Introducing Refined Agile Model (RAM) in the Context of Bangladesh'S Software Development Environment Concentrating on the Improvement of Requirement Engineering Process

International Journal of Software Engineering & Applications (IJSEA), Vol.10, No.4, July 2019
DOI: 10.5121/ijsea.2019.10402 9
INTRODUCING REFINED AGILE MODEL (RAM) IN
THE CONTEXT OF BANGLADESH'S SOFTWARE
DEVELOPMENT ENVIRONMENT CONCENTRATING
ON THE IMPROVEMENT OF REQUIREMENT
ENGINEERING PROCESS
Nirjhor Anjum1
and Anwarul Kabir2
1
Chief Analyst Officer, REVE Systems, Dhaka, Bangladesh.
2
Associate Professor, Department of Computer Science, American International
University Bangladesh.
ABSTRACT
The Software Companies of Bangladesh are using different types of agile models for software development.
Although theoretically these models are worthy for small and medium projects, in practical case they are
not so effective. In doing so, this paper tries to find out why do the agile models not suitable for
Bangladesh’s Software Companies and how do the problems that the Software Companies face for using
the models can be solved. To reveal the answers, this study is based on survey and interview methods.
Findings of this paper show that Bangladesh's Software Companies are facing different problems for
implementing traditional agile models, such as, Communicational gap, lack of Documentation,
unavailability of Prototype, Customer’s lack of knowledge in the area of IT and many more. The study
shows that if the Requirement Engineering Process is perfectly managed and some rules are modified in
the traditional agile models, these problems can be solved. In doing so, a new model has been proposed by
the study named Refined Agile Model (RAM) which is claimed to be better for Bangladesh rather than the
traditional Agile Models. This model proposes a process flow which consists of Prototyping Cycle,
Development Iteration Cycle and Additional Development Iteration Cycle. This new model also ensures a
Requirement Engineer at Client End, sufficient documentation, preparation of prototype and presentation
of frequent Demos. After ensuring these requirements in several real time projects, it was found that those
projects were completed more effectively compared to all other old project experiences. Eventually, the
paper concludes by mentioning that the Refined Agile Model (RAM) is the best model in the Bangladeshi
software environment.
KEYWORDS
Agile methodology, Requirement engineering process, Software development life cycle.
1. INTRODUCTION
Requirement Engineering is the basic part of any kind of software development. Requirement
Engineering involves requirement elicitation, requirement analysis, and requirement
management. Inaccurate collection of customer expectation, wrong definition of requirement,
improper analysis of requirement, and mismanagement of requirements etc. may cause
implementation of an inefficient system. Sometimes it may cause the total failure of a software
project.

10
The Requirement Volatility is such a problem that may occur in different phases of a Software
Development Life Cycle due to poor Requirement Engineering. In any pre-development phase if
customer expectation is collected inaccurately by the Business Analyst, it must bring requirement
volatility in the development phase. In the Project Analysis and Planning phase if the
requirements are defined wrongly by the Technical Analyst, it will bring requirement volatility as
well. If the System Architects do wrong analysis of the requirements it will also bring
requirement volatility in the project. Again, if the Project Manager fails to manage the
requirements in the development phase, it causes requirement volatility in both the development
and post-development phases too.
One of the main causes of requirement volatility is less communication between software
developer companies and client end’s main stakeholders who are the decision makers for their
projects. Though there are different well-known Software Development Methodologies like
Spiral, Prototype, RAD and Agile etc, these cannot ensure flaws, especially in Bangladeshi
software environment none of these models is found to be efficient due to constraints.
In this paperwe will be observing the present status of software development practices in
Bangladesh, especially we are going to look into the communication gaps between developers
and the client end users of the project which have severe impact on the entire project.
Section 1 of this paper focuses on the objectives for introducing Refined Agile Model (RAM) in
the context of Bangladesh's software development environment. Besides this section explains
elaborately about the background of the study. Furthermore, section 1 discuses methods that have
been used to do this empirical study. In the same vein, section 2 of the paper sheds lights on the
findings. On the other hand, Section 3 compares different methodologies that are used in
Software Development Life Cycle practices and also tries to focus on the problems of SDLC
practices in Bangladesh. Moreover, Section 4 of this paper introduces a new model named
Refined Agile Model (RAM) with the help of requirement engineering process. This section also
explains the efficacy of the new model to solve the problems of SDLC Practice in Bangladesh.
Eventually, Section 5 of the paper highlights some research initiatives that can researchers will
have been taken in this research arena in future.
1.1. OBJECTIVES OF THE RESEARCH
This paper aims at-
• To find the weaknesses of the models used in Bangladesh software development
environment.
• To explore the software process model used by the Software Company in Bangladesh.
• To investigate the problems that the Software Companies are facing in Bangladesh,
especially in requirement engineering while working for any domestic or offshore
software project.
• Based on the problems, especially in connection to the Requirement Engineering process,
this paper will try to introduce a process model that will suit in the context of
Bangladesh’s software development environment.
1.2. BACKGROUND STUDY
The Software Company is rather young compared to other industries in Bangladesh. Basically, in
the early 1980s software development started in Bangladesh (Moretaza, 2016). Though majority
of the Software Companies in Bangladesh were established in the late 1990s and onward, some of
them have started their business in early 80s. In 1997 the Bangladesh Association of Software and

