Software Development Processes and Methodologies

What software dev looks like:-

Waterfall – phase to phase to phase

 Requirements –
 Design - mis
 Implementation
 Verification
 Operations and Maintenance
Limitations of Waterfall – Difficult to respond to changes, misinterpretations of
requirements or design can remain undetected until the later development phases,
integration issues may remain undetected until the last phase.

V Model
Sashimi Model
RUP Model
Spiral Model

Agile – not a model – it is a mindset

Agile manifesto and principles

Scrum
Kanban
XP - Extreme Programming Project

Continuous Integration
Automated Testing
Automated Deployment
Automation
DevOps Culture

Can Agile work for big projects:-

Scaled Agile Frameworks
Disciplined Agile Delivery
Large Scale Scrum – LeSS

Smaller cylces – faster – cheaper?

Lean Startup
Design Thinking

Output => Outcome

Process is just part of the puzzle

Importance of Requirements
Software being intangible makes it hard
What is Requirement(s) Specifications?
2 things

Process
Create high level descriptions
Distinguis between ‘right’ and ‘wrong’ system
Capture WHAT not the HOW of the solution

Product (of the process)

What are the requirements specifications important?

Engineering Argument
Economic Argument

Poor requirements are the source of all evil

Requirements problems are

The most costly
The most difficult to identify
The most difficult to correct

Requirements vs Specifications

Writing the software requirements specifications (SRS)

 Users
 Developers

Natural language used to avoid miscommunication

 User requirements
 System specification

Requirements for the user – Specifications for the developer

Write your requirements in the user language
Write your specifications in the system language
Be sure that your specifications meet the requirement

Non Functional Requirements

Define System properties and constraints

Process requirements
Often more critical than functional requirements

Security, Performance, Usability

Non-functional requirements
  System Design Constraints.
 Quality related constraints like security, performance, and usability

Product Requirements
Organization Requirements
External Requirements

Product – Requierments which specify that the delivered product must behave in a
particular way
Organizational – requirements which are a consequence of organizational policies and
procedures
External – requirements which arise from factors which are external to the system and its
development process

Non-functional requirements
 Never be overlooked
 Very important on nearly every project
 Be sure to consider them separately from the functionality
 Consider each classification separately

A constraint to only use Microsoft Project during the system development is a non-
functional requirement.- True

Which of the following are non-functional requirements:

Some product requirements, like using a specific encryption protocol, are non-functional
requirements.
Organization requirements imposed by the company, like a specific coding style, are non-
functional requirements.
External requirements imposed by external organization, like using a specific development
style, are non-functional requirements.

WRSPM – Reference Model – The World Machine Model

Capturing the ‘Right’ Thing

- Requirement are always in the problem demain

- Software specifications is in the solution (computer) domain
- Several layers of abstraction can exist in between

Interface is the overlap between Environment & System

World – Assumptions to the ‘world’ situation

Requirements – problem to be solved
Specification (natural language
Program – what software devs write
Machine – hardware spec

eh – elements of the environment hidden from the system

ev – data visible to the system
sv – system elements visible to the environment
sh - system hidden all the background detail, not visible to the user

WRSPM Model is a reference model for how we understand problems in the real world

- Helps identify the difference between requirements and specification

- Requirements are in the user (problem) domain
- Specifications are in the system (solution) domain
- One must be careful to ensure that the specification meets the requirements

WRSPM stands for: World, Requirement, Specification, Program and Machine

Looking at the difference between user requirements and system specifications in the ATM
example, we know that swiping the card and prompting for a PIN are requirements, while
reading the card details and a 4-digit PIN are specifications. – TRUE

The purpose of the WRSPM model is to ensure that: Specifications meet the requirements

WRSPM – Real World Example – the model helps identify the elements involved in solving
problems

- Must be careful of world assumptions

- Requirements are met whne…
- Specification meet requirements when…
- If the W, S imply R and M, P imply S, your solution is ‘right’

Software Architecture: Definition

Parellels

- Architects are the technical interface between customer and contractor building <the
thing>
- Bad architectural design cannot be rescued by good construction
- Specialist types of projects require architect expertise
- Schools and styles have emerged

Partitioning large systems into smaller ones that can be created sepearetly and integrated
with one another and with existing systems

Good things are well architected

Good architecture is hard

Mistakes are the architectural level are hard to fix by just coding

Software Architecture mainly about decomposing the system into components

- Each component must have individual business value

- Helps organize workforce and resources
- Allows for parallelization
- Helps define the build vs buy question and getting funding
-

- A subsystem in an architecture must:

- be created separately and can operate individually.
- have business value. 
- be integrated with one another or with existing subsystems.

