input and output, and application complexity. Finding
available. Based on these situations, the demand of methods for
estimating - not counting - Function Points has risen from the Function Points consists of factoring unadjusted function
organizations involved in software business. This paper points and value adjustment factor together.
presents, the estimation methods (Early Function Points, ILF Function Points have two distinct purposes. The first one is
Models, Backfiring) and a general benchmarking model, useful basis for software measurement, comparison, and analysis.
for the evaluation of any additional method, as well. The second is to determine software size for input into
software cost estimation models are based on empirical cost
Index Terms— software measurement, size, function point,
estimating relationships (CERs) between Function Points and
estimation, benchmarking
effort.
B. Related Work
I. INTRODUCTION
FP Analysis is an efficient software sizing technique but, Estimation by expert [18][19], analogy based estimation
considering the standard IFPUG Counting Practices, it schemes [20], algorithmic methods including empirical
implies detailed set of descriptive documentation of the user methods [21], rule induction methods [22], artificial
functional requirements for any software application to be neural network based approaches [23] [24] [25], Bayesian
measured. There are two estimation method, compatible to network approaches [26], decision tree based methods [27]
the standard rules for FP, could be decisive. The first case and fuzzy logic based estimation schemes [28][29]. The
occurs when the software development is in an early phase estimation parameters are commonly derived from empirical
that it is simply not possible to perform a FP count.A standard data. Approximate effort and cost estimation of software
count requires the identification of elements need for effort, applications continues to be a critical issue for software
time and cost forecasts based on size evaluation in Functional project managers [32]. Expert judgment remains widely
Specification. The second case occurs when an evaluation of used in applying statistics and machine learning techniques to
software size asset is needed, but the required time and predict software project effort [33][34]. Hardware costs,
resources to perform FP calculation are not available. Various travel and training costs and effort costs are the three
estimating methods have been proposed to respond to the principal components of cost[36][37]. Many research papers
need of evaluating the size of a software project. is able to provide accurate effort/cost estimation is still a
challenge for many reasons. They include: (i) the
II. BACKGROUND AND RELATED WORK uncertainty in collected measurement, (ii) the estimation
methods have many drawbacks and (iii) the cost drivers is
A. Background
along with the development environment which might not be
Function Points is a cost estimating relationship (CER) to clearly specified [38]. The most popular algorithmic
software costs. Once determined,Function Points are widely estimation models include Boehm‟s constructive cost model
reported to be well suited for measuring the size of (COCOMO) [39]. Thus, accurate estimation methods, for
example, the FP method, have gained increasing importance
[40]. . The size is determined by identifying the components

46
 International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)
ISSN: 0976-1353 Volume 23 Issue 9 – DECEMBER 2016.

of the system as seen [41] by the end-user: the inputs, predictions are combined. In shortly, they provide iterated
outputs, inquiries, interfaces [42] to other systems and cycle of anonymous estimations by each expert, until the
logical internal files [43]. The components are classified estimates converge to an acceptable range. The conclusion is
as simple, average or complex. All values are then scored a group estimate arrived at by consensus. The group estimate
and the total is expressed in unadjusted FPs (UFPs) is typically a better overall estimate than any individual
performance and complexity of processing can be used to prediction.
weigh the UFP. The result of these computations is correlates
to system size. The FP metric does not correspond to any
C. Three-Point Estimation Techniques
actual physical attribute of a software system [47,48] (such as
lines of code or the number of subroutines) is useful for Three-Point Estimation Technique is
measuring productivity, and estimating the amount a to improve direct estimation, when more values are
development effort and time needed for a project [49,50]. provided by estimators. Given the Minimum, the Most Likely,
The total number of FPs depends on processing logic types and the Maximum Value for the size,
in the following five classes [51]. It is well documented the estimate is:
that the software industry suffers from frequent cost Est.Size = (Min + 4×MostLikely + Max) / 6
overruns [52]. A contributing factor is, we believe, the with standard deviation:
imprecise estimation terminology in use. A lack of clarity and s= (Max - Min) / 6
precision [53] in the use of estimation terms reduces the
interpretability of estimation [54] accuracy results, makes D. Simple Analogy Method
the communication of estimates difficult and lowers the Simple Analogy Method is in the historical database, that is
learning possibilities [55]. Number of enhancements to "similar" to the application under estimation. The found
adjustment factors is introduced. This model is grouping into implemented system provides a quick estimate of the new
three groups. They are “System complexity”, “I/O project size. Further investigation likely leads to Structured
complexity” and “Application complexity”. Analogy