11
Information Services was established as the national trade body for software and IT service
industry. The objective of BASIS was to promote the IT and Software Companies by providing
with different types of supports. Currently there are more than 954 registered Software
Companies with BASIS and around 2000 non-registered IT and Software Companies in
Bangladesh who are providing different software solutions (BASIS, 2016). A large part of these
Software Companies are working with foreign clients, therefore their main goal is to work with
offshore projects. On the other hand, other companies are involved in providing software solution
to the domestic clients.
In the domestic market most of the Software Companies generally provide customization and
configuration service of the software systems that are already developed software products.
However, other companies are involved with development of small to large website and software
for their local clients from the scratch. However, it may be noted that the market for software
system in Bangladesh is expanding. At present, maximum of these software projects are web
based. It has been found from analysis that maximum of these projects use PHP, other use
ASP.Net/C# platform, while rest few use JSP/J2EE platform.
It is for sure that, well defined requirement is the foundation of software project development
because it provides a big support in the project estimation, analysis and planning phases. The
success of a software project is also influenced by the quality of the requirements. Though the
primary requirements are well documented in the beginning, still changes will appear during the
software project development lifecycle. Requirement volatility is reported as one of the major
reasons behind changes that occur in projects(Tiwana & Keli, 2014). The main problem of
projects is due to increase in technical and business complexity now a days. Therefore, the
requirement volatility management ability is the better way to bring success to software projects
(Menachem, 1994).
Different studies on software projects have highlighted the growing concern posed by
requirement volatility, and have stressed the need of proper management techniques in dealing
with the situation. Various approaches have been discussed in the literature concerning
management of requirements changes. Jones presents a list of methods or techniques which can
be utilized to control the rate of requirements change which includes Joint Application Design,
Prototyping and Configuration Management Change Control Board (Jones, 2009). Wiegers
discusses about base-lining requirements with a purpose of controlling the requirements changes
(Wiegers, 1999). Sudhakar presents suggestions on the contribution of process models towards
management of requirement volatility (Sudhakar, 2005). His paper summarizes the advantages
and disadvantages of various process models in handling volatile requirements. It is mentioned by
Rajlich that adopting new practiceslike agile or iterative etc. can better manage the troubles
regarding to requirement volatility (Rajlich, 2006).
The main goal of this research paper is to propound a new Agile Development Model, especially
for Bangladesh, which can drive Requirement Engineering in a better way and can bring success
to the software projects.
1.3. RESEARCH METHODOLOGY
The paper is based on both survey and interview method along with several case studies and
analysis on real software projects to find out a solution for controlling Requirement Volatility.
At first, a structured set of questions was made where different questions were asked from each of
the phases of a software development life cycle. To get an effective feedback on this
questionnaire, 12 Business Analysts and Software Analysts were selected. Their valuable answers

12
on the questionnaire helped to identify Requirement Volatility as a major problem of Software
Development Life Cycle. But, only Survey Method was not sufficient to find out all the problems.
Then, to find the problems related to Requirement Volatility in details and the remedies that the
Analysts took, later on, Interview Method was taken. Through interview with the Analysts the
problem of Requirement Volatility was understood more clearly. Besides, the interviews gave
some ideas on how the Analysts tried to handle Requirement Volatility in their projects. But, the
found ideas were not sufficient to mitigate volatilities from the projects.
Next, as Case Study five best software projects were selected from a leading and renowned
software company of Bangladesh. These five software projects were taken from different
categories, different development model and different countries so that the paper may not have
any limitation to the type of software development or location.
All these survey, interview and case study are made on the project managers and on such projects
that faced requirement volatility even being renowned and large projects.
All these survey, interviews and case studies helped finally to find out the source, type, scope,
and reasons of Requirement Volatility. Finally all these helped this paper to find the way of
managing Requirement Engineering process effectively to bring success in software projects.
As the rate of Software Development in Bangladesh is not huge like foreign companies, this
paperanalyzed on 12 most popular software projects of different categories; however, interview
and questionnaire is taken from 25 people who are Senior Project Managers, Software Analysts,
Business Analysts and Team Leaders. It is expected that, this data will be sufficient to do research
on Bangladeshi Software Company.
2. EMPIRICAL STUDY AND PROBLEM IDENTIFICATION
Different methodologies were followed to examine and find the difficulties that Bangladeshi
Software Companies face in the software development life cycle. At first Survey with
questionnaire, next Interview from 12 Business Analysts and Software Analysts, and finally Case
Study on five best selected software projects were taken in this regard. This chapter explains
about the finding of problems that Bangladeshi Software Companies face while following
traditional development models.
Survey Findings
First of all, we will be going through the Surveys and will find out the statistics to understand the
findings from this survey:
Figure 1: IT Consciousness of Customer

