Software & Software Development

• Programmers work through a number of distinct phases collectively known as the software development lifecycle
• SDLC consists of:

  • Feasibility
  • Requirements
  • Analysis and design
  • Implementation
  • Testing
  • Deployment
  • Evaluation
  • Maintenance
• SDLC : Feasibility

  • Is the problem solvable
• SDLC: Requirements

  • Working out what the solution needs to do
• SDLC: Analysis and design

  • Working out how the solutions needs to do it
• SDLC: Implementation

  • Coding the solution
• SDLC: Deployment

  • Installing it in the target environment
• SDLC: Evaluation

  • Checking in with the user - is it complete
• SDLC: Maintenance

  • Ensuring it continues to function properly by the way of improvements, patches and updates
• Feasibility establishes whether or not a project can be done

  • Deciding how feasible a new system is relies on factors such as
  • Economic
  • Time
  • Technical
  • Political
  • Legal
• A requirement is a specific feature of the new system and is recorded in a document called requirements specification
• Requirements can be broken down into:

  • Interface: how the user interacts with it
  • Function: what the system will do
  • Performance: how well/fast the system will respond
• The five software development methodologies are:

  • Waterfall lifecycle
  • Spiral model
  • Rapid Application Development
  • Agile methodology
  • Extreme programming
• Waterfall lifecycle

  • Derives name from cascading effect from one phase to another
  • Each phase has a well-defined start and end point with identifiable deliverables
• Process of the waterfall model

  • Strict development stages
  • Identify problem
  • Access feasibility
  • Analyse system
  • Design solution
  • Implementation
  • Testing
  • Go live
  • Evaluation
  • Maintenance
• Advantages of waterfall model

  • Simplicity makes it easy to manage
  • Everyone on the project is clear on their responsibilities at each stage
  • Clear deliverables
  • Easy to see if project is running to schedule
• Disadvantages of waterfall model

  • Carries a lot of risk
  • User doesn't get to see the product for the first time until the project is near its end
  • Misunderstanding requirements can lead to a project that is not easy to fix
  • Requirements must be well understood, so model is not suitable for complex projects
• Spiral model

  • Risk-driven development methodology as any project has its own risks associated with it
  • A guide for development teams, allowing to adopt element of one or more methodologies e.g. waterfall
• Process of spiral model:

  • Uses approaches from waterfall but is risk-centered
  • Recognises that risk at the heart of many large scale projects
  • Deals with risks before they become major problems
  • 4 stages: Identity, Design, Construct, Evaluate (IDCE)
  • Prototype is developed and tested
  • Project goes through a number of builds until it's complete
• Advantages of the spiral model

  • Risk management is at the heart of the model
  • Excellent for projects that contain a high level of risk
• Disadvantages of the spiral model:

  • Complex nature of risk analysis increases costs - risk management is a highly specialised skill
  • If risk analysis is done badly, the project suffers
• Rapid Application Development (RAD)

  • Involves producing successive prototypes of software until a final version is produced and approved
  • Following the initial approval of a feasible program, increasingly refined prototypes are made with reduced functionality
  • These are designed, coded, tested and evaluated with the4 end user
• RAD process

  • SImilar to spiral model but emphasis is on rapid development
  • Prototype is shown to user which can influence future development with feedback
  • Iterative process may be carried out in which another prototype is developed until eventually, the program is complete
• Advantages of RAD

  • Requirements don't need to be entirely clear from the start
  • Focus groups including the user can be used to gather requirements without the need for full formal requirements document upfront
• Disadvantages of RAD

  • Focuses on usability rather than how the product works - not suited for projects where code efficiency is very important
  • Regular contact with client must be maintained at all times
  • Scales poorly for large projects with big teams
• Agile methodology

  • Focus on the idea that requirements will shift and change during development which can only be dealt with by producing software in a iterative way
  • Products built in a series of iterations named sprints
  • Sprints are short, time boxed periods when a team has focused goals to complete a set amount of work, taking no longer than 1-4 weeks
• Extreme programming

  • Aims to produce high quality code and promote developers' quality of life by encouraging them to adopt a set of common practices that focus on the values of simplicity, communication and respect
  • Considered agile framework as it encourages regular, small, iterative software releases
• Process of agile methodologies and extreme programming

  • In the design, coding, testing and evaluating stages:
  • Similar to RAD but less time between iterations
  • User becomes part of the development team to facilitate quick feedback
  • Paired programming: Two heads are better than one. Code partner to review , advise and bounce ideas off
• Advantages of agile and extreme programming

  • Emphasises programming so quality of end code is likely to be very high
  • Core principles and processes promote respect and collaboration, leading to a very productive development team
• Disadvantages of agile and extreme programming

  • Requires a team of programmers working in close collaboration - unlikely to work well if team is widely distributed geographically
  • Client must be able to commit to having a full time representative working with the development team
  • Some of the processes involved in extreme programming such as paired programming can be quite costly
• The most common test strategies are:

  • Black box testing
  • White box testing
  • Alpha testing
  • Beta testing
• Black box testing

  • Checks whether an input produces the expected output
  • Code efficiency is not important as long as inputs return the desired outputs
  • If all inputs produce the desired outputs, we can be satisfied the program works in all cases
• White box testing:

  • Involves testing the algorithms in the code and making sure all parts of those algorithms function as intended
  • Focuses on identifying and testing all the possible paths of execution through a program
  • Path of execution is noted so it can be compared with other runs
  • Checking the efficiency of the code
• Alpha and beta testing are carried out when software is nearly ready for release and can be tested as a complete solution
• Alpha testing occurs first and is limited to internal employees and their friends and family

  • Often an early version of the finished software and may contain a number of bugs
• Beta testing is usually opened to a wide community through a public beta program

  • Program is in an almost finished state
  • Developers look to test area such as load balancing and multiple hardware compatibility
• The different test data types are

  • No data
  • Erroneous data
  • Normal data
  • Boundary data
• Normal data is data that should be accepted by the program without causing errors
• Boundary data is data of the correct type that is on either edge of the accepted validation limits
• Trace tables have the ability to track down logic errors in a program

  • A new row should be added if the state of a variable changes