Professional Documents
Culture Documents
CHAPTER2
2. Literature Survey
A literature search identified 2000+ among which filtered to a 1341 conference and
258 journal articles, a total of both 1599 articles from 2001 to 2014. Figure 2.1 shows the
number of conference papers from standard symposiums, workshops, tutorials like
International Conference on Software Engineering (ICSE), Agile and Agile Software
Development (XP), Symposium on Empirical SE and Measurement (ESEM), LESS (lean
enterprise software systems), Product-Focused SPI (Profes), EuroSPI and EuroMicro
conference on SE and Advanced Applications (SEAA), and some other papers from
different reputed forums. Table 2.1 shows different search key terms by ORed and ANDed.
It seems that waterfall to agile transition, adoption and transformation is drastically
increasing and shows good attention in research communities both from academia and
industry.
However, an outline of journal publications which is shown in figure 2.2reveals that
IEEE Software has a greater number, followed by IST, JSS and IEEE Transaction on SE.
Among non-software publication articles leading a journal is an IBM Journal of Research
and Development, Computing in Science and Engineering. Thus, waterfall and agile seems
to be not just in SE but in other interdisciplinary too.
Figure 2.3 shows drastically increased research on waterfall and agile in both
journal articles and conference papers from 2001 to 2014. Similar fashion, done literature
27
search for agile and lean software development, which shows a high-quality vigorous
research is going on the combination of agile-lean principles, agile-lean practices, agile-
lean adoption, etc. and a good number of articles and papers released into the education
market (figure 2.4).
Fig: 2.2 Total number of Journal publications in scientific Journals from 2001-2014
28
Table 1 in chapter-1 describes with six aspects of each model: model name, year,
author, pros, cons and applications. In this survey, mainly focus on waterfall, agile and lean
software development. Broadly speaking on these models endeavored to address issues
such as in waterfall model: defects all too often are detected too late in the software
development process, not supporting in terms of business value, clear costs and limitations
and difficult to maintain and upgrade due to lack of integration between software
components. In an Agile model, trusts that each project should be taken care of
contrastingly and the current methods should be customized to the best suit the task
necessities.
Fig: 2.3 Publications on Waterfall and agile from 2001 to 2014, total number, conference and journal
papers (top, middle and bottom)
Some of the issues address by lean principles are unclear requirements, delay in a process,
a system of government, inevitable process repetition etc.
Fig: 2.4 Number of Publications with a combination of Agile and Lean Software
Development
2.1 Data Extraction
Table 2.2 shows the quality assessment and data extraction template.
Table 2.2 Quality Assessment and Data Extraction template
2.2 Contribution of Seminal Authors and their Clusters for Waterfall and
Agile
For better understand a field and their relationships, experts elaborated the use Co-
citation analysis, and also, it depends on authors, papers or co-citation, rests has they share
some theoretical similarity. This study is based on author documents similarity of work
done. The first and foremost step did is extracted the data from all 1341 conference and
258 journal articles, a total of both 1599 articles.
Table 2.3 and 2.4 shows the different categories and sub-categories of articles used in
waterfall and agile. Based on the cited references, relevance of two to three units of authors
work, a final cluster analysis is prepared in which included forty-nine authors. A 49 X 49
matrix of raw similarity was then computed in which 49 authors involving and mapping
between a pair of authors based on the data extracted. Instead of author co-citation analysis,
here used the Wards method by considering 5 level of conceptual similarity between two
or three authors. Figure 2.5 shows Cluster analysis of seminal authors for waterfall and
agile methodologies using the Wards method. For example, from figure 6 author’s
conceptual similarity can be expressed as, veeresh and susan[36] both of whom written
about comparison of agile and waterfall – implies that their works are closely and highly
related.