13
Figure 1 says that, maximum of the customers are unconscious or somewhat conscious about
what they are expecting. Therefore, when customer is explaining their requirement, it is for sure
that, any kind of wrong requirement can be received from customer. From this study we find that,
a Business or Technical Analyst should be concerned about how IT concern his customer is
actually, otherwise any mistake may occur while collecting requirements.
Figure 2: Customer’s Understanding on own Requirement
From figure 2 we can understand that, 60% of the customers are somewhat aware and 20% of the
customers are not aware of what they are expecting, while only rest 20% is aware of what they
are actually expecting. From this question it is found that, to avoid requirement volatility we
should know about what customer is expecting actually, because due to lack of technical
knowledge customer sometimes can’t understand what they are asking and whether it is feasible
or not what they are asking.
Figure 3: Documentation of Requirements from Customer
From figure 3 we can find that, 60% customers come with no documentation, while only 20%
comes with brief and rest 20% comes with rough documentation. It rarely happens that customer
comes with detailed documentation. This study makes us understand that, to protect a project
from requirement volatility collecting customer requirement through documentation is very
necessary. This might not cover all the requirements accurately but at least gives a concept on
how customer is defining his requirements.

14
Figure 4: Making Customer clear about the Deliverables in the Requirement Analysis Phase
From figure 4 it is very clear that, in 60% of the cases it wasn’t possible to make the customer
visualize about what they are going to receive, and in 40% of the cases all the Developer
Companies somewhat visualized the deliverables to the customers. Therefore, very clearly we
can understand that, customer can’t be sure about his entire requirement in the beginning, when
gradually customer watches some parts of the development and understands technical things, then
gradually he identifies and defines his further requirements more accurately. So, if the Analysts
make a clear view to their customers about what they are going to deliver ultimately, and to do so
if they make some prototypes or wireframes then easily requirement volatility will be avoidable.
Figure 5: Feasibility of Customer Requirements
Figure 5 makes clear that, in 80% cases customer explained requirements are almost feasible,
while in 20% cases few requirements are feasible while maximum are not. Therefore, it may
happen in some project cases that many requirements provided by customer might not be feasible.
If Analysts do mistake in evaluating feasibilities of customer requirements in the analysis phase,
the requirement volatility may occur in later phases of the SDLC. This means, proper feasibility
testing of customer requirement is highly important to protest requirement volatility.

15
Figure 6: Understandability and Analyzability of Customer’s Requirement by Analysts
According to figure 6, while 40% of the analysis and understanding of customer requirement by
analysts are perfect and 20% is overall, at the other hand, 40% of the analysis is poor. If
requirement is analyzed and understood poorly it must bring volatility in requirements. Therefore,
wrong analysis of requirement by the Analysts may bring requirement volatility in the
development phase, because when requirement is wrongly understood then requirement must
change in later phases once the wrongness is understood.
Figure 7: Frequency of communication in the Requirement Analysis Phase
The statistics of figure 7 shows that in 80% cases customer rarely communicates in the
development phase, while 20% communication is somewhat frequent. This study makes a clear
understanding that, due to rare communication in the requirement analysis phase it was not
possible to collect all the requirements properly from customer. Therefore, proper communication
is required between customer and analysts in the requirement analysis phase.

16
Figure 8: Arrival of New Requirements in the Development Phase
According to figure 8 it is found that, in 60% cases somewhat while in 40% cases huge
requirement arrived in the development phase. From it we learn that, care should be taken to
manage the requirements by the Project Managers. If a Project Manager fails in requirement
management it may cause requirement volatility. Furthermore, incomplete collection of
requirement in the analysis phase may cause volatility in the development phase.
Figure 9: Number of Newly Accepted Requirements in the Development Phase
Figure 9 shows that, in 80% cases we accept almost all new requirements, while in 20% cases we
accept all the requirements that customer provides while development phase is already running.
From it we learn a lesson that, we should not accept any new requirement while a development
phase is running if we really want to protest requirement volatility.
Figure 10: Number of Change Requirements that are accepted in Development Phase
Figure 10 makes us clear that, not only new requirements but we also accept huge change
requests in 40% cases, and somewhat change requests in rest 60% cases. If the requirements are
understood, analyzed, defined, documented, collected and managed properly, therefore if the
requirement analysis phase is successful then change request never appears in the later phases of
a SDLC.