Partitioning of a large system into smaller subsystems helps the buy-or-build decision because
we can examine each subsystem and reason about possible buy-or-build options for each. –
TRUE

A good software architecture is important because

- It helps organize the workforce and resources.

- It allows for parallelization in development. 
- It helps build-or-buy decisions.
- It helps with funding decisions. 

Software Architecture: Models

- Pipe and Filter

- Blackboard
- Layered
- Client server
- Event based

Pipe and Filter

Shared data between data sources which can be interspersed between data streams,
components talk to each other

Blackboard

Using a hub of data to share data or process, no interaction between components all feed to
main (blackboard) point
Layered

Triangle form, various forms top to bottom

Business
Data
Application
Tech / Infrastructure

Systems are partitioned between layers

Client Server

Cloud based computing is the big one. Computer – Cloud – AWS

Event Based

Similar to blackboard, Event manager is a central point components feed into, Event manager
has a separate connection to component

Software Architecture Models are best practice solutions to commonly encountered problems

- Start from the pure model, and adjust

- If your goals align with the goals of the model, use it
- Customize the ‘pure’ form to your needs
- Each models solves one particular large scale problem

A UNIX program where the output of one program is the input of another, is an example of which
of the software architecture models below: -- Pipe and Filter Model

An online banking system is best modeled by: Client Server Model

A vehicle identification and tracking system, where each moving vehicle is tracked and monitored
through a shared program, is best modeled by: Blackboard Model

Software Architecture: Process

System structuring
Control modelling
Modular decomposition

Subsystem vs Modules

Subsystem – independent system which holds business value

Module – component of a sub system which cannot function as a standalone system
Software Quality Attributes
- Performance
- Reliability
- Testability
- Security
- Usability

Software architecture process concerns itself with estimation, quality and partitioning

- Partitioning the responsibilities of the software is only the beginning

- Determine the costs and secure funding
- Divide work and system with quality in mind
- Apply software architectural models to common problems

The difference between subsystems and modules are: Subsystems can independently comprise
the business logic by itself while modules cant
Software architecture concerns itself with both estimation and quality but not partitioning. -
FALSE

Software Design: Introduction

What is Software Design?

A process
A product of the process

What is it not?

Architecture
Design

Stages of Design:

- Problem understanding
- Identify one or more solutions – There is More Than One Way To Do It - TMTOWTDI
- Describe solution abstractions
- Repeat process for each identified abstraction until the design is expressed in primitive
terms

Stages of design (formally)

 System Architecture
o Component Spec
 Component Interface Spec
 Component Design
o Data Structure Design
 Algorithm Design
Software Design takes abstract requirements and creates detail ready to be developed

 Decide classes, method, data types to be used in the solution

 Separate architecture from design (or don’t)
 Provide detail which is implementation ready
 But doesn’t include implementation detail/constrictions

Software design is the process of transforming the stated problem into a ready-to-use
implementation – FALSE
Abstract solutions do not require extensive domain knowledge and effectively reduce the costs
during the software design phase. – FALSE
It is often advised that abstract solutions do not provide optimization details regarding the
implementation. – TRUE
When it comes to software design, it is always best to follow a solution that is widely popular in
the industry – FALSE
While a solution coming from software design does not include implementation details, there are
still common cases where pseudocode may be provided to correctly capture the sense of a
complex algorithm. - TRUE

Software Design: Modularity

What are the aspects:

 Coupling
 Cohesion
 Information hiding
 Data encapsulation

Definition and principles of modularity

Complex systems MUST be broken down into smaller parts

3 primary goals:
 Decomposability – divide and conquer
 Composability
 Ease of understanding

Information Hiding:

Hide complexitiy in a ‘black box’

Examples: Functions, macros, classes, libraries

Example:

Void sortAscending (int *array, int length)

Don’t know which sort it uses

Don’t really need to
Know how to use it
Data encapsulation

 Encapsulates the data

 Helps find where problems are
 Makes designs more robust

Which of the four aspects of modularity is defined as: How well modules work together. –
Coupling
Which of the four aspects of modularity can be described as: Abstracting away implementation
details. – Information Hiding
Three aspects of ______ can be described as (1) Decomposability, (2) Composability, and (3)
Ease of Understanding. – Modularity
You have a sort function that provides no details on which sorting algorithm is used. This is an
example of which aspect of modularity? – Information Hiding
A benefit to using ______ is that you know if your data is corrupted, then it must have been
corrupted by the module. – Data Encapsulation
A benefit to using ______ is that you know if your data is corrupted, then it must have been
corrupted by the module. – Cohesion
Which of the four aspects of modularity can be described as: Containment of constructs and
concepts within a module. – Data Encapsulation

Software Design: Coupling

Measuring the strength of connections between (sub)-system components

Loose coupling allow for changes to be unlikely to propogate across components
Shared variables and control information lead to tight coupling
Loose coupling achieved by state decentralization and message passing

