Introduction to Software Engineering

Introduction to Software Engineering

• Software Engineering is an engineering discipline that deals with the cost-effective development of large and complex software-intensive systems
• Software development follows a systematic approach that includes design, build and test

Core Focuses of Software Engineering

• Software engineering focuses on the real-world goals, services, and constraints of the systems, as well as the precise specification of the systems' structure and behaviour
• It involves the activities required to develop assurance that the specifications and real-world goals are met
• The field addresses the evolution of these systems over time and across system families, including the methods, processes, and tools for developing software-intensive systems economically and in a timely way

Importance of Software Engineering

• It allows systems to be produced, reliably and in a trustworthy way, both economically and quickly
• Provides development methods and techniques for software systems

London Ambulance Service (LAS) example

• A computer-aided dispatching system was developed, including an automatic vehicle locating system (AVLS) and mobile data terminals (MDTs) to support automatic communication with ambulances
• The system was to replace the existing manual system.
• Features included receiving calls, dispatching ambulances based on call nature and resource ability, and monitoring response progress
• In October 1992, the computer-aided dispatch (CAD) system deployed by LAS failed
• One ambulance arrived to find the patient dead, and another answered a 'stroke' call after 11 hours
• The AVLS could not keep track of unit location and status, so incorrect messages were sent
• As a consequence, units were dispatched non-optimally and multiple units were assigned to some calls resulting in large numbers of error messages, and the system slowed because the message queue grew
• The IWSSD-8 case study, "Report of the Inquiry into the LAS" concluded the software was incomplete and untested, the implementation approach was high risk, and inappropriate assumptions were made during specification
• There was also a lack of user and client consultation during development, impacting system ownership, and the system had a poor fit with the organizational structure

Software Engineering Process

• This is the list of activities that develop a software system
• Consists of specification, design, implementation, validation, and evolution
• Specification includes identifying features/constraints, task allocation, and setting deadlines
• System design is to meet the requirements
• Validation occurs when the system does what the customer wants
• Evolution involves system modification in response to user needs

Software Development Life Cycle (SDLC)

• It's a systematic and iterative process to develop high-quality software to meets requirements
• Uses tools, phases (processes/deliverables), and duties. to minimise development risks and costs.
• Improves efficiency, reduces risks and overlapping responsibilities, which facilities communication among stakeholders and responds to changing requirements, plus solves problems early

SDLC Phases

• Phase 1: Planning defines the project problem, high-level features, feasibility, target users, and solution scope
• Phase 2: Analysis elaborates, analyzes, specifies, and validates functional and non-functional requirements of the system (Requirements Engineering)
• Phase 3: Design defines the solution architecture and components
• Phase 4: Implementation turns the design into code using business rules and application logic
• Phase 5: Testing validates the software system to ensure that it meets functional and non-functional requirements.
  • This includes unit, integration, system and user acceptance testing
• Phase 6: Maintenance and Evolution involves updating, modifying, and improving the solution after deployment to fix issues, including code improvements and addressing user interface issues

SDLC Personnel Roles

• Analysis phase involves product owners, project managers, business analysts, and CTOs
• System architects and UX/UI designers are typically active during design
• Front-end and back-end developers implement the system
• Solutions architects, QA engineers, and testers cover the testing phase
• Data administrators and DevOps engineers are part of the deployment phase
• Users, testers, and support managers get involved during maintenance

Software Development Methodologies

• Methodologies include formal, sequential, evolutionary, and informal strategies
• V model and Waterfall are examples of formal approaches
• Spiral and Iterative & Incremental models are considered evolutionary
• Scrum, XP, and Kanban are considered informal approaches

Waterfall Model

• Each activity is a separate phase and has a sequential approach that receives feedback from another phase.
• Used where requirements are well-understood, with little changes to the design process (critical, embedded and large software applications)
• Requires defining services and constraints, implementing executable code with unit testing, and testing the system with customer data before deployment, maintenance, and evolution
• Is well-planned and scheduled, estimates budgets easily, handles strict budgets, but lacks flexibility on changing requirements, and phases must be completed before the next, so is available only at the end of the process

Incremental Development

• Develop an initial implementation, get user feedback, and evolve through several versions.
• Requirements change during the development process, and is ideal for systems and software products.
• Requirements are split and incrementally delivered, with high-priority requirements in early increments, and these all act as a prototype and refined capabilities
• Risks are lower, and customer feedback incorporated for flexible debug and test, allowing customer feedback and delivery and deployment, however it is costly for planning, is costly, and needs very well defined module interfaces

Spiral Model

• Focuses on risk assessment and minimizing project risk by breaking a project into smaller segments, with each cycle passing through the same sequence of steps
• Involves identifying objectives, assessing risks, generic development for validation, and project review
• Is used when development is possible, has later features, avoids risks, good on critical project, however is costly for low risk and may be difficult due to endless analysis

Agile

• Breaks projects into phases and emphasizes collaboration and improvement and constantly cycles between planning, executing, and evaluating
• Very popular agile frameworks are Scrum, Kanban, Hybrid, Lean, Bimodal, and XP

Scrum

• Scrum breaks the end product into smaller components
• Team members inspect and adapt after each increment, with 3 team roles and 5 events and 3 artifacts
• A Sprint is a short development cycle to produce a shippable product increment, and can last 1-4 weeks
• Daily Scrum planning lasts around 15 min, and is for the day
• Sprint Retrospective reviews what happened in sprints to what can be improved
• The Sprint Backlog consists of tasks that delivers sprints
• Sprint Planning prioritizes that works gets done
• Agile requires ready, acceptance and outline of subtasks

Kanban

• Work items are represented as cards on the board to track work and maintain transparency for continuous communication

Scrum vs Kanban

• Scrum has a defined roles, sprint cycle and change policy, while Kanban has no defined roles and is continues. Kanban can be incorporated every time