17
Figure 11: Delay in Delivery Timeline while New/Change Requirements are accepted
Figure 11 explains that, in 80% cases deadline is overall missed and in 20% cases it is highly
missed due to accepting new/change requests from customer in the development phase. So, to
avoid requirement volatility we always should avoid accepting new/change requests widely in the
development phase. Therefore, we can accept any number of requirements in the requirement
collection phase but not when the contracted development phase is running.
Figure 12: Frequency of communication in the Development Phase
It is found from figure 12 that, in 80% cases customer communicates the development team either
rare and somewhat while the development phase is running. It is observed that where frequency
of communication is higher, the requirement volatility appeared less therein. Therefore, to
overcome requirement volatility problem, frequent communication between clients and
development team is required.
Figure 13: Number of Newly Accepted Requirements in the Testing Phase
From figure 13 it is found that, even in the testing phase customer gave huge requirement in 40%
cases while somewhat requirements in rest 40% cases. This happens because when customer
views a practical view of the development they understands more accurately about what they are
going to get and what they actually need. Therefore, in testing phase requirement management is
highly required by the Project Managers.
Figure 14: Number of Change Requirements that are accepted in Testing Phase

18
As according to figure 14, 80% cases the development companies accept new/change
requirements from the customers in the test phase to complete the project successfully. But such
acceptance of requirements lastly delays the timeline of the entire SDLC, and the project at last
gets delayed. Therefore, if any new requirement appears in the testing phase, we can mark that
another development phase but shouldn’t include those within the running development phase.
Otherwise the project will be victimized by requirement volatility.
Figure 15: Frequency of communication in the Testing Phase
From figure 15 we find that, in 80% cases customer-developer communication is either rare or
somewhat frequent. Though testers are overall aware of the requirements and testable checklists
but still they cannot understand the actual customer requirements due to less communication
opportunity with customers in the testing phase. If customer-developer/tester communication
remains deep in the testing phase as well, this will help to mitigate requirement volatility from the
end part of the SDLC.
Interview Findings
As questionnaire method is not sufficient to understand a software project, it’s life cycle, issues
and solutions, next interview method is added to learn more about how Requirement Volatility
appears in real time software projects, and how the Project Managers (both Business and
Technical) handle these volatilities practically. For the purpose of this paperinterviews were taken
from 6 Business Analysts and 6 Software Analysts who were involved with some selected and
most popular software projects.
From a set of interview it is found that, to avoid Requirement Volatility companies involve their
Software Analysts in the Requirement Collection so that they can hear and understand the
requirements from customer. Generally this phase is handled by Business Analysts, but to avoid
collection of wrong/incomplete requirement companies involve their Software Analysts as well.
It is also found from the interviews that, if even the Agile Method is followed, to reduce
Requirement Volatility proper and detailed documentation is maintained through the development
life cycle by some companies.
Some interviewee have informed that, to avoid Requirement Volatility customer was kept tagged
throughout the development process deeply and multiple demo is presented so that there might
have no new requirement at the end of the project.
3. A COMPARISON STUDY OF DIFFERENT METHODOLOGIES IN
BANGLADESH
There are different Software Development Life Cycles, among which the best known ones are:
• Waterfall Model
• Iterative Model

19
• Spiral Model
• V-Model
• Big Bang Model
• Agile Model
• Scrum
• XP
• RAD Model
• Software Prototype Model
Although the above mentioned models are used to find out the problems regarding the
information and technology sector of the world, Scrum, Agile Model (Agile XP and RAD) are
practiced in Bangladesh’s industries which are related to information and technology (IT) sector.
Due to this reason, this models are studied to identify the problems of the information and
technology based industries of Bangladesh.
Problem 1
The findings reveal that as Bangladesh’s software industries are growing rapidly from the last 10-
15 years, customers are not much concerned about IT products and services. In doing so, in many
cases customer expects things that are technically not feasible. As a result, the proper
requirements collection process becomes difficult for the Software/Business Analysts.
On the other hand, Agile Model advantageous to that customer who is overall technology
oriented. So, this is so much contradictory to the current scenario of Bangladesh.
Problem 2
From the paper findings it is very clearly found that the customers are very indifferent to study,
understand, analyze and plan the requirements of their projects through proper documentations in
Bangladesh. As a result, Software Companies compel to give less emphasis on the
documentation. And, for this lack of proper documentations, if any team member of the project
leave the project within development phase that makes an impediment to the new member who is
replaced instead of member who leaves the project to understand the scope of the project
properly. In this case, it is very clear that the Agile Model is not appropriate for Bangladesh
because of its preference on the less documentation.
Problem 3
As Agile Model requires an overall plan of the project from customer, and it is not feasible in
Bangladesh because here the maximum customers are not IT oriented and they cannot visualize
what they are expecting for their business. Later when the development of their project starts,
sitting with the vendor companies they begin understanding what they actually need. In doing so,
customer cannot give all requirements in the beginning, and his requirements increases
throughout the development phase. As a result, Software Companies cannot visualize the final
outcome of the project and unable to stay stick to a perfect project plan.
Problem 4
Agile and RAD methodologies demand an uninterrupted communication between customers and
Software Companies during the development phase of the project. But in the case of
Bangladesh’s Software Companies, after giving the work order, customers are indifferent to
communicate with the companies. When the deadline comes, they start to knock the developers
for delivering the project.

