How to Revise for GCSE Computer Science — A Complete Guide

GCSE Computer Science catches many students off guard. It looks like a practical subject — and it is — but the actual exam is entirely written. You cannot run code. You cannot look anything up. You need to write algorithms, trace programs and explain computing concepts from memory, in silence, against the clock.

This guide tells you exactly what to revise, when to revise it, and how to answer the questions that come up every year.

What is actually on the GCSE Computer Science exam?

Every major exam board — AQA (8525), OCR (J277) and Cambridge (0478) — splits the course across two written papers:

Paper 1 / Component 1 — Computer Systems (theory)

This covers the hardware and software side of computing:

CPU architecture — the fetch-decode-execute cycle, registers (PC, MAR, MDR, CIR), performance factors (clock speed, cores, cache)
Memory and storage — RAM vs ROM, types of storage (magnetic, optical, solid state), their characteristics and trade-offs
Networks — LAN and WAN, network topologies (star, mesh, bus), protocols (HTTP, TCP/IP, DNS, FTP), network security (malware, phishing, encryption)
Systems software — the role of the operating system, translators (compilers, interpreters, assemblers)
Ethics and law — the Computer Misuse Act, Data Protection Act/GDPR, ethical impacts of technology

Paper 2 / Component 2 — Computational Thinking, Algorithms and Programming

This is the practical side:

Algorithms — decomposition, abstraction, flowcharts, pseudocode
Searching — linear search, binary search (know when to use each and be able to trace)
Sorting — bubble sort, insertion sort, and for some boards, merge sort
Programming in Python — variables, data types, loops, selection, functions, lists, file handling
Boolean logic — AND, OR, NOT truth tables, logic diagrams

The single most important thing to know

Paper 2 cannot be revised by reading. You must practise writing code.

If you spend your revision reading Python rather than writing it, you will struggle in the exam. The questions ask you to produce working programs from scratch — with no autocomplete, no error messages and no chance to run and test.

Write code every day. Even if it is ten minutes of practice problems. That muscle memory matters.

A 12-week revision plan

Weeks 12–9 (Foundation)

Read through your notes on every Paper 1 topic once
Make flashcards for key definitions: CPU components and registers, network protocols, types of malware, data storage types
Start writing Python from scratch — variables, input/output, if/elif/else, for and while loops
Complete at least three past paper questions per week

Weeks 8–6 (Building depth)

Revisit weak Paper 1 topics — focus on anything you struggled to explain
Practise binary/hex conversion without a calculator (calculators are not permitted)
Trace algorithms on paper: bubble sort, linear search, binary search — write out every step
Write Python functions that take parameters and return values
Try the algorithm visualiser to see sorting and searching step-by-step

Weeks 5–3 (Practice-focused)

Work through complete past papers under timed conditions
Mark your own answers against the mark scheme — identify missing phrases
Focus on the verbs: “state” needs one word, “describe” needs features, “explain” needs a reason, “evaluate” needs advantages and disadvantages plus a conclusion
Fill in logic gate truth tables by hand until it takes under 60 seconds

Weeks 2–1 (Consolidation)

Cheat sheets only — one side of A4 per topic
Timed 20-minute bursts: write an algorithm in Python, then trace through it manually
Sleep, eat well, and trust your preparation

The topics that appear every year

These come up so reliably that you should know them perfectly:

1. The fetch-decode-execute cycle Every year. Usually 3–4 marks. Learn the register transfers in order: PC → MAR, memory[MAR] → MDR, MDR → CIR, PC incremented.

2. Binary to denary, denary to binary, binary to hex Every year. No calculator. Practice converting numbers like 45, 127, 200 in both directions.

3. Bubble sort trace Every year. Show every comparison in every pass, including passes where no swap occurs. One mark per correct comparison.

4. Logic gate truth tables Every year. Know AND, OR, NOT, NAND, NOR, XOR. The trick with NAND and NOR: they are the inverse of AND and OR — just flip the output column.

5. Compression: lossy vs lossless Every year. State the difference, give an example of each, and know when to use which.

6. Network security threats and defences Every year. Match the threat (malware, phishing, brute force, SQL injection) to the appropriate defence. Never answer “use antivirus” for phishing.

7. Write a Python program to… Every year. Common tasks: find the largest/smallest in a list, count items matching a condition, validate input with a loop, read from a file.

How to answer common question types

”State” / “Give” / “Name” — 1 mark each

One fact. One sentence maximum. Do not explain. “RAM is volatile” is a full answer to “state one characteristic of RAM”.

”Describe” — 2–4 marks

Features and steps. Use bullet points or numbered steps. “RAM stores the currently running programs. It is volatile, meaning data is lost when power is removed."

"Explain” — typically 4 marks

Fact + reason for every 2 marks. Use the word “because” or “therefore”. “Binary search is faster than linear search for large sorted lists because each comparison halves the remaining search space, meaning fewer comparisons are needed."

"Compare” — 4–6 marks

Similarities AND differences. Both things. Use “whereas”, “in contrast”, “both”. Never compare only one direction.

”Evaluate” — 6+ marks

Advantages AND disadvantages, then a conclusion. You will lose marks if you forget the conclusion — it is worth 2 marks on its own.

The most common mistakes

1. Not reading the question carefully enough “Describe” and “explain” are different. “State two advantages” means exactly two — not three. Re-read every question before answering.

2. Vague answers “It is faster” scores 0. “It is faster because each comparison eliminates half the remaining search space” scores 2. Always give a reason.

3. Confusing RAM and ROM RAM: volatile, read/write, stores currently running programs. ROM: non-volatile, read-only (or very slow to write), stores firmware/BIOS. This comes up every year.

4. Forgetting the binary point File size calculations: “pixels × colour depth gives the size in bits” — then divide by 8 for bytes, by 1024 for kilobytes, by 1024 again for megabytes. Students forget the division steps.

5. Tracing only swaps in bubble sort The mark scheme awards marks for every comparison, not just swaps. Write out every comparison on every pass.

Tools and resources on this site

Question bank — hundreds of exam-style questions with mark schemes, all free
Revision notes — structured notes per spec point for AQA, OCR and Cambridge
Flashcards — spaced repetition for key terms and definitions
Binary & hex converter — practise number conversion with instant feedback
Logic gate simulator — build circuits and check truth tables
Algorithm visualiser — see bubble sort, binary search and more step-by-step
Python Playground — write and run Python in the browser

One final thought

GCSE Computer Science rewards students who can explain, not just students who can do. A student who can write a perfect bubble sort in Python but cannot explain what it does in plain English will score fewer marks than one who can do both.

Revise the how. Revise the why. And then practise writing it all out on paper, from memory, without running the code.

That is what the exam asks for.