III. COMPARATIVE OF ESTIMATION METHODS

E. Structured Analogy Methods
Effort (and size) estimations are critical to the success of a The estimator compares the proposed application to one
software project [8]. Of the different techniques presented in or more existing applications: This type of application,
prior literature [9], the most popular include algorithmic and establish an initial prediction and passing from a Simple
parametric models, expert judgment, and reasoning by Analogy to a Structured Analogy, the differences and
analogy [10]. According to Heemstra’s survey, 29 different similarities are identified and used explicitly in a
software-based cost models have been proposed since 1966 mathematical way to adjust the estimate. A concept of
[11].These models usually take a software size as input to “distance” among systems may be defined and used to
estimate development effort. The LOC-based models are prioritize the choices.
mostly nonlinear, and their estimated effort is at least
quadratic to the size. The generation process of counting
function points involves nonlinear computation. Machine IV. EXPERIMENTAL RESULTS AND ANALYSIS
learning models for size and/or effort estimation are rather
A. Data Sets
more recent; they include case-based reasoning [13], fuzzy
logic [14], neural network [15], and many others [16]. Most
such approaches report results that are comparable to those of
other techniques [17]. However, because the data needed by
these models most can be applied only during the later stages
of the software development process. The most common
estimation methods are,
A. Direct Estimation Methods
Direct Estimation Methods approach to estimation may
results in very accurate estimates, although it is entirely
dependent on the experience of the expert(s). Sometimes, in
case of many experts collaborating to the same effort, it may
be difficult for the estimates to converge to a unique value.
TABLE.1. DEVELOPMENT EFFORT
The estimate itself may be influenced by subjective factors, as
personal relationships, contractual aspects, and so on.

B. Delphi Or Shang Techniques

The most commonly used procedures, by which individual

47
 International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)
ISSN: 0976-1353 Volume 23 Issue 9 – DECEMBER 2016.

programmers a method to build their own model, which

captures more of a given application domain’s characteristics
and is easier to use. The demonstration of this approach offers
an unique benefit in that it makes the function classification
more suitable for a particular application domain so that
function point counting can be conducted by programmers
themselves instead of by a certified FPA expert. We believe
such an approach should be applicable to domains when FPA
works since it still stays with the FPA spirit. However, it may
shares the limitations of FPA also. To verify its usefulness, it
is no doubt that more experimental data from a wider range of
application domains and environments should be collected to
Figure 1: chart
deepen the investigation into the trade-off between generic
versus specific. Atleast two research directions exist. First,
studies could evaluate whether the tailored FPA model always
fits a business application domain and still maintains
comparable estimation accuracy. Second, the ideas presented
here in might be adopted to another estimation model and
determine how it performs. Each direction will lead to more
detailed research findings..

