I m p r o v i n g A g i l e S o f t w a r e D e v e l o p m e n t : A C a s e 
S t u d y o n A d a p t i v e S y s t e m D e v e l o p m e n t 
T h e s i s P r e s e n t e d 
B y 
G I U S P e r e r a 
S u p e r v i s e d b y 
M r . M S D F e r n a n d o 
LIBRARY 
UNIVERSITY OF MORATUWA, SRI LANKA 
MORATUWA 
This thesis was submitted to the 
Department of Computer Science and Engineering 
of the 
University of Moratuwa Sri Lanka 
In partial fulfillment of the requirements for the 
Degree of Mas te r of Science 
University of Moratuwa ^ ° 
90106 
Department of Computer Science and Engineering 
University of Moratuwa 
Sri Lanka 
January 2008 
9 0 1 0 6 
D e c l a r a t i o n 
"I certify that this dissertation does not incorporate without acknowledgement any 
material previously submitted for a degree or diploma in any university; and to the 
best of my knowledge and belief it does not contain any material previously published 
or written by another person except where due reference is made in the text" 
Mr. M.S.D. Fernando 
Supervisor, 
Senior Lecturer, 
Dept. of Computer Science and Engineering, 
University of Moratuwa. 
G.I.U.S. Perera 
it 
• 
T o 
(Parents 
and 
teachers 
* 
iii 
A b s t r a c t 
Software development is getting more and more complex with the improvement of 
technologies and tools for relative domains. Applying software solutions became an 
inevitable fact in the modern society. However, the domain conditions and the 
business environments change rapidly, by making the software norms obsolete. 
Furthermore, due to the intrinsic properties of software, situation becomes worse. 
Though the history confines to a few decades, software engineering owns dozens of 
software process paradigms to date. Introducing new process paradigms may resolve a 
selected problem, but not the others, and often introduces more issues. Moreover, it is 
evident that the improvements for the software processes only from the technical 
orientation do not solve all the issues in rapid changing environments. 
Agile software process is a well known, lightweight, and flexible practice which was 
introduced as a remedy for the above crisis. In many aspects it serves its purpose. It 
introduced a paradigm shift to the software development for rapid delivery in 
uncertain environments. However, usage of well known Agile like software paradigms 
as it is in the practical conditions is arguable. The Agility concept endures certain 
flaws due to the native characteristics. Neither, it is the best software paradigm, nor 
the panacea for all software projects. 
This research was derived with new perspectives from rational concepts in different 
domains, which were not yet been introduced to the Agile software process 
improvements. The research objectives were focused towards identifying existing 
bottlenecks in the Agile practices and potential improvements to those. Additionally, a 
case study was carried out in the mutual benefiting manner to the system's 
improvements and give flavor to the research with practical essence. The conducted 
study was successful in many aspects, and shows a significant impact to the Agile 
process. Furthermore, within the limited resource constraints the outcome of the study 
is really promising for the future study in the research area. 
iv 
4 
A c k n o w l e d g e m e n t 
It is my great pleasure to acknowledge the contribution and assistance of a large 
number of individuals who made this research a complete and success. Without their 
help and guidance I won' t be able to have a successful research experience. Though it 
is hard to appreciate everyone individually, there are some people who should be 
appreciated at least from a word. 
First of all I would like to thank my supervisor Mr. Shantha Fernando, for accepting 
me as a research student and the guidance he provided. Despite from his busy 
schedule, he was always supportive and guiding me to the correct path. He was really 
helpful to me when I had doubts about the research activities. Specially, Mr. 
Fernando's valuable effort to drive me for an affluent research is highly appreciated. 
My special thank is extended to Prof. Gihan Dias, who was my first M.Sc. research 
coordinator for the Dept. of CSE during the research, and Dr. Sanath Jayasena, the 
second M.Sc. research coordinator for my research. It was really a privilege to have 
such senior faculty members, which indeed helped me to improve my work. 
I would also thank to Ms. Vishaka Nanayakkara, the Head of the Dept. of Computer 
Science and Engineering. She was always helped me with discussing LearnOrg related 
problems and potential enhancements. I specially thank her for allocating her busy 
time for discussing and helping me with my research. 