Couling Levels – Tight – Medium – Loose

Tight
Content Coupling
Common Coupling
External Coupling

Medium
Control Coupling
Data Structure Coupling

Loose
Data coupling
Message coupling
No coupling

Coupling measures communication

How well are we protected against requirements change propagation

Measuring the various types of coupling is important
Aim for data and message coupling
Justify anything tighter
Choose the most accurate answer: Low coupling aids in... – ensuring that changes don’t cross
boundaries of moduels
Which of the following statements about coupling are true? –
 Coupling measures the strength of connections between components.
 Loose coupling makes it unlikely that changes will be propogated across components.
 Shared variables and control information leads to tight coupling.
 Loose coupling is partly achieved through message passing.

The goal in low coupling is to ensure that changes don't cross the boundaries of modules. –
TRUE
he three types of tight coupling are: Control, Common, External
Module A relies directly on local data of module B. This is an example of what type of coupling? -
Tight content coupling
Modules A and B both rely on global data or a global variable. This is an example of what type of
coupling? – tight common coupling
Modules rely on externally imposed format (or protocol or interface). This is an example of what
type of coupling? – tight external coupling
The two types of medium coupling are: Control, data structure
Module A controls the logical flow of module B. This is an example of what type of coupling? –
Medium control coupling
Module A and B both rely on the same composite data structure. This is an example of what type
of coupling? – medium data structure coupling
The three types of loose coupling are: data message none
Modules only share parameters. This is an example of what type of coupling? – loose data
coupling
The loosest type of coupling; components only communicate through parameters or message
passing. This is an example of what type of coupling? – loose message coupling

Software Design: Cohesion

Measures how well a moduel components fit together

Implements a single logical entity or function
Represents a desirable design attribute
Divides into various levels of strength

Weak
Coincidental cohesion
Temporal cohesion
Procedural cohesion
Logical association

Medium
Communicational cohesion
Sequential cohesion

Strong
Object cohesion
Functional cohesion
Cohesion and Inheritance

 Technically speaking inheriting attributes from a super class weakens cohesion

 To understand the component one must examine the super class(es) as well

Cohesion is concerned with how well the components of a modules serve a single goal

 Just having high cohesion is not sufficient

 Measuring the various types of cohesion is important
 Aim for object and functional cohesion
 Justify anything less

Choose the most accurate answer: Cohesion describes... how well everying within a module fits
together
Which of the following statements about cohesion is true? -

 Cohesion measures how well everything within a module fits together.

 Describes the implementation of a single logical entity or function.
 Represents a desirable design attitude.
 High cohesion is the aim in software development.

The goal in real-life software development is to aim for object and functional cohesion and
anything less than this should be justified. – TRUE
The four types of weak cohesion are: Coincidental, temporal, procedural, logical association
Parts of modules are together in the same file. This is an example of what type of cohesion?
Coincidental cohesion
Different pieces of code are activated at the same time. This is an example of what type of
cohesion? – temporal cohesion
One piece of code runs after the other. This is an example of what type of cohesion? Procedural
cohesion
Components do similar but separate things. This is an example of what type of cohesion? –
logical association cohesion
The two types of medium cohesion are: communicational sequential
All elements of a component operate on the same input or produce the same output. This is an
example of what type of cohesion? Communicational cohesion
One component's output provides the input to another component. This is an example of what
type of cohesion? Sequential cohesion
The two types of strong cohesion are: object, functional
Each operation in a module is provided to allow object attributes to be modified or inspected.
This is an example of what type of cohesion? Object cohesion
Every part of a component is necessary for a single well-defined behavior. This is an example of
what type of cohesion? Functional cohesion
True or False: inheritance strengthens cohesion. – FALSE

Implementation

Software development implementation tips

  Program when you are alert
 Write code for people not for the computer – comments should be why its done now
how

 Write your comments, test and exception handling before you write functional code

 Break long methods into manageable pieces – google says no more than 40 lines

 Make each class represent single intuitive concept

 Methods focussed on a single task. Preferably without side effects

 Writing code for a second time? Extract into a method

 Optimize only when you’re sure it’s necessary- code profiler

Implementation skills are training well elsewhere

These tips will help you understand the difference between coder and developer
People, not computers are the primary focus
Let the computer handle optimization for computers
These tips make the developers using them more effective

Deployment

Deployment stage occurs at the end

Deployment means production deployment release to users

Occurs after testing/ QA and others have signed off on the release

Planning

Largely determined by project scope

Includes planned steps, problem areas and plans to recover

Deployment Plan

 Physical environment
 Hardware
 Documentation
 Training
 Database related activities
 3rd party software
 Software being deployed
Software deployment is the event of launching your product to users