20
Problem 5
Agile Model requires complete dedication of the software developers to the project. But findings
shows that in Bangladesh, the practical scenario is different. Here, it is impossible for any
developer to work dedicatedly for one or at two project. It frequently happens that each developer
is tagged with multiple software development and maintenance projects.
Table of Problem Findings on SDLC Practice in Bangladesh:
From the comparative studies of the different methodologies of Bangladesh, it is clear that
maximum agile and RAD methodologies attributes have some problems for which they are not
suitable for Bangladesh Software Companies. In doing so, the following table explains the
problems and their results and effectiveness of these Agile and RAD models in Bangladeshi
Industry:
ISSUES RESULT
Can the SDLC give Solution?
Scrum XP RAD
Customers are not much IT
concerned and can’t express
their exact requirement
accurately
• Many of customer
requirements become
unfeasible
• Wrong requirement is
collected from customer
No No No
Customer don’t provide
documentation for their
requirement
• Due to less documentation,
detailed requirements are not
collected from customers
No No No
Development Companies
rarely do detailed
documentation of the
projects
• If Team Member leaves it
becomes very difficult to let
new Members understand
about currently running
project
No No No
Customer can’t provide
their entire requirement in
Requirement Analysis phase
• More requirement arrives in
the Development and Testing
phase
• The final outcome remains
unclear and uncertain
No No No
Customers don’t keep much
communication in
Development and Testing
Phase
• Customer can’t understand
about the final outcome of the
development unless the final
phase reaches
Partially Partially No
New Requirement arrives in
the Development and even
in Testing Phase from
customer
• Timeline increases No No No

21
To keep customer satisfied,
the Software Companies
accept new requirements
even when a Development
and Testing Phase is
running
• Project delay occurs Yes Yes Yes
Developers are not fully
dedicated for one single
project, and is involved with
others
• Project delay occurs
• Project fails to reach quality
Partially Partially Partially
In this way, it becomes clear that, theoretical software development methodologies are failed to
suit with Bangladesh’s Software Environment. This outcome of the problem inspires to find a
better model which can be best for the Bangladesh’s Software Environment.
4. INTRODUCING REFINED AGILE MODEL (RAM) IN THE CONTEXT OF
BANGLADESH ENVIRONMENT CONCENTRATING ON THE IMPROVEMENT
OF REQUIREMENT ENGINEERING PROCESS
From the study on Bangladesh’s Software Companies it is found that, no traditional and
theoretical model can be suitable for the companies.
It is expected that the newly proposed Refined Agile Model (RAM) can be a fruitful model for
developing software in the Bangladesh’s software industry.
Requirement Engineering Process:
Refined Agile Model (RAM) proposes that Customer will not be responsible for giving plan,
sharing idea, describing requirements or keeping communication. The Software Company will
appoint a Requirement Manager and a Requirement Engineer who will be visiting Customer
premises, will know their business need, will understand what exactly they need and will act like
Customer with the Development Team. These Requirement Manager/Engineer are from IT
background with knowledge in Business Area. So, they will better understand the need of
Customers. In this cycle the Requirement Engineer/Manager will not only lock requirements but
also bring an approximately perfect Prototype of what customer is expecting. This will help he
Customer visualize that what he is going to receive finally.

22
Iteration Cycle Development Process:
Once the final Plan and Prototype is ready, next the entire Project Plan is parted into several
Iteration Cycles. In each iteration a phase of task will be completed.
Here as per Refined Agile Model (RAM) no new requirement will be accepted within the planned
iteration cycles. For additional requirements new iteration cycles will be defined after completion
of the actually planned iteration cycles. This will avoid increase of time and confusion within the
development.
Overall Process Flow:
In the first cycle “Prototype Locking” will be done, which will make the customer and developer
company sure about the final outcome.
And, next there will be ‘n’ number of Iteration Cycles to complete the project development which
will be defined by the Project’s Technical Lead discussing with Requirement Engineer/Manager.
No new requirement will be accepted within these cycles.
Finally, to meet customer desire, with the new requirements (that customer asked even after
development of the final Plan and Prototype) another `n` number of Iteration Cycles will be
cleated and continued.