My special thanks go to the progress committee members for this research, Mr. Raj 
Prasanna and Dr. Chandana Perera from the Dept. of Management of Technology. 
Their thoughts were really essential to add the management essence for this research. 
I would like to thank all the other lecturers and staff members of the department of 
Computer Science and Engineering for the support they have given in numerous ways. 
v 
• 
I should also thank my project members of the LearnOrg project and specially Mr. K. 
Sarveswaran, who was really helpful to me when 1 was doing my experiments. It was 
really a team experience when we work together. 
Students who participated in the experiments are worth to mention for their honest and 
guileless data, which were really, helped me to make this research trustworthy. 
I would like to thank my colleagues who studied with me for the M.Sc. at the 
Department. It was a wonderful experience to being with them. 
Finally, I would like to be grateful to my parents and my brother, for encouraging me. 
They also helped me from all the way they could, even helping me to cover up my 
daily work when I was busy with this research. 
In addition to the above, if anyone who is not mentioned there I 'm sorry for that. I 
appreciate everyone who helped me in my research. 
Perera G.I.U.S. 
Dept. of Computer Science and Engineering, 
University of Moratuwa, 
*J Sri Lanka. 
10 t h January 2008 
vi 
• 
T a b l e o f C o n t e n t 
Declaration ii 
Abstract iv 
Acknowledgement v 
Table of Content vii 
List of Figures and Tables xi 
Symbols and Abbreviations xii 
C h a p t e r 1 In t roduct ion 1 
1.1 Background 1 
1.2 Problem Statement 2 
1.2.1 Research questions 2 
1.2.2 Research hypothesis - For experiments 4 
1.3 Objectives of study 4 
1.3.1 Justification 4 
1.4 Significance of the study 5 
1.5 Scope of the study 5 
1.6 Structure of the thesis 6 
C h a p t e r 2 L i t e ra tu re review 8 
2.1 Introduction 8 
2.2 Software processes 8 
2.2.1 Software process - What is it? 8 
2.2.2 Why we need them? 9 
2.3 Software process improvements 10 
2.3.1 Problems with software processes 11 
2.4 Software lifecycle engineering 12 
2.4.1 Software lifecycle phases 12 
2.5 Agile software development 13 
2.5.1 Agile principle 14 
vii 
2.5.2 Why Agile methods are better? 15 
2.5.3. The areas to improve in Agile process 16 
2.6 LEAN principle 17 
2.6.1 LEAN principle - the framework 17 
2.6.2 LEAN software development 19 
2.6.3 Problems in LEAN production 21 
2.7 Managerial decision making 21 
2.7.1 Popular decision support methods 22 
2.7.2 Motivation for development 23 
2.8 Successful case studying 23 
2.9 Concluding remarks 24 
C h a p t e r 3 The case s tudy - System information 25 
3.1 E-learning 25 
3.2 Learning Management System - LMS 26 
3.3 Content Management System - CMS 27 
3.4 Generalized E-learning framework 28 
^ 3.4.1 The need of a generalized framework 29 
3.4.2 The architecture 28 
3.5 The LearnOrg system 31 
3.5.1 The history 31 
3.5.2 The current system 33 
3.5.3 System problems 34 
3.6 Concluding remarks 35 
C h a p t e r 4 Improvements 36 
4$y 4.1 Problem background 36 
4.1.1 Problems with Agile paradigm 36 
4.2 LEAN - Agile hybrid model 37 
4.2.1 Problems of LEAN software development 37 
4.2.2 The model 38 
viii 
4.2.3 The strengths of the model 39 
4.3 R-C Matrix 40 
4.3.1 The background 40 
4.3.2 The model 42 
4.3.3 Strengths of the model 44 
4.4 Case suggestions 45 
4.4.1 Solutions from other models 45 
4.4.2 Suggested improvements - case specific 46 
4.4.3 Rationalization of the improvements 46 
4.5 Concluding remarks 47 
C h a p t e r 5 Exper iment Methodology 48 
5.1 Experiment background 48 