 Primarily focused on how to deliver and how to revert on failure

 Deployment without rollback plans shouldn’t happen
 Document your rollback plans for error reduced action under fire
 Consider all aspects of the environment when writing the plan

Deployment occurs in the end stage of active development.- TRUE

Product deployment should encapsulate three main focus areas. What are they? – problem
areas, plan for recover in case of failure, planned steps

Software Deployment Rollback

The reversal of actions, completed during a deployment with the intent to revert a system
back to its previous working state

Why rollback

 Installation did not go as expected

 Problems would take longer to debug/fix than installation window
 Keep production system alive

Know your point of no return

 At some point rolling back with take longer and or cost more than pushing through
 If at all possible know this point prior to initializing the deployment
 Just prior to reaching the point verify that rollback is not the best course of action

Rollback is the primary concern of deployment

 A matter of ‘when’ not ‘if’

 Deployment without rollback plans shouldn’t happen
 Deployment without the possibility of rollback need special attention
 Be sure to determine your point of no return before you begin

Cutover Strategies

Cold backup (cold storage) – needs to be started up 24 hours delay

Warm standby – ready to be up and running in a short amount of time

Hot failover – virtual machine all machines ready but don’t receive the info

BCP – Business continuity plan

Software cutover is a trade-off between cost and preparedness

  Speed of recover is directly proportional to cost of recover
 Hot failover allows you to redirect live data with minimal
 Warm standby has services ready for initialization
 Gold backup is a replacement machine need full setup

Software Testing: Introduction

Testing Definitions

 The process of executing a program (or part of a program) with the intention of
finding errors( Myers vs Humphrey)

 The purpose of testing is to find erros

o Testing is the process of trying to discover every conceivable fault or
weakness in a work product (Myers via Kit)

 The process of searching for errors (Kaner)

What is a Test

Test Cases

Test Data Output - Correct Result? Expected Output

*Software under Test *

Oracle

Test Data and test cases

Test Data
Inputs which have been devised to test a system

Test Cases
Inputs to the system and the predicted outputs from operating the system on these inputs,
if the system performs to its specification

What is a Bug?

Failure
Error
Latent
Effective
Fault
‘dependable computing and fault tolerance: concepts and terminology – Jean Clause Laprie
1985
Programmers mistake is fault, leads to error, stays latent until it becomes effective, then
when the outcome is not as expected it causes failure

Axiom of Testing

Program testing can be used to show the presence of bugs but never their absence –
Dikkstra 1969

When to stop testing? No right answer, testing is to prove that it is not good

Software Testing is a process to find (and fix) defects in a systems implementation

 There is no way to prove a program correct through testing

 Still an essential part of the process
 Must ensure quality if input selection and correct expected output
 Understanding testing is key to a successful project

Software Testing: Definitions

Verification: the process of evaluation a system or component to determine whether the

products satisfy the conditions imposed

Checking a program against the most closely related design documents or specifications

An attempt to find errors by executing a program in a test or simulated environment

Confirm that the software performs and conforms to its specification ‘Are we building the
right thing’

Validation: The process of evaluating a system or component to determine whether it

satisfies specified requirements

Checking the programme against the published user or system requirements

An attempt to find errors by executing a program in a real environment

Confirm that the software performs to the users satisfaction. Assure that the software
system meets the users needs ‘Are we building the right thing’

Dynamic vs Static Verification

Dynamic V & V
 Concerned with exercising and observing product behaviour
 Testing in all forms

Static V&V
  Concerned with analysis of the static system representations to discover problems
 Proofs and inspections – formal testing

Verification and Validation

 Building the thing right and build the right thing

 Regardless of definition an important aspect of testing
 Dynamic and static V&V aren’t exclusive
 Verification is cheaper/easier than validation

Software Testing: Strategies

Incremental Testing – regression testing

Top down testing – stubs are for levels not created yet, single or few lines of code that
return a set value. Mock is similar but not returned value, just line works

Bottom up testing – Driver is the reverse to a stub, level above current

Back to back testing – using previous versions for testing

Test Scaffolding
Goal - setup the environment for executing test
 Driver
 Stub
 Program Unit
 Oracle

Tradeoff – Effort in testing and regression testing v effort in developing drivers/stubs

Who should Test?