23
Figure: Overall Process Flow
Therefore, the followings are the salient features of the proposed model:
1) In the Requirement Analysis Phase a Requirement Engineer from Software Company
will be sent to Customer-end (under the supervision of a Requirement Manager) who will
be giving sufficient time at the Customer Premises to understand customer’s need. Point
to be noted here that, this Requirement Engineer and Requirement Manager will act on
behalf of the Client always throughout the project development and will deal
requirements on behalf of Client. Their role will be understanding Client’s need and
acting like Client with the Software Developing Company.
2) After collection of the requirements by Requirement Engineer, these requirements will be
listed down and locked by the Requirement Manager who will be vetting and confirming
these requirements from Customer.
3) Then, the Requirement Manager will make a Prototype with the help of an UI Engineer
or Documentation Expert. This will help the customer to visualize about what he/she is
ultimately going to get in the delivery. Thus, the Prototyped Model is practiced in this
phase.
4) Next, the Project Manager will make an almost complete plan of the entire development.
As per the proposed rule, the maximum possible plans should be done in the beginning,
and the Unseen and Undefined Requirements will be added later on in different Phases.
5) Project Manager will make different Managed Iteration Cycle of Development to break-
down the entire development into different Phases. The Requirement Manager will help
to set a priority sequence of Phases and Tasks.
6) In each Managed Iteration the Cycle-Plan, Development, Testing, Feedback (UAT),
Fixing, and Deployment of each Phase will be done. In each Phase’s UAT phase the
Requirement Manager will be locking UAT feedbacks sitting with customer, if required
in the Customer’s premises.
7) If one Managed Iteration Cycle has any missing or additional requirement, those are
developed and managed after the Contracted Development completes. Therefore, the
Cycles will be added totally separately. Therefore there will be no addition of
requirement within an ongoing cycle. Thus the entire process is Incremental as well.
Some other recommendations for the proposed model:
1) Understanding the IT knowledge depth of customer is very important, because this will
help to evaluate, deal and collect requirement from customer accordingly.
2) Practically it is very tough to gather the entire requirement in the analysis phase, and
some requirements will remain uncollected, therefore new requirements will appear while
development is ongoing or even later for sure. But still, Analysts will try to collect the
maximum possible requirements in the Requirement Analysis phase.

24
3) Technical Analysts should also be tagged while collecting requirement from customer if
the Requirement Engineer is a non IT person. This will help to get vetting on the
feasibilities of what customer is asking for.
4) Making customer clear about the feasibility of his expectations is necessary, because this
helps to avoid future impractical expectation of customer in the project. Therefore, it is
very important for the Analysts to identify the feasibilities of client requirements
accurately.
5) Proper documentation is required to collect requirement, prepare final deliverable list,
generate Software Requirement Specification (SRS), make QA and UAT checklist etc. so
that in each phase can be specified and locked through proper documentation.
6) Presenting prototype, wireframe and demo of the final product is necessary to let
customer have a clear view on what they are going to get once the development is
completed. This will help the Analysts help to avoid wrong analysis of customer
requirements as well.
7) Frequent communication from the beginning phase to deployment phase is highly
required to avoid any type of Requirement Volatility. This frequent communication will
be done by the Requirement Manager. If number of demo presentation is increased
through the entire development phase, then the risk of wrong development decreases;
because each demo presentation phase gets vetting from customer.
8) The Project Managers have to be bold and logical to reject any new/change requirement
within a running development phase if it is not feasible accept those within the running
one.
9) If any new requirement appears in the development or testing phase, it should be tagged
with the next development phase, but not with the running one.
10) QA Testers should be tagged with the project in the Requirement Analysis phase to
prepare accurate checklist for testing. QA Tester should also be tagged with customer
while they are providing feedback in the testing phase so that UAT of customer and
feedback of QA Tester matches perfectly.
4.1 Validation of the Methodology
To validate that “Refined Agile Model” works better than theoretical agile models and other
models, it was implemented on two projects named:
• Training Management System, Implemented for Sudokkho NGO
• MicroERP System for E-Commerce Business, Implemented for Proshiddho Limited
RAM worked wonderfully and a table is presented below that explains how RAM solved the
problems that theoretical agile couldn’t solve:
PROBLEM SOLUTION BY RAM
Customer can’t provide their
entire requirement in
A Requirement Engineer (who is from Engineering background
with concepts of Business) visited client’s premises 5 times and
understood their business need. Finally he created the
Requirement List on behalf of the customer and explained it to the
Development Lead
Customer don’t provide
documentation for their
requirement
The Requirement Engineer and his senior, the Requirement
Manager prepared the Requirement List and made proper Project
Documentation on behalf of the customer.