5.2 Experiment 1: LEAN - Agile model 49 
5.2.1 Experiment environment 49 
5.2.2 Performance measures 50 
5.2.3 Experiment rationalization 51 
* 5.3 Experiment 2: R-C Matrix 52 
5.3.1 Experiment environment 52 
5.3.2 Performance measures 53 
5.3.3 Experiment rationalization 53 
5.4 Case study observations 54 
5.4.1 Case study environment 54 
5.4.2 Problem tracking 55 
5.5 Concluding remarks 55 
0 C h a p t e r 6 Data and Analysis 56 
6.1 Experiment 1 - LEAN-Agile hybrid process 56 
6.1.1 Collected data 56 
6.1.2 Analysis - Hypothesis testing 1 57 
6.1.3 Analysis - Hypothesis testing 2 60 
ft 
6.1.4 Analysis - Defects rate behavior 63 
6.2 Experiment 2 - R-C Matrix evaluation 64 
6.2.1 Collected data 64 
6.2.2 Analysis - Fairness of the data 66 
6.2.3 Analysis - Hypothesis testing 66 
6.3 Case study information 68 
6.3.1 Collected data 68 
6.3.2 Analysis 69 
6.4 Concluding remarks 70 
C h a p t e r 7 Discussion 71 
7.1 Study outcomes 71 
7.2 Experimental limitations 73 
7.3 Summary 75 
Conclusion 76 
References 78 
Annexes 87 
x 
• 
L i s t o f F i g u r e s 
Fig. 2.1 LEAN principles for success 18 
Fig. 3.1 The architecture of the Generalized e-learning framework 30 
Fig. 3.2 The University of Moratuwa e-learning system view 33 
Fig. 6.1 MINITAB output - ANOVA results for hypothesis testing with LOC 59 
Fig. 6.2 MINITAB output - ANOVA results for hypothesis testing with 
work levels 62 
Fig. 6.3 Average defects rates in the examined time period for the two samples 63 
Fig 6.4 MINITAB output - ANOVA results for the design phase times 66 
Fig 6.5 MINITAB output - ANOVA results for hypothesis testing using T0/T2 67 
Fig. 6.6 problem count for the LearnOrg during the study 70 
L i s t o f T a b l e s 
Table 2.1 The seven wastes of production 19 
Table 2.2 Respective seven wastes in software development 20 
Table 4.1 The R-C Matrix 43 
Table 5.1: Project parameters and their results; collecting data 50 
Table 6.1 a weekly data sample - normal agile practicing groups 56 
Table 6.2 a weekly data sample - improved agile practicing groups 57 
Table 6.3 sample A (normal agile) weighted sums of LOC 58 
Table 6.4 sample B (hybrid agile) weighted sums of LOC 58 
Table 6.5 sample A - deviation percentage from expected work amount 61 
Table 6.6 sample B - deviation percentage from expected work amount 61 
Table 6.7 Selected functional implementations without the matrix 65 
Table 6.8 Selected functional implementations with the matrix 65 
xi 
• 
A b b r e v i a t i o n s 
AMI Agility Measurement Index 
ANOVA ANalysis Of VAriance 
BCG Boston Consultation Group 
CD - ROM Compact Disk Read Only Memory 
CMM Capability Maturity Model 
CMMI Capability Maturity Model Integration 
CMS Content Management System 
COCOM02 Constructive COst MOdel 2 
CSE Computer Science & Engineering 
DEMP Distance Education Modernization Project 
DEPP Distance Education Partnership Program 
DSS Decision Support Systems 
1A-LMM IT Aligned Learning Maturity Model 
ICE Information and Communication Engineering 
ICT Information and Communication Technology 
ICTA Information and Communication Technology Agency 
IEEE Institute of Electrical and Electronic Engineers 
JIT Just In Time 
LAMP Linux, Appache, MySql, PHP 
LM3 Learning Maturity Management Model 
LMS Learning Management System 
LOC Lines Of Codes 
MOODLE Modular Object Oriented Distance Learning Environment 
NPV Net Present Value 
PEST Political, Economical, Social, Technological 
SCORM Shareable Content Object Reference Model 
SEI Software Engineering Institute 
SPICE Software Process Improvement and Capability dEtermination 
SPIN Software Process Improvement Network 
SWOT Strengths, Weaknesses, Opportunities, Threats 
VE Value Engineering