Developer
 ‘This is my baby’
 Understands the system
 Will test gently
 Driven by deadlines

Tester
 ‘smashy smashy’
 Must learn the system
 Will attempt to break it
 Driven by ‘quality’

Axioms of testing
  As the number of detected defects in a piece of software increases the probability of
the existence of more undetected defects also increases
 Assign your best programmers to test
 Exhaustive testing is impossible
 You cannot test a program completely
 Even if you do find the last bug you’ll never know it
 It takes more time than you have to test less than you’d like
 You will run out of time before you run out of test cases

Strategies of tesing drive the actual act of testing units

Pure top down or bottom up doesn’t exist

Stubs are created to allow higher level units to execute
Drivers organize and execute the units under test
Programs cannot be tested completely

Software Testing: Perspectives

Blackbox Testing
 Designed without knowledge of programs internal structure
 Based on functional requirements

White box Testing

 Examines the internal design of the program
 Requires detailed knowledge of its structure

V&V Process

 Must be applied at each stage of the process

 Has 2 principal objectives
 Discover of defects in a system
 Assessment of whether or not the system is usable

Stages of testing
Unit Testing – single unit
Module Testing – collection of units
Sub system testing – collections of modules
System testing – testing complete system, made up of sub systems
Acceptance testing – testing by users, alpha / beta testing

Many software testing perspective exist

 Each has its own benefits

 Use them all to get a better picture of how testing works
 It is never on or the other, just use each in isolation
 Even all of them combines will likely not be sufficient
Software Development Models

Predictive vs Adaptive

Predicitive – understand what the requirements are from the outset – Design,
implementation, testing – product delivered

Adaptive – Has an idea but not fully – initial version built, feedback gathered, version
updated etc etc

Incremental vs Iterative

Incremental – understand what is required, however each part is built part at a time

Iterative – Idea is known, version is built, next version is built upon & enhanced and end
may be totally different to original idea

Incremental vs Iterative vs Both vs None

All models are going to have characteristics, use cases, pros & cons, where it fits adaptive vs
predictive through the normal flow:
requirements,design,implementation,testing,deployment,maintenance

Waterfall Model

+ Simplest model. Very predictive. Simple and easy to understand. Efficient

<- Not fexible for change. First release takes a long time
Use for predictable repeatable work

requirements, design, Implementation, testing, deployment, maintenance

Feedback loop between next stage and previous exists should any problem occur

Later a bug is found the more costly it becomes, up x50

#1 We know requirements very well, they wont change

#2 Team has experience building similar software
#3 Translation from requirements to product is going to be perfect

V Model
Very similar to waterfall but as a V

Concept Operations & Maintenance

Requirements Verification & Validation
Design Integration Tests
Implementation
-------------------------------------------------------------------------------------------------------------------

Left side is project definition

Right side is project test & validation
Along the bottom is project completion time

Emphasis on validation earlier in the process

Very predictive – doesnt allow feedback or changes

+ Earlier detection of potential defects/issues

<- More upfront work

Ambiguity in requirements and early validation would be useful

Sashimi Model

Phases of the lifecycle are allowed to overlap

Requirements
Design
Implementation
Testing
Deployment
Maintenance
3months 6months 9 months

A phase can start before previous phase ends

Less predictive that previous model but still fairly predictive

+ Shorten the development time. People with different skill can start without waiting. Can
do learning spike early
<- May result in rework

Use to shorten the time. For resource utilization

Incremental Models

Built in parts, multiple mini waterfalls. May overlap. Can use different model for each
increment.
Predictive with some movement towards adaptive depending on model associated
alongside

+ Deliver value early. Get feedback and make necessary changes between increments
<- May result in rework. May cost more

Use. If organisation may benefit from early delivery of part of product. If building one
increment will help define future increments

Incremental Model Variations

Each increment follows all steps of the process

Unified Process and its Variants

Inception Elaboration Construction Transition

Business modelling
Requirements
Analysis and design
Implementation
Test
Deployment
Time

Each top section are broken down into iterations and there can be various iterations per
phase

Inception
 Short phase
 What happens?
 Establish business case
 Feasibility
 Build vs Buy
 Preliminary schedule and cost
Milestone: lifecycle objective

Elaboration
 What happens
o Capture requirements
o Address known risks
o Validate the system architecture ->
o Executable architecture baseline ->
Milestone: lifecycle architecture
Construction
 Largest Phase
  What happens?
o Software is built
o Multiple iterations – each results in release
o Iterative and incremental
 Milestone: lnitial operational capability

Transition
 What happens
o Deployment
o Get feedback and refine
 System conversion and user training

It is a framework. It incorporates other methods / models

Any step may involve different kinds of work (requirements, design etc) but relative effort
and emphasis might be different

Architecture centric

Use case centric

Focus on risk mitigation

Unified process: much more adaptive than predicitive

+ Adaptive. Quality and reuse. Focus on risk mitigation increases chances of success. Flexible
to incorporate other software development models
<- Complicated. Need more resources. Too much overheard for smaller project

Uses – Bigger and riskier projects. All requirements not known early in the project. Desire to
deliver value earlier