25
Customer are not much IT
concern and can’t express their
exact requirement accurately
throughout the Development
phases
Throughout the entire Development Phase the Requirement
Engineer was dealing with the Development Lead as Customer on
behalf of the Customer. The Requirement Engineer checked the
development frequently and gave his feedback when he found
mismatch with Customer’s requirement.
New Requirement arrives in the
Development Phase
As a Prototype of the entire project was prepared in the beginning
and was acceptance and signed by the Customer, therefore there
were be no scope of requirement volatility.
Customer’s new/change requirements were moved to make new
iteration cycle after completion of the planned iteration cycles.
Customers don’t keep much
communication in Development
and Testing Phase
As there were Requirement Manager and Engineer who acted as
Customer to the Development Lead/Team there were no need of
Customer to sit with the Development Company.
However, this happened because Requirement Manager/Engineer
kept frequent communication with Customer to synchronize them
about the Development.
To keep customer satisfied, the
Software Companies accept new
requirements even when a
Development and Testing Phase
is running
At first all possible requirements were collected, then these were
divided into some Development Phases (Iteration Cycle). The
overall plan was done in the beginning and prototype was made as
per that plan.
As the Prototype was signed, and Requirements were locked
properly, customer didn’t pressurize to add much new/change
requirements.
Developers are not fully
dedicated for one single project,
and gets involved suddenly with
other projects, therefore there
remains scope of change in
project team.
As Requirement Engineer and Manager documented the projects
properly and listed all Requirements perfectly, it helps new team
member to understand about requirement and from where what to
do.
So, finally, the above results were evaluated and the theoretical agile models and RAM was
compared, and it is found that RAM performs better. Here goes the comparison chart among
theoretical Agile Models and Refined Agile Model (RAM):
PROBLEM
Can the Model give solution?
Scrum XP RAD
RA
M
Customer can’t provide their entire requirement in
No No No Yes
Customer don’t provide documentation for their
requirement
No No No Yes
Customer are not much IT concern and can’t
express their exact requirement accurately
throughout the Development phases
No No No Yes

26
New Requirement arrives in the Development
Phase
No No No Yes
Customers don’t keep much communication in
Development and Testing Phase
Partially Partially No Yes
To keep customer satisfied, the Software
Companies accept new requirements even when a
Development and Testing Phase is running
Yes Yes Yes Yes
Developers are not fully dedicated for one single
project, and gets involved suddenly with other
projects, therefore there remains scope of change in
project team.
Partially Partially Partially Yes
So, this way, implementation of Refined Agile Model (RAM) in the two selected projects it is
found that this newly proposed model truly works better in the Bangladesh’s Software
Development Environment.
5. FUTURE WORK
In this paper different type, price and volume of software projects were analyzed to understand
different SDLC and their implementations and obstacles in the Bangladesh’s Software Company.
All these experiments have found that, the requirement engineering process is not being proper
and requirement volatility in huge and uncontrollable in many cases.
However, all the observations of this paper were made in the Bangladeshi Software Company
regarding to their SDLC practices. Therefore, this paper is limited within the Bangladeshi
Software Company. This same study can be made on the foreign companies and the projects as
well. In this way this study will have a global and wider solution for the problems occurring in
the requirement engineering process.
Besides, the proposed Methodology will be best suitable for Medium to Large scale Software
Development, but is not good for small application development. So, further study can be done to
suit the Development Methodology for all type of developments.
As a further scope of study, this same research can be done for the developments that were done
on the open source CMS and other regular type of web based projects.
6. CONCLUSION
This paper tries to find out that why does the Agile Model not suitable for Bangladesh’s Software
Companies and how do the problems that the Software Companies face for using the
methodology can be solved. Methodologically, this paper collected data through survey and
interview. The findings of this paper reveal that agile development model is good for small and
medium projects and in favor of the incremental model of delivery, many Software Companies in
Bangladesh think it is suitable for local software projects. But the working environment of
software development of Bangladesh does not very compatible for using Agile Model.
In the case of Agile Model, it is required to keep frequent communication with client from the
beginning of the project up to its end. But, in practical it becomes difficult for Software