REFERENCES
[1] AD/M Productivity Measurement and Estimate Validation; Allan
Albrecht, IBM Corporate Information Systems and Administration;
Figure 2: graph 1985
[2] Function Point Analysis: Difficulties and Improvements, C. R.
V. CONCLUSION Symons, IEEE Transactions on Software Engineering, 1985.
[3] Software Function, Source Lines of Code and Development Effort
A number of different models and effort estimation Prediction: A Software Science Validation; Albrecht A.J., & Gaffney
methods have been developed in the past four decades. This J.E.; IEEE Transactions on Software Engineering, Vol SE -9, No. 6;
clearly indicates the awareness among the researchers of the November 1983.
[4] Empirical Studies of the Assumptions Underlying Software Cost
need to improve effort estimation in software engineering. Estimation Models; B. A. Kitchenham, NCC Ltd; 1991.
Many factors have impact on the software development [5] Function Point Counting Practices Manual: Release 3.0 ; IFPUG
process. These factors are not only human, technical but also Counting Practices Committee; 1990.
[6] Software Sizing and Estimating: Mk II FPA; Symons C.R.; John Wiley
political and their impact can never be fully predicted. The & Sons; 1991
even insufficiently accurate estimates are far better than none. [7] A Cooperative Industry Study: Software Development/Maintenance
We have illustrated result of two approaches for measuring Productivity; Xerox Corporation;1985
[8] M. Jorgensen and D. I. K. Sjoberg, “Impactof effort estimates on
the size in the estimation process in our work. If the estimation
software project work,” Information and Software Technology, Vol.
is done accurately, it decreases error. We have analyzed the 43, 2001, pp. 939-948.
values of different matrix under function point as well as use [9] C. Gencel and O. Demirors, “Functional size measurement revisited,”
case point method. The relationship among the matrix of ACM Transactions on Software Engineering and Methodology, Vol.
17, 2008, pp. 1-36.
these methods shows the practical results. Hence we [10] B. Barry, C. Abts, and S. Chulani, “Software development cost
conclude that the estimation process through use case point is estimation approaches −A survey,” Annals of Software Engineering,
better than the function point. The requirement & design is Vol. 10, 2000, pp. 177-205.
[11] A. L. Lederer and J. Prasad, “A causal model for software cost
more clearly shown in use case point model than in function estimating error,” IEEE Transactions on Software Engineering, Vol.
point model. The estimation using use case 24, 1998, pp. 137-148.
point shows practical reality of development. Hence the [12] A. J. Albrecht and J. E. Gaffney Jr., “Software function, source lines of
code, and development effort prediction: A software science
accuracy of estimation depends upon method used for validation,” IEEE Transactions on Software Engineering, Vol. 9, 1983,
estimation also. Although effort estimation is a critical step pp. 639-648.
for the success of a software project, in practice, many [13] R. Bisio and F. Malabocchia, “Cost estimation of software projects
projects still use ad hoc methods to conduct this task. This through case-base reasoning,” in Proceedings of Case-Based
Reasoning Research and Development, 1995, pp. 11-22.
phenomenon may result because most of the generic [14] O. D. Lima, P. M. Farias, and A. D. Belchior “Fuzzy modeling for
algorithmic estimation models are difficult to be adopted, in function points analysis, ”Software Quality Journal, Vol. 11, 2003, pp.
that they require the collection of many detailed items that 149-166.
[15] J. Hakkarainen, P. Laamamen, and R. Rask, “Neural networks in
may affect the size (and effort) associated with an application. specification level software size estimation,” in P. K. Simpson, ed.,
This research presents a different approach that trades Neural Network Applications, IEEE Technology Update Series, 1993,
“generic” with “specific” and greatly simplifies the estimation pp. 887-895.
[16] A. Heiat, “Comparison of artificial neural network and regression
model to enable common programmers to use it. Although the models for estimating software development effort,” Information and
new approach may not be as formal as a normal estimation Software Technology, Vol. 44, 2002, pp. 911-922.
model, it does reduce the difficulties of using it by giving