2.3 Contribution of Seminal Authors and their Clusters for Agile and
Lean Software Development
Similar fashion of waterfall and agile methodologies, A 39 X 39 matrix of raw
similarity was then computed in which 39 authors involving and mapping between a pair
of authors based on the data extracted. Table 2.5 shows different categories of articles on
agile software development and figure 2.6 shows Cluster analysis of seminal authors for
agile and lean software development using Wards method in which similarity concepts such
as lean in education, lean software development, agile-lean practices and principles, agile-
lean transformation, agile-lean adoption, agile-lean combination, defects detection,
elimination of waste, LSD in six sigma, etc.Jan Bosch (2013) [37] suggested and layouts a
set of lean development principles rather than agile practices.
31
Wards method
Scale Factor
Fig: 2.5 Cluster analysis of seminal authors for waterfall and agile methodologies using Wards method
33
Table 2.4 Different categories with their sub-categories of articles used in waterfall and agile
Wards method
Scale factor
Fig: 2.6 Cluster analysis of seminal authors for agile and lean software development using Wards method
35
three conductors in the wake of examining the whole paper.The following studies were
considered that met atleast one for the inclusion criteria:
Studies presenting the process models that are used for monitoring the work.
Studies reporting lean principles integrating with software development.
Studies related to Poka-Yoke principles in a software development life cycle.
Studies related to Fuzzy logic performance improvement in software development.
The following studies were excluded that met at least one for the exclusion criteria were:
Papers that are not focused on the process models.
Papers presenting only ideation without any practical orientation.
Papers presenting only quality attributes like reliability, security and their related
metrics.
Papers presenting only accessibility studies.
Papers presenting fuzzy logic in other disciplines.
Papers presenting techniques on how to improve performance and usability without
measures.
Papers which are not justifying with industrial projects / another category of case
study(s).
Preliminary papers for workshops, books, and special issues.
Different sources published the same study i.e. Duplicate reports.
Papers which are not written in English language.
The references of the selected studies (just those which had been observed to be most
significant in each digital library)were followed with a specific end goal to check whether
other applicable studies could be incorporated into the inquiry. This procedure allowed
validating the starting date of proposed systematic mapping study.
The following table 2.6 shows inclusion criteria and critical study for which few data
extracted fields shown to claim the research gap. The fields are title, author and year, details
of the paper, objective, methodology and paper future directions/limitations.
37
Fig: 2.7 Software monitoring framework (Courtesy: Ho-Leung Tsoi et al. [23])
To know the progress and success of the software projects during the operation of
development, SunartWanapaisan et al. [22] proposed a timesheet approach. This approach
monitored the progress by using staffs’ timesheet and compare with the work product that
have been completed by the staff and if any deviation in the project performance then alert
the project manager, so that improved the project plan too. As its’ future work authors
suggested developing a tool that can compare the staffs’ timesheet for different products
40
and for analyze. This approach is quite different than the proposed approaches with respect
to methodology, collection of data, metrics.
M. Kuhrmann et al. [23] presented an idea in terms of a drafted model by integrating
three aspects: influencers, project characteristics and criteria to monitor work in progress
throughout the project management life cycle of all activities. But the model does not
support for different dimensions like completion of work, success factors, and prevention
mechanisms.
Fig: 2.8 Governance, Management andProcesses (Courtesy: Tjan-Hien Cheng et al. [24])
Marco Kuhrmann et al. [24] presented an artifact model for agile methods to monitor
projects and to track its progress on the basis of quality. From a given reference model, it
defines artifacts with its dependencies and this definitions serves as process interfaces to
41
coordinate among different projects in a distributed environment. Still the artifact model
need to refine and generalized as a means to local software processes.
Xiaofeng Wang et al. [25] gave a superior comprehension of lean software
development approaches and how they are connected in agile software development using
30 experience reports published in past conferences of agile software.TorgeirDingsøyr et
al. [26] analyzed publications and citations to represent how the research on agile has
progressed in the 10 years taking after the articulation of the manifesto.
Authors also used wards method and categorical approach for easy mapping the
research undergone in agile concepts. One of the research directions gives that: at agile
conferences, a developing interest is evident on recognizing approaches to combine lean
principles with the development of software. It is one of the motivated lines to propose the
idea of this thesis and integrated lean manufacturing principle (Poka-Yoke) into the
software development life cycle. Other articles in table 2.6 (inclusion criteria) related to
Poka-Yoke principles are discussed in the next chapter 3.
for the waterfall model. Tested activities are characterized as tasks that were finished, yet
went over spending budget, time, or both.