27
Companies to sit and communicate very frequently with their customers in Bangladesh. This
ultimately leads to a communication gap between clients and Software Company. For such
communication gap neither Software Companies get scope to understand client’s need properly
nor clients understand that whether the Software Companies are understood his expectation
accurately or not.
Another big problem with traditional Agile Model in Bangladesh is proper documentation. In
Agile Model detailed documentation is not required. But in the case of Bangladesh, maximum
clients change their requirements randomly. To manage a project understanding and managing
the change-requests properly is highly required. And, due to lack of documentation in Agile
Model project managers sometimes fail to understand and manage the changing behavior of the
projects. Besides, as in maximum of the Bangladeshi Software Companies resources are limited,
therefore in case of large projects the Companies shifts their developers suddenly from one
project to another based on their need. But due to lack of proper documentation, newly joined
developers cannot understand the project suddenly which also hampers the development of the
project.
Moreover, in Agile Model there is no distinction between small or big changes. As every change
has a cost and in Bangladesh as random change requests appear from client-end within a
development life cycle, therefore the traditional Agile Model can’t fully identify the effort against
the large changes. This ultimately increases the cost of project which may bring loss for the
project.
All these studies mainly highlighted the problems of various models, especially Agile, that
increase Requirement Volatility. But, there are some other minor SDLC practices which are not
studied and analyzed. Moreover, these studies are not done on the Bangladeshi software projects.
For this reason, the paperhas given priority in finding the remedy of Requirement Volatility and
thus finding a way to manage the Requirement Engineering process effectively in the context of
Bangladeshi Software projects.
In this paper, Requirement Volatility is found as a major problem that appears in the traditional
software development life cycles (SDLCs) that are practiced in Bangladesh.
To ensure the solutions of the problems a better solution of new Software Development
Methodology is proposed which is named as prototyped Agile Model.
Refiled Agile Model prepares a Prototype that defines an almost complete Project Plan in the
beginning. Next, the entire project plan is placed in different iterations and are delivered time to
time as per customer requirement. In each iteration the requirement collection, development,
testing and deployment will be incremental but will be within the scope of the designed
Prototype. Thus the entire model becomes a combination of Incremental Iterative (Agile) and
Prototype Model.
In conclusion, by using the new model named “Refined Agile Model” practically, positive
outcome was found. So, it is expected that overcoming all the bindings of traditional and
theoretical models this model will be able to server the purposes of managing Software
Development Life Cycles in Bangladeshi Software Company.
REFERENCES
[1] Moretaza, T. (2016) Ensuring priority for local Software Companies first challenge. The Independent.
[Online] 27th June. p.14. Available from: http://www.theindependentbd.com/post/49371. [Accessed:
03rd July 2016].

28
[2] BASIS (2016) BASIS Members Directory 2016. [Online] Available from:
http://www.basis.org.bd/index.php/members_area. [Accessed: 28th July 2016].
[3] A. Tiwana, and M. Keil, “The one minute risk assessment tool,” Communications of the ACM, 2004.
[4] M. Ben-Menachem, Software Configuration Management Guidebook, McGraw-Hill International
(UK) Limited, 1994.
[5] C. Jones, Software Engineering Best Practices: Lessons from Successful Projects in the Top
Companies, McGraw-Hill Osborne Media, 1st ed., 2009.
[6] K. Wiegers, Software Requirements, Microsoft Press. 1999.
[7] M. Sudhakar. Managing the Impact of Requirements Volatility. Master Thesis. Department of
Computing Science, Umeå University, Umeå, Sweden. 2005.
[8] V. Rajlich, “Changing the paradigm of software engineering,” Communications of the ACM, vol. 49,
no. 8, August 2006.
[9] Costello, R. and Liu, D. (1995), “Metrics for Requirements Engineering”: Journal of Systems and
Software, Vol 29 (No. 1), pp. 39-63 MIL-STD-498. 1994. Software [10]Development and
Documentation. U.S. DoD.
[11] T. JavedManzil, M. Quiser, and S. Durrani, “A study to investigate the Impact of requirements
Instability on Software Defects”, ACM SIGSOFT Software Engineering Notes, 29 (3), May 2004,
pp:1-7.
[12] Zowghi, N. Nurmuliani, A study of the Impact of requirements volatility on Software Project
Performance, Proceedings of the Ninth Asia-Pacific Software Engineering [13]Conference , APSEC
2002, Gold Cost, Queensland, Australia,04-06 Dec 2002, pp:3-11.
[14] Lamsweerde, A. Requirements Engineering in the Year 00: A research perspective. In proceeding of
the 22nd International conference on Software Engineering (ICSE’2000), Limerick, Ireland, 5-19,
ACM Press.

Introducing Refined Agile Model (RAM) in the Context of Bangladesh'S Software Development Environment Concentrating on the Improvement of Requirement Engineering Process

More Related Content

Similar to Introducing Refined Agile Model (RAM) in the Context of Bangladesh'S Software Development Environment Concentrating on the Improvement of Requirement Engineering Process (20)

More from sebastianku31 (20)

Recently uploaded (20)

Introducing Refined Agile Model (RAM) in the Context of Bangladesh'S Software Development Environment Concentrating on the Improvement of Requirement Engineering Process