International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),

ISSN 0976 - 6375(Online), Volume 5, Issue 7, July (2014), pp. 67-72 IAEME
67











SOFTWARE REQUIREMENT COLLECTION ENHANCEMENT USING
SAMPLING TECHNIQUE AND APPLYING T-DISTRIBUTION


Swarnalatha K S
1
, G N Srinivasan
2
, Rakesh R
3
, Vipin Dwivedi
4


RV college of Engineering,
Mysore Road, R V Vidyanikethan Post, Bangalore, Karnataka 560059





ABSTRACT

In software engineering, effective development phase can only takes place if the requirements
gathering from client side are clearly defined and should be approachable. Clients requirements play
a crucial role for completion of the successful software projects in all of the development all the
software methodologies. If requirements from the client are not properly defined it can lead to failure
of the project. Now a days some information technology organization are not aware about the proper
way to collect the requirement from the clients and fulfill the requirements of the clients. Effective
Software development can only takes place if the requirements are properly defined and modeled.
Most of the Clients have difficulty in explaining what they need exactly, and the problems increases
when software developers failed to convert the clients requirements into working code. The role of
the business analyst to effectively communicate with clients to understand the requirements .Many of
project fail due to inappropriate information gathering process.

Index Terms: Requirement Engineering, SDLC, Gathering, Sampling Theory, t-distribution.

1. INTRODUCTION

Software Engineering is the very important and its a integrated part of the any software or
hardware industry from past several years. Software industry have become significant efforts based
activity from the last few years. Around 99.2% of software development companies which are very
small i.e. less than 100 employees in a software company are working towards significant products,
for those software firms which needs efficient software requirement engineering practices that are
suitable for their defined size and any type of business based the client requirement. Developing and
implementing any efficient software is a difficult and extremely software developer intensive activity
[1]. Developing and deploying the software is error prone because of many software developer
INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING &
TECHNOLOGY (IJCET)



ISSN 0976 6367(Print)
ISSN 0976 6375(Online)
Volume 5, Issue 7, July (2014), pp. 67-72
IAEME: www.iaeme.com/IJCET.asp
Journal Impact Factor (2014): 8.5328 (Calculated by GISI)
www.jifactor.com

IJCET
I A E M E
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 7, July (2014), pp. 67-72 IAEME
68

intensive activities. Day by day software based devices are controlling functions that are very crucial
to human survival. The chances of failures of these software based devices have increased rapidly.

Problem Definition
Effectively requirement gathering from the clients requirements is an important and first step
of any software projects and one of the most challenging requirement project management skills. It
very important to build the software project on well-defined platform and understandable client
requirements to avoid cost burden, unsatisfied client, or even cancellation of the project. Clients
typically find difficulty in explaining requirement to the business analyst or project manager[2]. The
main problems increase when software developers fail to translate requirements collected from
clients into working software program. Main focus of study is on the analysis of clients
requirements gathering or collection practices using sampling theorem.

Background
Various studies highlights the requirement collection is vitally important to build the software
project on well-formed platform and verifiable client's requirements to avoid cost burden, unsatisfied
clients, or even cancellation of the project. Clients typically have difficulty in explaining what they
need exactly and the problems rapidly increases when software developers fail to translate
requirements from the clients into working software programs. American Standish Group published
a survey report called chaos report on software requirement collection. The failure rate of all type of
software companies are presented below:

The success rate was only 24% in large software companies, 37.2% in medium scale software
companies and 48% in small scale software companies.
69.5 % of large software company projects were very challenging, compared to 52.7% in
medium software companies and 60.4% in small software companies.
39.5 % projects were cancelled in large software companies, compares to 45.1% in medium
software companies and 31.6% in small scale software companies

2. MATHEMATICAL MODEL

This sampling theorem is used in estimating population. We are using the same theorem in
estimating software requirements [3, 4]. We are estimating the mean for a particular requirement for
the entire population i.e., the mean of responses of a huge number of people. Instead of collecting
requirements from a very large number of people, we have to just collect the requirements from few
people and this collected requirements is a sample. We have to collect few more samples in the
similar method. Samples have to be verified before it is taken into consideration (Samples cannot be
biased).To employ t-distribution the samples under consideration have to be in normal distribution or
binomial distribution [6]. Here we are employing binomial distribution. The questions are of type
yes or no (most preferred since we are using binomial distribution), if not yes or no has to be
assumed internally by the requirement collector[7].
For the development of a food application software for a mobile, a survey was conducted by
us out of which one of the question was Do you like to Cook? This was the question which was
for analysis of the applications download [5]. The answers to this question was yes or no.
The mean for the entire population estimated through taking samples will lie within a
particular range (range will be calculated) and the mean for the entire population can never go
beyond this range (having a confidence level 99% for two sided which means one of the range
calculated for 100 entries lies outside the means value.).Let us divide the input collected into
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 7, July (2014), pp. 67-72 IAEME
69

different samples of size 10[9]. The following table shows the response of people saying yes to the
question.