The outcomes in figure 2.9 obviously demonstrate that the agile approaches convey more
fruitful activities than the waterfall model.
In the year 2010, 35% of the associations that utilized agile; just three years after
the fact, Actuation Consulting reports that, that figure has dramatically increased, to right
around 74%! That as well as this development is sponsored by genuine advantages:
Actuation Consulting study demonstrates that 86.9% of an agile client’s credit expanded
benefits to the adoption of agile. Agile is by a wide margin the most well-known strategy,
and in a larger part of cases, it is accepted to convey more esteem to organizations.
AGILE WATERFALL
8%
18%
Successful Successful
24% 49%
Challenged Challenged
There is developing an overview prove that agile works superior than traditional. The
review demonstrated that when individuals characterize accomplishment in their own
particular terms that agile tasks had a 72% achievement rate, contrasted and 63% for
conventional and 43% for off shoring. These figures are abridged and demonstrated that
individuals' involvement with agile software development was extremely positive.
2.6Waterfall People
A few sorts of organizations will dependably be more qualified to waterfall or different
methodologies, yet information lets us know that even those organizations are looking over
the wall. Actually, an Actuation Consulting survey demonstrates that 65.8% of waterfall
clients partner agile with expanded benefits; just 13.2% believe that an unadulterated
43
waterfall methodology would prompt expanded benefits. The following figure 2.10 shows
how business needs meet the target using agility.
Productivity
Somewhat High
Much Higher
4% 13%
1% Much Lower
No Change
Figures from 2.11 to 2.14 shows some results that are published in the last few
years regarding the effectiveness of agile software development compared with traditional
approaches.
44
Quality
3% 6%
Much Lower
29% 14% Somewhat Lower
No Change
Somewhat Higher
Client Satisfaction
3% 4%
Much Lower
31% 18% Somewhat Lower
No Change
Somewhat Higher
Much Higher
47%
Much Higher
18% Somewhat Higher
32% No Change
Somewhat Lower
40% Much Lower
Figure 2.11 shows comparison with respect to productivity in which 60% is high,
comparison of quality shows in figure 2.12 in which 48% is high, from figure 2.13 shows
47%of client satisfaction is high and 32% of system development cost is lower which
shows in figure 2.14.
Following figure 2.15 indicates the challenges faced by organizations adopting
agile. The most difficult challenges are listed at the top.
Figure 2.16 portrays current agileexperience with team size. Majority of agile teams
are small, with 20 or less individuals (this is valid with different standards as well). As
anyone might expect, it creates the impression that associations are discovering it less
demanding to succeed with littler teams than with bigger teams, additionally valid for
different paradigms.
Taking after figure 2.17 shows current agile experiences with geographic
distribution. At the end of the day discovered individuals reporting accomplishment at all
levels of geographic distribution.
46
10 or Less
11 to 20
21 to 30
31 to 50
Failures
51 to 100 Success
101 to 200
201 to 500
501 Plus
Co-located
Distant Locations
The questionnaire was conducted and categorized on the basis of following agile factors:
a. Specification
b. Project Schedule
c. Team Work
d. Client Collaboration
e. Challenges/Risk Involved
Figures from 2.18 to 2.22 show, questionnaire templates for the agile factor’s
specification, project schedule, team work, client collaboration and challenges/risks
involved.
Project Schedule
Sr. No. Questions
1. Was the project divided into short (same length) cycles/iterations during which the team was focused on
individual functions (or set of functions) of the product
2. Do you try to make important project decisions rapidly within short timeframes?
3. Does your software development team rely on internalized, informal, undocumented plans (as against
formal documented plans)?
4. Do you believe in short, iterative, test-driven, and people-centric development?
5. Would you shift from lifecycle-based development to feature-driven evolutionary and iterative
development?
6. Do you measure and track progress based on working software?
7. Do you practice simple designs, processes, and approaches in your software development
methodologies?