Rational unified process – 9 disciplines. 6 best practices + tools

Enterprise unified process
Open UP – lighter version
Agile UP – Lighter version, Agile focussed

Spiral Model

Cyclic model
Four basic steps in each cycle:
 Determine objectives
 Identify risks
 Double up
 Plan for next iteration
Not every step needs to be performed

Determine Objectives
 Objectives
 Constraints
 Alternatives

Identify and Resolve Risks

 Identify Risks
 Resolve what you can

Dev and Tests

 Work done to meet objectives

Plan for next Iteration

 Review work done and commitment for next iteration

Then starts the next cycle

How to track process

 Life cycle objective – sufficient definition of a technical and management approach
 Life Cycle architecture – sufficient definition of the preferred approach and
significant risks eliminated or mitigated
 Initial Operational Capability – sufficient preparation of the software, site, users,
operators and maintainers
 Risk Drive Model

 Effort and detail driven by risk

 Process model generator. Incorporates other models

Summary
 First described in 1986 by Barry Boehm as process model generator
 Risk driven
 4 basic activities in every cycle
 Risk determines level of effort and degree of details
 Not every activity in the diagram need to be performed

More adaptive focussed

+ Adaptive. Risk focus increases chances of success. Flexible for using any model. Minimizes
waste. Options for go / no go
<- Complicated. Coss more to manage. Need stakeholder engagement
Use – Very large high risk projects
Phase Gates / Stage Gates

Requirements Design Implementation Testing Deploy

Gates or checkpoints that exist between the various phase

Gate Requirements gate Design gate Implementation gate Testing gate Deploy

At any gate if it doesn’t come up to scratch the idea gets binned

What do you check at a gate;

 Check quality of execution of previous step

 Business case still makes sense
 If yes does the action plan for the next step reasonable and sound?

Gate structure and elements

Deliverables of previous phase

Criteria: questions or metrics on which the project is judged
Results of the gate review – a decision, go yes or no. Approved action plan for the next gate.
List of deliverables and date for the next gate

Sumary:

Technique to provide opportunity to reflect on curreny progress and decide if we need to

change/continue or terminate. Organization customise it to their needs

+ Poor projects can be quickly rejected. Cost and fiscal analysis provides quantitative
information around feasibility. Provides opportunity to validate the updated business case.
<- Potential for structural organization to interfere with creativity and innovation

Applying Software Development Models

Example:

Company need to install a well known HR Management System at a big retailers HQ

Company has done the same instsall with other companys

Known problem, known solution

Waterfall model – simple

Sashimi model – perhaps, testing various parts as go along
V model – if there is any additional aspect regarding needing testing

Example 2
Very large hospital with locations all over want to automate processes
Hired a company that is expert in this area and has done similar automation but not on this
scale
Hospital management want consistency across the globe byt not sure about the nuances in
different parts
Few of the place can benefit greatly from the immediate automation

Fairly known problem

Fairly known solution except at scale
Benefit from a phased delivery
May tweak a bit based on early iterations
Want consistency

Incremental Model

Assignment Topic:

As a software development professional, you will run into all kind of projects and situations within
those projects. This assignment is designed to present you with a fictitious situation and ask you
to recommend a software development approach suited to the situation. 

Here is the situation: 

Zenith Healthcare is a new company in the market and has launched its product two years ago.
The product is loved by clients and is growing in popularity. The level of product demand was not
anticipated, and the current system architecture cannot support the rising demand. To support
the anticipated demand, the company needs to re-architect the system and provide the exact
same functionality. Thus, the requirements from the client perspective are very well known and
do not need to change. What needs to be changed in the system to support the growing demand
is also well understood. The product has 4, fairly independent components. All 4 components
need to be re-architected. Out of the 4, one of them has caused the most pain and the
organization could benefit greatly if that component could be replaced first with a new, highly
scalable architecture. The work of migrating to a new platform is a tedious job and the
deployment of a new system will involve a lot of external communication, managing customer
expectations, etc.

The technical architect and one project manager will be working from the corporate headquarters
in Germany, but most of the team who will be doing the coding for the migration will be offshore
in Belarus. The testing team will also be in Belarus.

What software development methodology would you suggest for this situation and why? 

 Step 1: Start analyzing the scenario by identifying the characteristics of this situation

and specify the logic behind the selection of characteristics. For example, you may identify
"User Needs Unknown" as a characteristic based on statement X, Y and Z in the scenario.
 Step 2: Select a model that best fits the characteristics you identified in step 1. Justify
your choice by providing by the logic behind your selection. For example, you may say that
 since the scenario has characteristics X and Y, models A and B are potential candidates.
Additionally, since the scenario has characteristic Z, model A is the best option.