Table 1.0
Sl.No 1 2 3 4 5 6 7
No. of People saying yes 6 8 7 8 9 7 6
Probability of saying yes 0.6 0.8 0.7 0.8 0.9 0.7 0.6
Probability of saying no 0.4 0.2 0.3 0.2 0.1 0.3 0.4

From binomial distribution
The mean can be calculated by using the formula

The Standard Deviation can be calculated using the formula



n = number (here the number of people in the sample)
p = probability (here the probability of saying yes)
q = 1-p (probability of saying no)
For the first sample

(Mean)

= 0.6 x 10 = 6 and (Std. Deviation) = 10x0.6x0.4 = 1.549

Similarly

and for all the samples can be calculated and the values are shown below.

Table 1.1
Sl.No 1 2 3 4 5 6 7

6 8 7 8 9 7 6
1.549 1.265 1.449 1.265 0.949 1.449 1.549

After this we can apply the t-distribution knowing the sample mean and standard deviation of
the sample the range in which the mean of yes over the entire population lies can be estimated. The
curve obtained will be of bell shaped with population mean in the center [10]. In order to do this first
we need to decide on the confidence level let us take confidence level as 99%. The curve will look
similar to as shown below.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976
ISSN 0976 - 6375(Online), Volume 5, Issue 7, July (2014), pp.

From t-tables value of t

for 99% confidence level for two sided curve is 3.250 (here degrees
of freedom for calculating t is n-1 = 9)
The formula for range can be calculated using


Where, = Sample Mean, S = Standard deviation sample
t = 3.250 (from the table), n = sample size (here its 10)
Calculating for 1
st
value,


Similarly for calculating range for all values and tabulating

Sl.No
1
2
3
4
5
6
7

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976
6375(Online), Volume 5, Issue 7, July (2014), pp. 67-72 IAEME
70

for 99% confidence level for two sided curve is 3.250 (here degrees
1 = 9)
The formula for range can be calculated using

= Standard deviation sample
= sample size (here its 10)

Similarly for calculating range for all values and tabulating
Range
6 1.549 (4.408,7.592)
8 1.265 (6.70,9.30)
7 1.449 (5.511,8.489)
8 1.265 (6.70,9.30)
9 0.949 (8.025,9.975)
7 1.449 (5.511,8.489)
6 1.549 (4.408,7.592)
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
for 99% confidence level for two sided curve is 3.250 (here degrees
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 7, July (2014), pp. 67-72 IAEME
71

These values can be plotted and a curve as in Fig. 1.0 will be obtained, the population mean
lies in the center and ranges will be as shown in the diagram, since we have used a confidence level
of 99% only one range will lie outside the mean value for 100 samples.
To get a good approximation for the range, we can take average of lower order values and
higher order values in the range column in the table 1.3 [11]; after calculating the average, good
approximation for the range for population mean can be obtained the range obtained is
(5.895,8.678). So, the population mean lies between this range and these ranges can be averaged to
get the population mean having confidence level of 99% and the value is 7.332 this is true for the
entire population.

CONCLUSION

From the above method, instead of collecting requirements from lakhs of people (using
questionnaire) we can simply collect the requirements from few thousands of people and know the
entire populations effect from small amount of input taken and hence can save lots of time
compared to the normal method of questionnaire.