48
 International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)
ISSN: 0976-1353 Volume 23 Issue 9 – DECEMBER 2016.
[17] A. R. Gray and S. G. MacDonell, “A comparison of techniques for IEEE/WIC/ACM Int. Conf. on Web Intelligence and Intelligent Agent
developing predictive models of software metrics,” Information and Technology, Milano, Italy, 2009, vol. 3, pp. 77–80
Software Technology, Vol. 39, 1997, pp. 425-437. [41] B.W. Boehm, “Software Engineering Economics,” Prentice Hall,
[18] SaleemBasha, Dhavachelvan.P. “Analysis of Empirical Software 1981.
Effort Estimation Models” (IJCSIS) International Journal of [42] B.W. Boehm, E. Horowitz, R. Madachy, D. Reifer, B. K. Clark,
Computer Science and Information Security, Vol. 7, No. 3, 2010. B.Steece, A. W. Brown, S. Chulani, and C. Abts, “Software
[19] Jorgen MS joberg D.I.K, “The Impact of Customer Expectation on CostEstimation with COCOMO II,” Prentice Hall, 2000.
Software Development Effort Estimates” International Journal of [43] F. J. Heemstra, "Software cost estimation," Information and Software
Project Management, Elsevier, pp 317-325, 2004. Technology, vol. 34, pp. 627-639, 1992
[20] Chiu NH, Huang SJ, “The Adjusted Analogy-Based Software Effort [44] N. Fenton, "Software Measurement: A necessary Scientific
Estimation Based on Similarity Distances,” Journal of Systems and Basis,"IEEE Transactions on Software Engineering, vol. 20, pp.
Software, Volume 80, Issue 4, pp 628-640, 2007. 199-206, 1994.
[21] Kaczmarek J, Kucharski M, “Size and Effort Estimation for [45] Barry Boehma, Chris Abtsa andSunitaChulani, “Softwaredevelopment
Applications Written in Java,” Journal of Information and Software cost estimation approaches -A survey” Annals of Software
Technology, Volume 46, Issue 9, pp 589-60, 2004. Engineering, pp 177-205, 2000.
[22] Jeffery R, RuheM,Wieczorek I, “Using Public Domain Metrics to [46] James Nelson, H., Monarchi, D.E.: „Ensuring the quality of conceptual
Estimate Software Development Effort,” In Proceedings of the 7th representations‟, Softw. Qual. J., 1997, 15, (2), pp. 213–233.
International Symposium on Software Metrics, IEEE Computer [47] Khoshgoftaar, T.M., Allen, E.b., Naik, A., Jones, W.D., Hudepohl,
Society, Washington, DC, pp 16–27, 2001. J.P.:„Using classification trees for software quality models: lessons
[23] Heiat A, “Comparison of Artificial Neural Network and Regression learned‟,Int. J. Softw. Eng. Knowl. Eng., 1999, 9, (2), pp. 217–231
Models for Estimating Software Development Effort,” Journal of [48] Kitchenham, B.A.: „Cross versus within-company cost estimation
Information and Software Technology, Volume 44, Issue 15, pp 911- studies:a systematic review‟, IEEE Trans. Softw. Eng., 2007, 33, (5),
922, 2002. pp. 316–329
[24] K. Srinivasan and D. Fisher, "Machine learning approaches to [49] Hannay, J.E., Sjøberg, D.I.K., Dyba, T.: „A systematic review of theory
estimating software development effort," IEEE Transactions on use in software engineering experiments‟, Softw. – Pract. Exper.,
Software Engineering, vol. 21, pp. 126-137, 1995. 2007,33, (2), pp. 87–107
[25] A. R. Venkatachalam, "Software Cost Estimation Using Artificial [50] Jack E. Matson, Bruce E. Barrett, and Joseph M. Mellichamp,
Neural Networks," Presented at 1993 International Joint Conference “Software Development Cost Estimation Using Function Points” IEEE
on Neural Networks, Nagoya, Japan, 1993. TRANSACTIONS ON SOFTWARE ENGINEERING, VOL. 20, NO.
[26] G. H. Subramanian, P. C. Pendharkar, and M. Wallace, "An Empirical 4, APRIL 1994