8. Is agile project management applicable in the context of product development?
9. Can certain agile practices be utilized for projects that are still carried out in the traditional manner?
Client Collaboration
Sr. No. Questions
1. Should customers closely collaborate with the development team members?
2. In software development projects, customers are committed to the project, i.e., they are motivated, active,
and consider themselves to be responsible elements of the project.
3. In the projects, a very high priority is given to achieve customer satisfaction.
4. Customer/client should propose and evaluated (cost, value added) changes together with the team.
5. Customer/client user should participate in the development of test procedures.
6. Software development project team follows continuous attention to technical excellence and good design
for development.
7. Customer satisfaction has given a high priority through early and continuous delivery of valuable
software.
similar time zone leads to better communication, 100% suggest business and developer
team should work together and as well self-organizing teams ready to take challenges.
The main aim of the study is how research in moving drastically in different
classifications of waterfall, agile and lean and issues which are not resolve up to the
satisfaction of the customers. Challenges: After the study of different articles, it seems that
for developing large or complex software application, the challenges are effort intensive,
schedule slippage, high cost, long development time, kept up-to date, changing needs for
users, high risk of failure, user acceptance, performance, poor maintainability, ad hoc
software development, efficiency, usability, etc. and some discussion is done in [26], [43].
To design, develop the needs of customer specifications efficiently, cost effectively
and to ensuring quality need an artifact. The artifact need to contain: injected by scientific
principles with well-defined approach like engineering approach (guidelines), plan and
schedule-work, involvement of a user in defining requirements, schedule reviews, define
milestones, identify stages in development, plan extensive testing, define deliverables. For
success in large software development it is important to follow an engineering approach,
consisting of a well-defined process.
55
Poka-Yoke (PY) is a mistake proof technique used in lean manufacturing. The technique
is being explored for the entire software development life cycle. A PY in a lean
manufacturing process helps an equipment operator to avoid (which means yokeru)
mistakes (means Poka). By anticipating, adjusting and/or drawing consideration/alarm to
human errors as they happen eliminate product defects. Shigeo Shingo (figure
2.28)formalized the concept and adopted the term as part of the Toyota Production System
in the 1960s’. It was originally called as baka-yoke but which gives the meaning as "fool-
proofing" or also called as "idiot-proofing," so, the name was changed to Poka-Yoke and
56
the process of zero defects is known as “Mistake Proofing” or also known as “Fail-Safe”.
To pursue more value added activities, PY can free workers’ time and mind by accepting
control tedious endeavors or exercises that depend upon watchfulness or memory. For more
information refer [27] [28] [29].
Background:
Poka-Yoke Detection:
Figure 2.32 shows Poka-Yoke detection example in which if any nut is misplaced under
the welder then automatically buzzer sound raised.
58
Poka-Yoke Elimination:
Figure 2.33 shows Poka-Yoke elimination example in which if any one insert the chuck
into the machine in reverse order then it’s not works properly.
Table 2.8 tells different principles of mistake proofing and their objectives, when
any mistake happens then apply the procedure to avoid it. For example, when defining the
potential mistakes, list the errors which could be exchanged to the following stride in the
process and apply the tools such as brainstorming etc. When it needs to identify root causes,
investigate and analyze the root causes using 5 whys or Fishbone techniques. Detailed
procedure with guidance, risks occur and how to avoid it is shown in table 2.9.
Zero Quality Control (ZQC) is the concept of quality to manufacture / develop zero defects’
product by eliminating waste associated with defects.ZQC consists of four elementary
components: 100% Audit checks, Point of Origin, Poka-Yoke, and immediate feedback.
Shigeo defines “A method that uses sensor or other devices for catching errors that may
pass by operators or assemblers is said to be PY”.
PY effect two elements of ZDQ: PY and Point of Origin Inspections which is used to
identifying defects immediately (Proactive approach), other is PY and Informative
Inspection used for correcting the action with the help of quick feedback.
60
Control Systems (SQCS), Self-Check Systems (SeCS), and Successive Check Systems
(SuCS) are three categories under Informative inspection method. Amongst SeCS is a
higher order approach than SuCS because it avoids abnormalities immediately during the
process of software development by installing Poka-Yoke (PY, In Japanese Mistake-
Proofing) devices physically within the process.