Imagine that you were the lead or project manager for this project. For the selected model, take
us through a simulated / fictitious journey on how this project will be completed all the way from
defining requirements to deployment. You are free to make up characters as you feel appropriate
to fit your story. Please watch the video on" Model Selection" to get an idea. The video stays at
high level, but you can go in further details as you feel necessary. In your story, please make
sure to talk about artifact and practices followed by the team on this project.

Assume that you are the quality lead or technical lead on this project. What kind of testing would
you suggest the team to do? Be sure to justify your answer. To answer this question, first list
down the key things from the use case above that are really important. For e.g. scalability,
performance, usability, integration between components etc. After that, identify what type of
testing would you want the team to do to make sure that upgraded product is high quality and
deployed defect free. Please refer to the "Testing and Verification" section in module 2. Also,
please watch following videos to learn about various types of testing
methods: https://www.coursera.org/learn/software-processes/lecture/G30EZ/software-testing-
perspectives

Continuing your role as a quality lead or technical lead for the project. Write a few example of
test cases or a descriptive narrative for what you expect the testing team to use when testing this
product. Please refer to the "Testing and Verification" section in module 2. Feel free to make
assumptions about the functionality of the system to come up with a scenario.

Software Development Models: Agile and Lean

Why Agile?

Verification phase a lot of errors would occure in the waterfall method. Expensive to fix as
found late in the process. Errors in what delivered vs what was needed / required

Predicting requirements / needs is hard

Market shift may have changed from concept through to delivery during the time line

Necessity is the mother of invention

Crystal, SCRUM, FDD, DSDM, XP – other methodologies

Agile is 4 values, 12 principles. Defines a mindset

Whether it's scrum, kanban, xp, lean, dsdm, or another kind of agile methodology, they all rely on
the 4 Agile Value Statements.

While both the items on the right and the items on the left are valued, the items on the left are
valued more:

 Individuals and interactions over processes and tools

 Working software over comprehensive documentation
 Customer collaboration over contract negotiation
  Responding to change over following a plan

Agile Manifesto and Principles

Agile is 4 values, 12 principles

Values:

Principles
1. Our highest priority is to satisfy the customer through early and continuous devliery
of valuable software
2. Welcome changing requirements even late in development. Agile processes harness
change for the customers competitive advantage
3. Delivery working software frequently from a couple of weeks to a couple of months
with a preference to the shorter timescale
4. Business people and developers must work together daily throughout the project
5. Build projects around motivated individuals. Give them the environment and support
they need and tryst them to get the job done
6. The most efficient and effective method of conveying information to and within a
development team is face to face conversation
7. Working software is the primary measure of progress
8. Agile processes promote sustainable development. The sponsors developers and
users should be able to maintain a constant, pace indefinitely
9. Continuous attention to technical excellence and good design enhances agility – cost
of exploration
10. Simplicity the art of maximizing the amount of work not done is essential
11. The best architectures requirements and designs emerge from self organizing teams
12. At regular intervals the team reflects on how to become more effective then tunes
and adjusts its behaviour accordingly

How does it solve problems of waterfall model?

Adaptive
 Deliver working software frequently
 Welcome chang
 Technical excellence and good designs
 Continuous improvement
Detect translation issues early. Validate user needs early. Detect integration issues early
People and Interaction
 Business and developer work together
 Face to face conversations
 Self organizing teams
 Promote sustainable development
 Motivated individuals
Detect translation issues early

What new problems does it bring?

Adaptive
  Deliver working software frequently
 Welcome change
 Technical excellence and good designs
 Continuous improvement

Architecture, design, database modelling is a challenging. Lack of control, unpredictable

journey, very uncomfortable for leaders/ organizations
People and Interaction
 Business and developer work together
 Face to face conversations
 Self organizing teams
 Promote sustainable development
 Motivated individuals
Requires participation from customers throughout the development process

Practices
Principles
Values

Agile Frameworks

Agile is a mindset. Frameworks should be customised

Scrum – 1 to 4 week cycle – most popular framework

Kanban – continuous flow model
Scrumban – continuous flow in a cycle
XP – Extreme Programming
Scrum XP hybrid
Lean startup
Feature Driven Development – FDD
Crystal
Custom Hybrid
Dynamic Systems Development Methodology – DSDM

Scrum
1 to 4 week sprint
Define, design, build, test

Framework contains: Product owner, Scrum master and The team

Product owner – Inputs from executives, team, stakeholders, customer users to create what
is to be done, list of what is required, definitions etc.
Product backlog – Sprint Planning meeting, work out what order from the product backlog
things will be done.
Sprint backlog, what exactly needs to be done for each task, during each sprint.
Daily scrum meeting (stand-up) to review what was done yesterday and for today. Sprints
are 1-4 weeks, and the sprint end date and team deliverable do not change. – Finished
work.
Sprint review – full team gets together with all involved, customers, stakeholders etc to go
through what was done and if any changes are needed
Sprint retrospective – how can it be better next time, what went well, badly, how it can be
improved.
Burndown / Burnup charts – breaksdown what daily work is needed and to be finalised for
that sprint

Kanban
Set or properties and principals so long as there is a continuous flow.

