Cambridge IGCSE · Exam Tips

Computer Science (0478) Exam Tips

Master the Cambridge IGCSE Computer Science (0478) syllabus with examiner-verified advice on binary arithmetic, precise register definitions, Paper 2 algorithm construction, and scoring full marks in the 15-mark scenario-based pseudocode.

5 min readUpdated: 21 Jun 2026

Exam at a Glance

Papers
2
Total Marks
150
Time Limit
3h 30min
Question Types
4
PaperDurationMarksQuestionsWeightingQuestion Types
Paper 1 Computer Systems1h 45min752750%Short Answer, Structured/Explain, Calculations / Tables
Paper 2 Algorithms, Programming and Logic1h 45min751150%Trace tables & Diagrams, Pseudocode writing, Scenario-based coding (2D Array), SQL and Database tasks
Grade Scale
A*ABCDEFGU
Calculator Policy

A silent scientific calculator may be used on papers where calculators are permitted (some papers are non-calculator). It must not be graphical or programmable and must hold no stored information.

  • AO1: AO1: Recall, select and communicate knowledge and understanding of computer science principles and concepts (40%)
  • AO2: AO2: Apply knowledge and understanding of computer science principles and concepts to a given context, including the analysis and design of computational or programming problems (40%)
  • AO3: AO3: Provide solutions to problems by evaluating computer systems, making reasoned judgements and presenting conclusions (20%)

Built from real past papers and marking schemes (2023–2025).

Tips & Strategies

Unlocking the Code: Where the Marks Really Hide

Stepping into a Cambridge IGCSE Computer Science exam can feel like entering a maze of precise keywords and strict pseudocode rules. Candidates often lose critical marks not because they lack technical ideas, but because they express them in vague, everyday language. In Paper 1, writing that clock speed is "how fast the computer is" will instantly cost you marks. Top scorers know that the examiner is looking for precision: the number of cycles or instructions processed per second. Knowing exactly where the marks hide and how to communicate with examiner-level accuracy is the master key to achieving an A*.

The 105-Minute Countdown: Your Strategy for P1 and P2

You have exactly 105 minutes (1 hour 45 minutes) for each paper, but the way you allocate this time must differ dramatically between Paper 1 and Paper 2. In Paper 1 (Computer Systems), with 75 marks distributed over short-answer and structured explanation questions, a steady pace of 1.4 minutes per mark is your baseline. Do not linger on a tricky hardware or security description. If you are stuck on a 4-mark question about proxy servers, move on and come back.

For Paper 2 (Algorithms, Programming and Logic), time management is a high-stakes balancing act. You must reserve at least 30 to 35 minutes at the end specifically for the 15-mark scenario-based programming question (typically Q10). It is common to feel rushed on this section. A great strategy is to tackle the SQL query, database, and truth table questions first. These are highly structured, high-yield marks that can be completed efficiently, leaving you with a calm, focused block of time to write and comment your logic.

The Binary Carry Rule: No Denary Safety Nets Allowed

In Paper 1, you will encounter binary addition questions, such as adding two 8-bit registers like \( 11110101_2 + 00111001_2 \). A dangerous, recurring pitfall is converting both values to denary, performing the calculation in denary, and converting the sum back to binary. Do not do this.

Examiners look explicitly for your column-by-column binary carry markings. If you only write down the final binary string, you risk losing all working marks if you make a single transcription error. Draw a clear "carry row" above your addition, show the overflow bit if it exceeds 8 bits (generating a 9th bit, causing an overflow error), and systematically write your carries. For example, show that \( 1_2 + 1_2 = 10_2 \) (0 write down, 1 carried) and \( 1_2 + 1_2 + 1_2 = 11_2 \) (1 write down, 1 carried).

The Double-Entry Trap: Unraveling Validation and Verification

Year after year, candidates lose easy marks by confusing validation with verification. When asked to evaluate data entry security or input correctness, you must keep these two systems strictly separated:

  • Validation is an automatic computer-driven check. It uses programmed rules to check if input data matches set constraints. Examples include range checks (e.g., ensuring a score is between \( 1 \) and \( 100 \) inclusive), type checks (checking if a value is an integer or string), and length checks.
  • Verification is a method to ensure data matches its original source during entry or transmission. This is done via visual checks (manually proofreading text against paper documents) or double-entry verification (forcing a user to input a password twice, checking for an exact match). It does not check if the data itself is sensible.

If you recommend a visual check to validate that a postcode has exactly 7 characters, you will receive zero marks. Match the exact check to the correct concept.

The 15-Mark Masterpiece: Comments, Constraints, and Zero Arrays

The final question in Paper 2 is your crowning glory or your biggest pitfall. You are asked to write a substantial program in pseudocode, Python, VB, or Java. To guarantee full marks, top scorers follow three non-negotiable rules:

  1. Read the Scenario's Required Variables: If the question specifies that you must use a 2D array named NumberGenerated[], you must use that exact spelling, capitalization, and array structure. Inventing your own variable names (like my_random_array) immediately breaks the assessment criteria and drops your band score.
  2. Never Skip Comments: Under Assessment Objective 3 (AO3), up to 2 marks of the structure criteria are reserved specifically for how well your code is commented. Write brief, meaningful comments using double slashes (//) to explain your logic loops, validation checks, and output steps.
  3. Always Program the Validation Loop: If the scenario requires user input (like a player's name or ticket count), you must write a complete loop that continuously prompts the user until a valid value is entered. Simply writing an IF statement is insufficient because it only checks the input once. Use a condition-controlled loop (e.g., REPEAT ... UNTIL or WHILE ... DO) to trap invalid data.

