Skip to content
Tutorials Center
My learning

Topics

Lesson 3 of 50 ยท What Software Engineering Is

Software Processes and the Life Cycle

Learn the activities every software process contains, the difference between a process and a methodology, and how work flows from idea to maintenance.

Every piece of software that gets built and used went through the same basic activities, whether or not anyone wrote them down: someone worked out what was needed, someone decided how to build it, someone built it, someone checked it, it was put in front of users, and it was fixed and improved afterwards. A software process is a named, repeatable way of organizing those activities. Teams that have no explicit process still have one; it is just accidental.

This lesson names the activities that every process shares and separates three words that are often mixed up: process, model and methodology. The following chapter then compares specific process models. The code examples are plain JavaScript stand-ins that model work flowing through phases; they are teaching models, not real project-management tools.

The six core activities

Different processes arrange these differently, but you will find all six in any serious project.

Plan and understand

Requirements What must the software do?
Design How will it be structured?
1 / 3
  1. Requirements. Find out who the users are, what they need and which constraints apply. The output is an agreed description of the problem, such as user stories or a requirements document (gathering requirements).
  2. Design. Decide the architecture, the main components, the data and the interfaces. The output is diagrams and decisions the team can build from (software design).
  3. Construction. Write, review and integrate the code, usually with automated builds.
  4. Testing. Verify that the software does what the specification says and validate that it solves the real problem (testing levels).
  5. Deployment. Package the software, release it to an environment and make it available to users.
  6. Maintenance. Correct defects, adapt to new platforms, add features and eventually retire the system (evolution).

Supporting activities run alongside all six: project management, risk management, configuration and version control, documentation and quality assurance.

Process, model and methodology

These words overlap in everyday speech, but the distinctions are useful.

  • A process is the actual set of activities a team performs, in the order and manner it performs them.
  • A process model is an abstract description of a family of processes, such as waterfall, iterative or spiral. It says what the phases are and how they relate.
  • A methodology (or framework) is a model plus practices, roles, artifacts and advice, for example Scrum or Extreme Programming. It tells you not just the shape of the work but also who does what and which habits to follow.

So "iterative development" is a model; "Scrum" is a methodology built on an iterative model. A team's process is what they really do on Tuesday, which may differ from what the handbook says. When improving a team, always look at the real process first.

Two shapes of work

Almost all models are one of two shapes, or a mix:

  • Plan-driven: do each activity once, in order, with documents handed from phase to phase. This works when requirements are stable and the cost of mistakes is high.
  • Iterative and incremental: repeat all the activities in short cycles, each producing a working slice of the product. This works when requirements are uncertain or learning is valuable.
flowchart LR
    subgraph P["Plan-driven"]
      direction LR
      R1["Requirements"] --> D1["Design"] --> C1["Build"] --> T1["Test"] --> Z1["Deploy"]
    end
    subgraph I["Iterative"]
      direction LR
      L["Plan a small slice"] --> M["Design and build it"] --> N["Test it"] --> O["Release it"] --> L
    end

Plan-driven work moves once through the phases; iterative work loops through all of them for each slice.

Neither is "better". Real projects often mix them: a regulated medical device may use a plan-driven backbone with iterative development inside each phase.

A process as a pipeline

You can model a process as a pipeline: work items enter one end, each phase takes some time to process them and passes them on. This stand-in follows three features through the six activities, one day per step, and shows what is finished on each day. It is a simplification (a real phase has people with limited capacity, rework and delays) but it illustrates how a process shapes when value arrives.

JS Features flowing through phases

Press Run (or Ctrl + Enter) to see the console output here.

No output.

Notice that the first feature is in the deploy phase on day 4 (the fifth day, counting from day 0) and that on days 2 to 5 several phases are busy with different features at the same time. That overlap is the basic idea behind incremental delivery.

A process with rework

Real phases do not always pass. A test can fail and send work back to build. The second stand-in uses a deterministic list of test results (not random) to show how rework changes elapsed time.

JS Rework loops make a process longer

Press Run (or Ctrl + Enter) to see the console output here.

No output.

The plan said 13 days (2 + 3 + 5 + 2 + 1); the process took 21 because two attempts failed, each adding two days of fixing and two of retesting. When you see schedules fail, look for hidden rework. Finding defects earlier, in requirements and design reviews, shrinks these loops.

Why we choose a process on purpose

A conscious process gives a team:

  • Shared vocabulary, so "done" means the same thing to everyone.
  • Visibility, so managers and customers can see progress and problems.
  • Repeatability, so success does not depend on one person's memory.
  • A basis for improvement: you can only improve what you can describe.

It also has costs. Too much ceremony slows small teams; too little leaves large ones in chaos. The right amount depends on size, risk and how well the problem is understood. Choosing a process model gives a decision method.

Check your understanding

Quick check ยท 5 questions

of 5 answered You scored / 5 ยท perfect!

  1. 1. Which of these is a methodology rather than just a process model?

    A methodology adds roles, practices and artifacts to a process model. Scrum does so on top of an iterative model.
  2. 2. In which situation does a plan-driven process fit best?

    Plan-driven processes depend on stable, well understood requirements and benefit when late changes are expensive.
  3. 3. Why did the rework example finish later than planned?

    Each failed test attempt added fix time and test time, so elapsed time exceeded the plan.
  4. 4. Which activity usually costs the most over a successful product's lifetime?

    Long-lived products spend most of their total effort after release, fixing, adapting and extending.
  5. 5. What is a software 'process' as the term is used in this lesson?

    The process is the real work as performed. A model or handbook describes or prescribes it, and the two can differ.

Summary

  • Every project performs requirements, design, construction, testing, deployment and maintenance, whether or not it names them.
  • A process is what a team really does; a process model is an abstract pattern; a methodology adds roles and practices.
  • Work is either plan-driven (phases once, in order) or iterative and incremental (loops of small slices), and many projects mix both.
  • Rework loops are a hidden source of delay, so catching defects early shortens schedules.
  • The right amount of process depends on project size, risk and uncertainty.

Next: The Software Development Life Cycle.