Study of the Effect of Complexity, Platform, and Program Type on [51] CHRIS F. KEMERER “An Empirical Validation of Software Cost
Software Development Effort of Business Applications, "Empirical Estimation Models”
Software Engineering, vol. 11, pp. 541-553, 2006. [52] Putnam. L.H. General empirical solution to the macro software sizing
[27] R. W. Selby and A. A. Porter, "Learning from examples: generation estimating problem. IEEE Trans. Soffw. Eng. SE 4, 4 (July 1978),
and evaluation of decision trees for software resource analysis," IEEE 345-361.
Transactions on Software Engineering, vol. 14, pp. 1743-1757, 1988. [53] Putnam. L.. and Fitzsimmons, A. Estimating software costs.
[28] S. Kumar, B. A. Krishna, and P. S. Satsangi, "Fuzzy systems and Datamation 25, lo-12 (Sept.-Nov. 1979)
neural networks in software engineering project [54] B Boehm, C Abts, and S Chulani. "Software Development Cost
management,"Journal of Applied Intelligence, vol. 4, pp. 31-52, 1994. Estimation Approaches – A Survey”, Technical Report
[29] Huang SJ, Lin CY, Chiu NH, “Fuzzy Decision Tree Approach for USC-CSE-2000-505", University of Southern California – Center for
Embedding Risk Assessment Information into Software Cost Software Engineering, USA, (2000).
Estimation Model,” Journal of Information Science and [55] S. chulani, B. Boehm, and B. Steece, “Bayesian Analysis of Emperical
Engineering,Volume 22, Number 2, pp 297–313, 2006. Software Engineering Cost Models,‟ IEEE Trans. Software Eng.,
[30] M. van Genuchten and H. Koolen, "On the Use of Software Cost vol.25, no. 4, pp.573-583, 1999.
Models," Information & Management, vol. 21, pp. 37-44, 1991. [56] M.Pauline, P.Aruna and B.Shadaksharappa “Fuzzy-Base Approach
[31] T. K. Abdel-Hamid, "Adapting, Correcting, and Perfecting Using Enhanced Function Point to Evaluate the Performance of
softwareestimates: Amaintenance metaphor " in Computer, vol. 26, pp. Software Project”, The IUP Journal of Computer Sciences, Vol. VI,
20-29, 1993 No. 2, 2012
[32] K. Maxwell, L. Van Wassenhove, and S. Dutta, "Performance
Evaluation of General and Company Specific Models in Software
Development Effort Estimation," Management Science, vol. 45, pp.
787-803, 1999.
[33] H. Azath,and R.S.D. Wahidabanu “Efficient effort estimation system
viz. function pointsand quality assurance coverage” IETSoftw., 2012,
Vol. 6, Iss. 4, pp. 335–341 335, doi: 10.1049/iet-sen.2011.0146.
[34] Deng, J.D., Purvis, M.K., Purvis, M.A.: „Software effort
estimation: harmonizing algorithms and domain knowledge in an
integrated data mining approach‟, Inf. Sci. Discuss. Pap. Ser., 2009,
2009, (5), pp. 1 –13
[35] Idri, A., Khoshgoftaar, T.M., Abran, A.: „Can neural networks be
easilyinterpreted in software cost estimation?‟. 2002 World
Congress on Computational Intelligence, Honolulu, Huwaii,
12–17May 2002, pp. 1–
[36] Mittal, H., Bhatia, P.: „A comparative study of conventional effort
estimation and fuzzy effort estimation based on triangular fuzzy
numbers‟, Int. J. Comput. Sci. Secur., 2002, 1, (4), pp. 36–47.
[37] Benton, A., Bradly, M.: „The International Function Point User
Group(IFPUG)‟, in „Function point counting practices manual
–release 4.1‟(SA, 1999).
[38] Aljahdali, S., Sheta,A.F.: „Software effort estimation by tuning
COOCMOmodel parameters using differential evolution‟. Int. Conf.
on Computer Systems and Applications (AICCSA), 16–19 May 2010,
pp.1-6.
[39] Boehm, B.W.: „Software engineering economics‟ (Prentice
Hall,Englewood Cliffs, NJ, 1981).
[40] Peischl, B., Nica, M., Zanker, M., Schmid, W.: „Recommending effort
estimation methods for software project management‟. Proc.

49