What the Top 10% Do: Precise Terminology and Active Practice

Top scorers do not just read notes; they practice with active recall. In Computer Science, synonyms can be hazardous. Avoid using brand names like "Windows", "Excel", or "WinRAR" in your answers; doing so will invalidate your response. Instead, refer to software categories such as "operating systems", "spreadsheets", or "compression utilities".

When reviewing computer architecture, draw your own diagrams of the Fetch-Decode-Execute cycle. Ensure you can accurately trace the directional paths of the system buses, highlighting that the Data Bus and Control Bus are bidirectional, while the Address Bus is strictly unidirectional. When asked about CPU performance upgrades, always write down the exact consequence: stating "increasing cores allows more instructions to be processed concurrently/simultaneously" scores, whereas "it makes it run faster" does not. Study with precision, and the grade you want will follow naturally!

Calculator Programmes

Table mode for roots & turning points

Scientific calculator (e.g. Casio fx-991 series)

Purpose: Tabulate \(y\) across a range of \(x\) to locate sign changes (roots) and approximate maxima/minima.

When to use it: Solving or sketching a function when you want to find where its graph crosses or turns.

Steps
Enter the function in TABLE mode, set the start, end and step, then read where the sign of \(y\) changes or where it peaks.

Exam note: Allowed on papers where a calculator is permitted; use a silent scientific calculator with no stored content and show your method.

Statistics mode (mean, SD & regression)

Scientific calculator (e.g. Casio fx-991 series)

Purpose: Read the mean \(\bar{x}\) and standard deviation directly, and the gradient/intercept (and \(r\)) of a linear regression for bivariate data.

When to use it: Any data-handling, statistics, or required-practical analysis question.

Steps
Enter the data in STAT mode (1-VAR or A+BX), then recall \(\bar{x}\), \(\sigma\) or the regression coefficients.

Exam note: Allowed on papers where a calculator is permitted; use a silent scientific calculator with no stored content and show your method.

Carry exact values with Ans & memory

Scientific calculator (e.g. Casio fx-991 series)

Purpose: Keep full-precision intermediate values to avoid rounding errors.

When to use it: Multi-step calculations where premature rounding loses the final accuracy mark.

Steps
Use Ans, STO/RCL or the M+ memory to reuse the unrounded result of each step; round only the final answer.

Exam note: Allowed on papers where a calculator is permitted; use a silent scientific calculator with no stored content and show your method.

Equation solver — to CHECK your working

Scientific calculator (e.g. Casio fx-991 series)

Purpose: Use the built-in EQN/SOLVE mode to verify roots of quadratics or simultaneous equations you have already solved by algebra.

When to use it: As a check only, after solving by hand.

Steps
Enter the coefficients in EQN mode (or use SOLVE) and confirm they match your worked solution.

Exam note: Allowed on papers where a calculator is permitted; use a silent scientific calculator with no stored content and show your method.

Common Mistakes

  1. 1highMarks at stake: 4Number systems

    Converting binary operands to denary first, performing the addition in denary, and converting the sum back to binary.

    How to avoid it: Always perform the binary addition column-by-column directly in binary. Ensure you write down explicit carry bits (e.g., 1s carried to the next left column) on the paper to secure working and method marks.
  2. 2highMarks at stake: 3Algorithm design and problem-solving

    Confusing validation with verification during database or user input scenarios.

    How to avoid it: Remember that validation is an automated, rules-based computer system check (e.g., range check, type check). Verification is checking that data matches its original source, done via double-entry or manual visual checks by human users.
  3. 3mediumMarks at stake: 1File handling

    Omitting the CLOSEFILE statement after executing operations in file-handling tasks.

    How to avoid it: Always match every open file procedure with a corresponding CLOSEFILE command before completing your algorithm or subroutine.
  4. 4highMarks at stake: 1Computer architecture

    Writing generic, simple summaries for CPU performance terms, like defining clock speed strictly as 'how fast the CPU runs'.

    How to avoid it: Use precise definitions: Clock speed is the number of cycles or instructions processed per second by the CPU, measured in Hertz (Hz) or Gigahertz (GHz).
  5. 5highMarks at stake: 2Programming concepts

    Omitting explanatory comments in the 15-mark scenario-based pseudocode question in Paper 2.

    How to avoid it: Use double slashes (//) to add brief comments explaining loops, variable tracking, and output tasks. Up to 2 of the structure marks in the level-of-response grid are reserved specifically for code commentary.
  6. 6mediumMarks at stake: 1Types of software and interrupts

    Using brand names (e.g., Windows, Excel, WinRAR, Photoshop) instead of generic terms.

    How to avoid it: Never use commercial brand names in explanations of software categories. Use generic classifications like 'operating systems', 'spreadsheet applications', 'compression software', or 'image editors'.

Turn these tips into top grades

thinka turns your weak spots into targeted practice, with instant marking and exam-style feedback. Study smarter, not longer.

Practise real exam questions with instant AI feedback and marking.

Start Practising Free