Backlog – Software development pipeline – end result

Backlog is a list of work to be done

Backlog, Analyze, Develop, Rest, Release – all steps are put into columns, with Doing, Done
breakdown within each

From the backlog each task with the breakdown is moved from Doing to Done as it
progresses.

1 – Visualize your workflow

2 – Limit WIP (work in progress) to a set number so to prevent log jam in testing for
example, once work is done if there are too many items in a WIP then it should be looked as
to why, code is wrong, testing taking too long etc.
3 – Manage the flow, are there issues, is the code wrong, does it require more people.
4 – Make process policies explicit – When is something done to move to testing

Kanban Principles
 Start with what you do know
 Agree to pursue incremental evolutionary change
 Respect the current process, roles, responsibilities and titles

Kanban Properties
 Visualize the workflow
 Limit WIP – work in progress
 Manage flow
 Make process policies explicit
 Improve collaboratively

Agile and Lean Summary

Mindset not process / model

Incremental and iterative
Very short dev cycle
Very adaptive vs predictive
+ flexible to change, increases the chances of building the right product. Speed to market
<- may result in rework. Requires close collaboration with clients and users
Use – where requirements may change. Team using unproven technology

Lean Startup

Unique approach, how to learn faster about market or user need. Validated learning.

Assumption Build
Learn Experiment

Measure
Metric

Cycles should be completed as quickly as possible

‘We must learn what customers really want, not what they say they want or what we thing
they should want. We must discover whether we are on a path that will lead to growing a
sustainable business’ Eric Ries

Incremental and iterative

Very short dev cycles
Very adaptive vs predictive
+ helps you learn faster and build the right project. Speed to market
<- may result in rework, requires you to experiment iteratively with clients and users
Use – doubtful business case / user need. Lots of high probability risks

Model Selection: When to use which model:-

Example – big retailer wants to sell online, understand they need to but not sure what to
offer, need to offer something asap, ecommerce is new, company stakeholders of the
company will be driving it.

Analysis and recommendation

- Company fairly confident of the idea
- Need to iterate for the following reasons
o Speed to market – get something out quick
o Learn about what customers like and want
o New stuff for the company so need to learn about technology as well
- Business stakeholders are dedicated and hence have bandwidth to collaborate with dev
team regularly
AGILE – allow them to be flexible, tweak plans, design – work alongside stakeholders
Example
Unknown expected use
Unexpected user need

Lean Statup

Example
Existing successful company has a new disruptive idea. Like to create a new company for it

Untested new idea

Unexpected user need

Learn Startup

Assignment Topic:

As a software development professional, you will run into all kind of projects and situations within
those projects. This assignment is designed to present you with a fictitious situation and ask you
to recommend an approach to software development for that situation.

Here is the situation:

Georgia school of Arts has been the choice for higher education in the state. The college has
been serving the state for last 50 years. College has mostly operated using manual processes
and state has been expecting the college to match the other colleges in the country in terms of
automation. College leadership also sees great benefit by automation as it hopes to keep tuition
in check and provide better service to student and faculty. 

The college wants to automate most of its processes in next 5 years. College do not have budget
to shorten the timeframe and wants to do this automation mostly in-house with hired consultants
when needed. 

The IT department has lot to learn in terms technology to build this system. Also, college
leadership has some idea on what to build but not sure what exactly are college needs in terms
of automation. College has formed a group from various departments of college that will help
define and drive the development. This group will work closely with IT department on this effort.

College wants to reap the benefit of automation as it is being built and use feedback from system
users to guide future automation efforts. 

The automation will start with processes that impact the students for e.g. admissions, class
registration, grading, learning management system etc. Some of the key things identified for this
software are the privacy and security concerns. College wants to make sure that systems can't
be hacked and only the right people have access to the info. 
To start with, the team working on this project consists of 5 developers, 3 QA and a Team Lead.
Organization had signed a contract with a local company to provided additional resources when
needed.

What software development methodology would you suggest for this situation and why?

 Step 1: Start with analyzing the scenario and identifying characteristics of this situation

and specify logic behind the selection of characteristics. Example: You may identify "User
Needs Unknown" as a characteristic based on statement x, y and z in the scenario.
 Step 2: Map the characteristics to selection of model and provide your logic to make
that conclusion. For e.g. you may say that since scenario has x and y characteristic, model
A and B would be potential candidate. Additionally, since scenario has characteristic z,
model A would be best option.