B8 · Computational Thinking and the AI Project Cycle

Decomposition

Decomposition means breaking a large, complicated problem into smaller parts that are easier to understand, solve and check. Each part can be worked on separately, sometimes by different people, and the solutions are then put back together.

6 min Beginner Jump to the activity

What you need: Nothing. This page starts from the beginning, with no programming and no formulas.

Definition

"Organise the school's annual day." Where would you even begin? Now try: "Book the hall. Make a list of performers. Print the invitations." Suddenly it feels possible.

Decomposition means breaking a large, complicated problem into smaller parts that are easier to understand, solve and check. Each part can be worked on separately, sometimes by different people, and the solutions are then put back together.

In simple words

You cannot eat a whole roti in one bite, but you can tear it into pieces. Big problems are the same. Making a birthday party becomes: decide the guests, plan the food, decorate the room, choose the games. Each small job is easy; together they make the party.

Why it matters

Decomposition is the first step of almost every project, from building a bridge to building an AI system. It turns a frightening task into a list of manageable ones, lets teams share the work, and makes mistakes easier to find.

Key terms

Computational thinking
A way of solving problems that a person or a computer could carry out, built on decomposition, pattern recognition, abstraction and algorithms.
Decomposition
Breaking a problem into smaller, more manageable parts.
Sub-problem
One of the smaller parts a big problem is broken into.
Dependency
When one part cannot start until another is finished, such as printing invitations only after the date is fixed.
Task tree
A diagram showing a big task at the top and its smaller parts branching below.
Integration
Putting the solved parts back together into one working whole.

Explained in detail

The four pillars of computational thinking

Computational thinking is not about thinking like a computer. It is a set of problem-solving habits that people use to describe problems so clearly that a computer, or another person, could carry out the solution. It has four main parts: decomposition (breaking the problem down), pattern recognition (spotting what is similar), abstraction (ignoring details that do not matter) and algorithms (writing clear steps). This page covers the first; the next pages cover the others.

Breaking down the annual day

A task tree makes decomposition visible. At the top is "Annual day". Below it come four branches: venue, programme, guests and publicity. Each branch breaks down again. Programme becomes: choose performances, make a running order, arrange rehearsals, find a host. Rehearsals become: book practice rooms, make a timetable, inform teachers. You keep breaking parts down until each one is a single job that one person can clearly do and check.

Watch for dependencies

Some parts depend on others. Invitations cannot be printed until the date and venue are fixed. The running order cannot be finished until performances are chosen. Spotting these dependencies tells you what must be done first, and what can be done at the same time by different people. Good decomposition produces parts that are as independent as possible, so that a delay in one does not stop all the others.

Decomposition in AI and science

Big problemSome of its parts
A self-driving carSee the road, recognise objects, predict what others will do, plan a path, control steering and brakes
A voice assistantTurn sound into words, understand the request, find the answer, turn the answer into speech
A Moon missionRocket, orbiter, lander, rover, communication, landing site selection
A crop disease appTake a photo, check the quality, recognise the disease, suggest treatment, show it in the farmer's language

India's Chandrayaan-3 mission, which landed near the Moon's south pole in August 2023, was built by many teams each responsible for one part: propulsion, the lander's legs, its cameras, its software, the rover's wheels. Each part was designed and tested separately, then brought together. A failure in testing could be traced to one part instead of the whole spacecraft.

How it works

  1. Write the big problem at the top of a page.
  2. Break it into three to five main parts.
  3. Break each part again until every piece is one clear job.
  4. Mark which parts depend on others and must come first.
  5. Solve or assign each part, then put the results together and check the whole.

Try it yourself

Breaking down the annual day

A student committee has three weeks to organise the school's annual day.

Which part of making a class newspaper?

Sort each job into the part it belongs to.

  • Choose a font and layout

  • Hand out copies at assembly

  • Draw the masthead

  • Write a report on the sports day

  • Interview the school gardener

  • Photocopy 200 copies

Common mistakes

  • Starting work without breaking the problem down first.
  • Stopping too early, leaving parts that are still too big.
  • Forgetting to put the parts back together and check the whole.

Check yourself

Question 1

1. What is decomposition?

Question 2

2. Invitations can only be printed after the date is fixed. This is a…

Question 3

3. How do you know you have broken a task down enough?

Summary

  • Decomposition breaks big problems into smaller, manageable parts.
  • Spot dependencies to know what comes first and what can run together.
  • Solve the parts, then put them together and check the whole.

Finished reading?

Sources: BBC Bitesize: Introduction to computational thinking · Indian Space Research Organisation