REFERENCES

[1] Damian, D., The study of requirements engineering in global software development: as
challenging as important, Proceedings of International Workshop on Global Software
Development ICSE 2012, Florida, USA, 2012.
[2] Zowghi, D., Does Global Software Development Need a Different Requirements Engineering
process?, Proceedings of International Workshop on Global Software Development ICSE
2012, Orlando, Florida, USA, 2012, 53-55.
[3] Yin, R. K, Case study research: design and methods, Sage, 2013.
[4] Erikladnicki, R., Audy, J., and Evaristo, R., Distributed Software Development: Toward an
understanding of the relationship between project team, users and customers, to be
presented in ICEIS 2013, Angers, France, 2013.
[5] G. Ropohl, Philosophy of Socio-technical Systems, In Society for Philosophy and
Technology, 2012, pp. 59-71.
[6] Pandey, U. Suman, A. K. Ramani, Social-Organizational Participation difficulties in
Requirement Engineering Process- A Study, National Conference on Emerging Trends in
Software Engineering and Information Technology, Gwalior Engineering College, Gwalior,
2013.
[7] N. Juristo, A. M. Moreno & A. Silva, Is the European Industry Moving Toward Solving
Requirements Engineering Problems? IEEE Software, 2010, pp. 70-77.
[8] S. Komi-Sirvio & M. Tihinen, Great Challenges and Opportunities of Distributed Software
Development - An Industrial Survey, Fifteenth International Conference on Software
Engineering and Knowledge Engineering, SEKE 2003, San Francisco, 1-3 July 2013.
[9] J. Siddiqi, Requirement Engineering: The Emerging Wisdom, IEEE Software, 2013,
pp.15- 19.
[10] G. Kotonya & I. Sommerville, Requirements Engineering: Process and Techniques. John
Wiley & Sons, 2010.
[11] J. D. Sailor, System Engineering: An Introduction. IEEE System and Software Requirements
Engineering, IEEE Software Computer Society Press Tutorial. IEEE Software Society Press,
2011.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 7, July (2014), pp. 67-72 IAEME
72

[12] Pandey, U. Suman, A. K. Ramani, Design and Development of Requirements Specification
Documents for Making Quality Software Products National Conference on ICIS, D.P. Vipra
College, Bilaspur, 2009.
[13] R. H. Thayer & W. W. Royce, Software Systems Engineering, IEEE System and Software
Requirements Engineering. IEEE Software Computer Society Press Tutorial. IEEE Software
Society Press. Los Alamos, California, 1990.
[14] W. W. Royce, Managing the Development of Large Software Systems Proceedings of
IEEE Wescon, Reprinted in Proceedings 9
th
International Conference Software Engineering
IEEE Computer Society Press, Los Alamitos. California. USA, 1987, pp. 328-338.
[15] M. Dorfman, System and Software Requirements Engineering IEEE System and Software
Requirements Engineering, IEEE Software Computer Society Press Tutorial. IEEE Software
Society Press. Los Alamos, California, 1990.
[16] S. W. Ambler, Process Patterns. Building Large-Scale Systems Using Object Technology.
Cambridge University Press, 1998.
[17] P. Parviainen, H. Hulkko, J. Kaariainen, J. Takalo & M. Tihinen, Requirements
Engineering, Inventory of Technologies, VTT Publications, Espoo, 2003.
[18] Leffingwell & D. Widrig, Managing Software Requirements - A Unified Approach, Addison-
Wesley, 2003.
[19] D. Nelsen, System Engineering and Requirement Allocation, IEEE System and Software
Requirements Engineering, IEEE Software Computer Society Press Tutorial, IEEE Software
Society Press Los Alamos, California, 1990.
[20] R. S. Pressman, Software Engineering, A Practitioners Approach. Third Edition, McGraw-
Hill Inc, 1992.
[21] S. Blanchard & W. J. Fabrycky, Systems Engineering and Analysis. Prentice-Hall, 1981.
[22] S. Lauesen, Software Requirements: Styles and Techniques, Addison-Wesley, 2002.
[23] H. Berlack, Software configuration management. John Wiley & Sons, 1992.
[24] O. Gotel, Contribution Structures for Requirements Traceability. Ph.D. Thesis, Imperial
College of Science, Technology and Medicine, University of London, 1995.
[25] M.E.C Hull, K. Jackson & A. J. J Dick, Requirements Engineering. Springer-Verlag. Berlin,
2002.
[26] Hooks & K. Farry, Customer-Centred Products, Amacom, 2001.
[27] M. Lang & J. Duggan, A Tool to Support Collaborative Software Requirements
Management, Requirements Engineering Journal 6(3), 2001, pp. 161172.
[28] S.Saira Thabasum, Need for Design Patterns and Frameworks for Quality Software
Development, International Journal of Computer Engineering & Technology (IJCET),
Volume 3, Issue 1, 2012, pp. 54 - 58, ISSN Print: 0976 6367, ISSN Online: 0976 6375.
[29] Dhirendra Pandey, U. Suman and A. K. Ramani,, Performance Measurement of Different
Requirements Engineering Process Models: A Case Study, International Journal of
Computer Engineering & Technology (IJCET), Volume 1, Issue 2, 2010, pp. 1 - 15,
ISSN Print: 0976 6367, ISSN Online: 0976 6375.