Check
and
Feedback
Defects
Action
Cause
Check&
Action
Feedbac
k
Errors
can be a compelling method for giving more documentation and however the sign ought to
be clear and unambiguous, all together not to delude clients or customer engineers.
Programming must be intended to keep a complete accident, even notwithstanding
framework failure. Auto-save components are a decent sample. "It's not frequently that the
power gets cut, but rather when it does, our clients will unquestionably value that saved a
large portion of their work [40]". A great part of the exploration center is for ZOC, quality
control, recognizing defects. Then again, the restriction that the related exploration brings
is not making a difference PY in whole.
Adzic (2012) has discussed mind-map (impact map) creation collaboratively with
senior technical and business people for answering questions like why, who, how and what.
This visualization map gives a big impact over software products and projects [40].
The PY principles also applied in education ZbigniewPrusak (2004) and Aaron
Sprunger, ChinweikeEseonu (2014) addressed the concepts of sensitivity, variability and
robustness and listed mistake proofing principles done by the students [35]. Also,
addressed lean principles integrated in PY for effective leadership in student organizations
[41]. In 2010, Mukesh Jain described a Six-Sigma based improvement model to analyze
end-user expectations, the effectiveness of the online services provided by Microsoft like
Hotmail, Bing and MSN [34].
ArashShahin and Maryam Ghasemaghaei has proposed a framework with the
technique PY for calculating the effectiveness of quality at design and delivery stages of a
product by assuming service recovery and service PY are post and pre solutions on service
fail-safing[42].
Highlight of the current state of Microsoft online service performance in major regions,
challenges. Five steps PY based approach used in Microsoft to improve performance across
their online services.Refer appendix-B for detailed usage of PY in Microsoft online
products.
64
OBJECTIVE 1:
How to design an alternative model for mistake proofing in a small and expansive
scale development of software which strengthens for services and products with high
quality desires? To tackle the above problem, detailed the genuine proposed procedure of
High Quality in a Large Scale development of software improvement using PY principles
(HQLS-PY) which solves the following Research Questions (RQ):
RQ-1: How to influence principles of PY for inspecting each of the phases in software
development life cycle and how things can be arranged and outlined upfront to avoid
from finding issues late in the SDLC?
RQ-2: How to create a model for implementing the variabilities in the connection of
Software Product Lines furthermore meets industry principles?
66
RQ-3: How to quantify design, oversee procedures, enhance the performance of the
business, meeting the requirements of customers and to monitor achievementby using
standard metrics?
RQ-4: Which opportunities infuse to inject product monitoring at the right place to
capture the user experience, product quality and alerting at the right time?
OBJECTIVE 2:
When analyzing the activities involved in a software development of thesmall scale
companies, found out that lot of mistakes can happen at every stage and these mistakes
cannot be identified by the top level management. Also, management should aware about
the progress and mistakes happening in the software development at every stage as per
Poka-Yoke concept. The software company struggle, the following operations which cause
the overall delay of the operations. i) Daily activity evaluation and finding the goal reached,
ii) A progress bar to easily monitor the improvement level of a product, iii) Hierarchical
monitoring of the work activities and mistakes, iv) No numerical values to know about the
progress of the project, v) Human errors which come from the need to prevent human action
from a leading system to undesirable states. These attributes are important for the company
to improve software performance in terms of cost, time and energy. Based on this objective,
the following research questions are formulated for monitoring the process of software
development using Poka-Yoke concept with fuzzy theory.
RQ-5:Does the monitoring phase play a major role in a software development cycle
to improve the software quality?
RQ-6: Can the fuzzy logic integrated with the Poka-Yoke concept improve software
performance using the traditional metrics?
RQ-7: Does the proposed FPYM model applied to the different companies have good
correlation in the improvement?
RQ-8: Choosing four metrics that are considered in the research work, what is the
performance that is forecasted for future products in terms of said four metrics?
67
The following are the list of chapters which solves the above research question: