What software and paper format are required for 2027 SEC G3 Computing K349? SEAB’s official 2027 G3 Computing K349 syllabus has two compulsory examinations: Paper 1 Written, 2 hours, 80 marks and 60%; and Paper 2 Lab-Based, 2 hours 30 minutes, 70 marks and 40%. Paper 2 contains one spreadsheet question and four to five programming questions, testing the Algorithms and Programming and Spreadsheets modules. Candidates use a computer with access to spreadsheet software, Python and JupyterLab; the centre provides suitable infrastructure. SEAB supplies a Python quick-reference guide, and the time allowance includes saving the required files for electronic submission.
The official 2027 G3 private-candidate directory lists software including Microsoft Excel, Python and JupyterLab Desktop for K349. That does not mean candidates are free to use any personal laptop, external packages or internet resources during the examination. The precise computing environment, installed tools, file-saving arrangements and candidate instructions must come from the authorised centre. Effective preparation develops programming logic, data-handling accuracy, spreadsheet reasoning and the habit of testing solutions—not simply familiarity with menu buttons. Students should also learn the wider consequences of computing, including security, privacy, bias and responsible data use.
- Subject code: G3 Computing K349, earlier O-Level code 7155; distinct from new G1 Computing K127 and G2 Computing K237.
- Paper 1: written, two hours, 80 marks, 60%; includes questions across all five Computing modules.
- Paper 2: lab-based, two hours thirty minutes, 70 marks, 40%; one spreadsheet question and four to five programming questions.
- Examination software: spreadsheet application, Python and JupyterLab in the authorised computer environment.
- Python reference: quick-reference guide provided; candidates still need to design and debug actual solutions.
- Paper 2 submission: required softcopy files must be saved according to instructions within the examination time.
- Calculators: SEAB permits an approved calculator in Papers 1 and 2.
- First action: confirm the subject code, read the official syllabus and practise an unfamiliar programming problem using deliberate testing.
1. K349 is not the same as G1 K127 or G2 K237
The first Singapore-Cambridge Secondary Education Certificate takes place in 2027, and G1, G2 and G3 subjects appear under a common qualification with different levels. Computing appears in more than one level, but the subject entries are distinct. SEAB identifies G1 K127 and G2 K237 as new subjects in their respective 2027 lists. G3 K349, by contrast, has the earlier O-Level reference code 7155. A parent should not treat every new SEC code as proof of an entirely new programme or assume that all Computing papers share the same software and depth.
The learner’s actual school entry matters. A student who takes G3 K349 should follow its precise syllabus, including Python programming and spreadsheets as examined. Using a G1 lesson in a visual programming environment may be helpful for a concept, but it is not an adequate substitute for G3’s own assessment requirements. Conversely, a new G1 learner need not be pushed through K349 programming simply because a parent believes harder material is automatically better.
Computing education develops abstraction, decomposition and systems thinking. Those abilities can grow across levels without pretending the content is identical. A child who understands a loop in one visual context may be ready to learn its Python form; another may need more practice connecting the idea to a real problem. A good teacher identifies the capability that transfers and the new representation the student must learn.
2. Exact 2027 K349 paper structure
| Paper | Mode | Duration | Marks | Weight |
|---|---|---|---|---|
| Paper 1 | Written examination | 2 hours | 80 | 60% |
| Paper 2 | Lab-based computing examination | 2 hours 30 minutes | 70 | 40% |
Both papers contain compulsory questions. Paper 1 tests all five Computing modules and can include multiple-choice items with one or multiple correct answers, short answers, matching, cloze and structured work. Paper 2 assesses the programming and spreadsheet modules through hands-on tasks and saved digital submissions. The syllabus supplies relevant formulae for Paper 1 and a Python quick-reference guide for the lab examination.
According to the 2027 specification table, the overall assessment balance is about 30% knowledge and understanding, 40% application and 30% design, development, testing and refinement. That matters because software familiarity is only a tool. An answer may run without fulfilling the problem’s requirements, and a confident explanation may be insufficient if the candidate cannot produce the needed functional file. The subject rewards understanding translated into working, tested solutions.
3. The five modules: what each contributes
The syllabus contains five core modules: Computing Fundamentals; Algorithms and Programming; Spreadsheets; Networking; and Impact of Computing. Paper 1 can examine all five. Paper 2 focuses on Algorithms and Programming plus Spreadsheets. A student who spends every revision hour coding may become more fluent in Python but leave fundamental computing concepts, network understanding or ethical questions weak. Conversely, a pupil who memorises definitions without building software may struggle in the lab.
Computing Fundamentals explores how information is represented and processed, the functions of computer components and ideas such as logic gates and data representation. Algorithms and Programming turns problems into sequences, decisions, repetitions and testable code. Spreadsheets connect tabular data with formulas and decision-making. Networking examines how computer systems communicate and what limitations or security implications arise. Impact of Computing considers the social, legal, economic and ethical consequences of digital technology.
These modules reinforce one another. A program uses data representations, relies on underlying hardware, may exchange information across a network and can affect real people if designed badly. Teaching them as unrelated vocabulary lists wastes that connection. Encourage students to explain how a specific computational solution works at more than one level.
4. What “lab-based examination” actually means
SEAB states that the examination centre provides adequate hardware and software and that each candidate has sole use of a personal computer for Paper 2. This wording describes the computer assigned for examination use; it is not permission to bring a private laptop. Students should follow the school’s or SEAB centre’s actual candidate instructions for equipment, software and file access.
Paper 2 lasts two hours thirty minutes, and that time includes saving the required work. Candidates submit softcopies of the requested files for marking. A functional result that is saved under the wrong name or location may create an avoidable submission problem. A good practice routine therefore includes reading the required file format, checking that the intended work is saved and verifying that the saved version is the correct one, within whatever procedures are officially permitted.
Do not assume the exam computer will have internet access, arbitrary Python libraries or plugins chosen by the student. The official subject and candidate instructions establish the environment. Practise with the tools the syllabus specifies and with school guidance about what is available. Useful study materials should not promise that importing a third-party package will solve tasks if the installed examination environment does not support it.
5. The spreadsheet question is about relationships, not formatting
Paper 2 includes one spreadsheet question. Spreadsheet competence is more than entering numbers into cells or making a colourful chart. A student should understand how formulas represent a relationship, how references behave when copied and how data can be checked for errors. A sheet that produces one correct number after manual adjustment may be less reliable than a formula that updates correctly when inputs change.
Consider an original example. A class records four weekly attendance counts in cells B2:B5: 18, 22, 20 and 24. A formula such as =SUM(B2:B5) returns 84, while =AVERAGE(B2:B5) returns 21. These are simple operations, but they teach the difference between total and mean. If the question asks for the maximum weekly attendance, the appropriate calculation is different. Choosing the function depends on understanding the task.
Another exercise tests references. A worksheet calculates a price after a discount using a rate in a fixed cell. A student who copies a formula and unintentionally shifts the rate reference can obtain different answers across rows. The correct approach is to understand relative and absolute references when relevant to the taught syllabus. More importantly, the student should test the formula with changed inputs and explain why the output behaves as expected.
6. Validate inputs before trusting a spreadsheet
A real spreadsheet can contain inconsistent or missing data. A cell may hold the text “twenty” instead of the number 20, or a date may be stored in an unexpected format. A student should know that data type and validation affect calculations. A neat display is not proof that every formula is meaningful. The question is whether the sheet’s logic matches the intended problem and handles the provided cases correctly.
For an original teaching case, imagine a list of school-event expenses with quantity and unit price columns. A student must calculate each line total, sum the expenses and compare the result with a budget. Before writing formulas, identify what the table represents and which entries should be numeric. Then test a row with a quantity of zero and a row with a larger value. The learner develops the habit of checking assumptions instead of treating the first successful calculation as proof.
Spreadsheets are used in finance, science, public administration and everyday planning. Accuracy has consequences. A mislabeled unit or shifted reference can propagate a mistake into many rows. Computing education should teach students to notice such risks and explain how their formulas preserve the intended meaning.
7. Algorithms begin with a clear problem specification
Before writing Python, identify the input, output, constraints and conditions of success. For example, a simple program might calculate a total price from quantity and unit cost. It should know which inputs are allowed, whether quantities can be zero and how values should be presented. Without these definitions, code can run correctly according to its syntax while failing the intended task.
The syllabus’s programming module includes problem analysis, control structures and other Python concepts. Decomposition means breaking a large task into smaller parts that can be understood and tested. Abstraction means retaining the relevant details while ignoring those that do not affect the solution. Both are practical reasoning tools. A student who has learnt only to copy a function may struggle when the input format changes; a student who understands the requirements can adapt the design.
Ask the learner to describe the algorithm in ordinary language before coding. What needs to happen first? When will the program make a choice? Which part repeats? What would count as an error? A clear specification makes later debugging easier and reduces the temptation to modify random lines until the output happens to look plausible.
8. Original Python example: a simple transparent calculation
The following short program is an original educational illustration, not a past-year exam problem. It calculates the price of an order and handles a basic invalid quantity.
def order_total(quantity, unit_price):
if quantity < 0 or unit_price < 0:
return None
return quantity * unit_price
print(order_total(3, 4.50))
print(order_total(0, 4.50))
print(order_total(-1, 4.50))The outputs are 13.5, 0.0 and None, respectively, in ordinary Python 3 execution. The code illustrates a function with parameters, a conditional and different input cases. It does not establish how a real shop should deal with taxes, refunds or currency rounding, and it does not claim to reproduce an official K349 question. The teaching value is in asking why a negative quantity is handled differently and whether the chosen behaviour satisfies a specified problem.
Next change the task: instead of returning None, suppose the program must display a meaningful error message or require a new input. The solution design must change because the requirement changed. That is the essence of programming judgement. A student should not modify code without recognising which output or behaviour the revised task actually demands.
9. Selection, iteration and functions: understanding control flow
Programming often combines sequence, selection and repetition. A sequence executes steps in order. Selection chooses a path based on a condition. Iteration repeats a block while a rule applies or across a specified collection. Functions group related work into reusable units. These concepts are more important than memorising a few popular Python statements.
For an original exercise, give a list of scores and ask the student to count how many meet a stated threshold. The first conceptual step is to identify the condition, then examine each item and update the count only when the condition holds. A learner can implement this through a loop and an if statement. The teacher should then ask for test cases: an empty list, values exactly on the threshold, and values just above or below it.
Off-by-one errors are common when counting, indexing or iterating. A program may work with three entries but fail with an empty list or a boundary value. Testing only the example supplied in the question can create false confidence. Encourage students to explain what the loop does on every iteration and why it terminates. That clarity supports reliable debugging.
10. Testing and debugging: the difference between running and working
A program that displays an answer is not necessarily correct. It may fail on unusual inputs, repeat a calculation or produce the wrong output format. Testing compares observed behaviour with explicit requirements. Debugging investigates why the behaviour differs. Randomly editing code until a familiar example passes is inefficient because the underlying mistake may remain.
Use a simple test register with columns for input, expected output, observed output and conclusion. Include ordinary cases, boundaries and inappropriate inputs where relevant. For the original order-total function, useful tests include quantity three, zero and a negative value. The required handling of an invalid input depends on the task specification. A real exam question may define its own valid range, so students must read it carefully.
When a test fails, ask which assumption is wrong. Did the algorithm misunderstand the data, did a condition use the wrong comparison, or did a variable update occur at the wrong time? The answer determines the correction. Debugging is a process of reasoning about cause and effect, not a reflection of the learner’s worth.
11. File input, output and saving examination work
The K349 syllabus includes file handling and submission of softcopy work in Paper 2. Students should understand why files must be opened, read or written according to a specification and why saving the intended result is part of completing a computing task. A program may produce the correct output on screen while failing to create the required saved file.
For classroom practice, use a dedicated folder with innocuous test data and clear filenames. Ask the learner which file is the input, which is the output and how they will confirm that the saved file contains the correct content. Avoid using personal accounts, passwords or sensitive student information in practice examples. The ability to handle files responsibly is relevant beyond school assessments.
Examination file paths, submission naming and permitted storage locations are defined by centre instructions. An online guide should not invent a path or tell candidates to use external storage during a supervised examination. Teach the underlying responsibility—save and verify accurately—then follow the actual instructions on the day.
12. The supplied Python quick-reference guide is not a substitute for learning
SEAB supplies a quick-reference guide for Python during the Paper 2 lab examination. The document helps candidates recall permitted syntax and functions, but it does not write an algorithm for them. A learner must still understand input requirements, choose control flow, represent data and debug the resulting solution. Looking up a function name is useful only when the student knows what they need the function to do.
A common mistake is to practise using libraries or features from a general online programming tutorial without checking whether they are within the assessed syllabus or installed in the examination environment. Some tools can be valuable for enrichment, but the core revision plan should follow the current K349 document and the school’s approved environment. The goal is reliable competence under the actual conditions, not the largest collection of Python packages.
Students should also develop basic fluency so that every line does not require a search. Knowing how conditions, loops, data structures and functions work frees attention for the problem’s reasoning. The reference guide then supports accuracy rather than becoming a crutch.
13. Networking appears in Paper 1, not the Paper 2 lab focus
Networking is one of the five K349 modules. Students need to understand relevant concepts and their consequences rather than only recognise a router in a diagram. A network connects devices and services, but the details of communication, addressing, data transfer and security create real questions about reliability and privacy. The precise syllabus defines which concepts are assessed.
Paper 2 focuses on Spreadsheets and Algorithms and Programming, while Paper 1 can examine networking and the other modules. This matters for time allocation. If a child practises only Python after school, they may become confident in the lab while neglecting the written paper worth sixty percent. A balanced approach uses small retrieval sessions for networking and computing fundamentals alongside practical programming.
Teach safe and responsible behaviour when discussing networks. Students should understand why passwords, access permissions and careful handling of personal information matter. They do not need to experiment with another person’s devices or accounts to learn these concepts. Classroom simulations and lawful examples are sufficient.
14. Impact of Computing: ethics, privacy, bias and responsibility
The Impact of Computing module makes the subject broader than writing code. Digital systems affect individuals and society. An algorithm may unintentionally disadvantage a group if its training or input data are unrepresentative. A poorly designed form may collect personal information that the service does not need. A fast system can still produce harmful decisions if nobody checks its assumptions. Students should learn to ask who benefits, who might be harmed and what evidence shows a system works responsibly.
Consider an original hypothetical school application that recommends reading books based only on past borrowing records. A student who has never borrowed from the library may receive poor recommendations, not because they lack interests but because the dataset does not represent them. A responsible designer might allow users to state preferences and avoid treating missing historical data as proof of disinterest. This is a safe example of bias, data quality and human-centred design.
Computing literacy includes knowing the limits of automation. A program can apply rules consistently, but the rules may be wrong, incomplete or inappropriate for the user. Ethical thinking is therefore connected to testing and validation. The same habit that checks a spreadsheet formula can support careful scrutiny of a system’s effect on real people.
15. Data representation: bits and bytes still matter
Computing Fundamentals includes binary, hexadecimal, representations of signed values, logic gates and storage concepts within the official scope. These may feel distant from a spreadsheet but explain how digital systems represent information. A student who understands binary place values can interpret how a computer stores numbers instead of treating every digital display as magic.
For a simple original example, binary 1011 corresponds to decimal 11 because 8 + 0 + 2 + 1 = 11. The computation is a place-value exercise. More advanced representations include fixed-width signed integers and data encoding, where the number of bits and chosen convention matter. The syllabus determines how much detail is examined, so use its listed outcomes rather than adding unrelated university-level computer architecture.
Logical reasoning also links to programming conditions. An AND condition requires both propositions, while an OR condition accepts one or the other according to its defined meaning. A poorly designed condition can reject a valid input or accept an invalid one. A truth table makes the relationship explicit and helps students debug decisions in a program.
16. Time management in the 150-minute lab
Paper 2’s two-and-a-half-hour duration includes time for saving submissions. Students need to read all task requirements, plan a sensible order and leave time for testing. Spending most of the period polishing one spreadsheet’s appearance can jeopardise the programming section. Conversely, rushing into code without understanding the task may produce a program that runs but answers the wrong problem.
A reasonable study method is to practise estimating the complexity of a task before starting. Which inputs are known? What output must be saved? Which part can be tested independently? If a programming task is stuck, can the pupil isolate a small example that reveals the problem? These are transferable habits. The exact examination sequence and submission procedure must follow school and SEAB instructions rather than an arbitrary generic schedule.
Saving and verification should become routine in school practice. Students should know whether the required file has been created, contains the intended result and has been placed according to the authorised instructions. The goal is confidence based on a reliable process, not panic in the final seconds.
17. Case study: a student knows syntax but cannot design an algorithm
A Secondary 4 candidate can copy and adapt simple Python examples but freezes when the wording changes. The tutor finds that the pupil starts coding before identifying the required input and output. Lessons begin with problem decomposition, a short algorithm in ordinary language and explicit test cases. After this planning, the student writes code more accurately. The improvement is in computational thinking, not memorisation of more functions.
18. Case study: a good spreadsheet fails when copied
Another student produces the correct total in one row but obtains wrong results after copying the formula. The teacher identifies a misunderstanding of cell references and helps the learner test the behaviour across multiple rows. The formula is then revised to preserve the intended reference. The pupil learns to distinguish a correct first output from a robust reusable spreadsheet model.
19. Case study: the family practises G1 tools for a G3 examination
A parent searches “SEC Computing practical” and purchases beginner lessons that focus only on visual programming. The school confirms G3 K349, whose lab examination requires Python, JupyterLab and spreadsheet tasks. The family keeps introductory exercises where they teach useful control-flow concepts but reorganises the core study programme around K349’s actual syllabus. Correct information reduces wasted time and avoids unfairly blaming the child for difficulty with the wrong materials.
20. A four-week K349 learning reset
Week 1: verify the subject and assess starting skills
Download the official K349 PDF and check the student’s actual registration and school software guidance. Record Paper 1 and Paper 2 weights and tasks. Give short exercises in Python control flow, spreadsheet formulas and explanation of a networking or ethics concept. Identify the earliest recurring error: misunderstanding requirements, weak syntax, data-type confusion or incomplete testing.
Week 2: rebuild program logic
Teach decomposition, inputs and outputs, conditions, loops and functions at the relevant scope. Have the pupil write a short algorithm before coding, then test ordinary and boundary cases. Correct the first failure mechanism and retry with an unfamiliar problem. Use the official Python quick reference as a learning aid without replacing understanding.
Week 3: integrate spreadsheets and file handling
Practise a fresh spreadsheet problem involving data types, formulas and a decision. Test the formula with changed inputs. Practise saving harmless school-created output files correctly and verifying that they contain the intended result. Mix in conceptual written questions from the other syllabus modules so preparation remains balanced.
Week 4: test transfer and strengthen independence
Give an unfamiliar programming and spreadsheet challenge using the school-approved environment. Ask the student to describe the requirements, design, tests and result before looking at a model answer. Review errors precisely, then adjust the next learning target. Keep a sustainable routine that allows normal school and family life.
21. Frequently asked questions
What is the 2027 G3 Computing code?
K349, with previous O-Level reference 7155. It is separate from G1 Computing K127 and G2 Computing K237.
How many papers are there?
Two compulsory papers: Paper 1 Written and Paper 2 Lab-Based.
How long is Computing Paper 2?
Two hours thirty minutes, including time to save required work. It carries seventy marks and forty percent of the overall subject grade.
What software is used in Paper 2?
The official syllabus specifies a spreadsheet, Python and JupyterLab. The 2027 private-candidate directory lists Microsoft Excel, Python and JupyterLab Desktop. Confirm the authorised centre environment and current instructions.
How many spreadsheet questions are in Paper 2?
One spreadsheet question, alongside four to five programming questions under the published 2027 scheme.
Is a Python quick-reference guide provided?
Yes. It helps with syntax and specified functions, but candidates must still design, implement, test and refine their own solutions.
Can candidates use a personal laptop?
Do not assume so. The centre provides the computer and software facilities for the lab examination. Follow the actual centre and SEAB candidate instructions.
Do students need to submit electronic files?
Yes. The paper requires saved softcopies of specified work, and the time allowance includes saving them. Check the exact file instructions during the examination.
Does Paper 1 assess programming only?
No. Paper 1 covers all five modules, including Computing Fundamentals, Networking and Impact of Computing, not just algorithms.
Are calculators allowed?
The 2027 K349 syllabus permits an approved calculator in both papers, subject to current SEAB rules.
Is K349 entirely new in 2027?
The SEC subject has a new K-code but a previous O-Level reference of 7155. G1 K127 and G2 K237 are separately marked new in the 2027 directories. Do not generalise across levels.
What is the best way to prepare for Python tasks?
Practise analysing the problem, choosing data structures and control flow, writing clear code and testing unfamiliar inputs. Repeating one memorised program is weaker evidence of readiness than solving a new problem independently.
22. Official sources and further eduKate reading
Official sources, checked 7 November 2026: SEAB 2027 G3 Computing K349 full syllabus, Python quick-reference guide and Paper 2 specification; G3 school-candidate subjects; G3 private-candidate software listing; SEC overview. The programming and spreadsheet exercises above are original learning examples, not reproduced examination tasks.
Related eduKate guides: new G1/G2 Computing versus existing G3 Computing; Computing across G1, G2 and G3; G3 Computing Quick Notes; and digital examination preparation across SEC subjects.
The value of Computing is not simply the ability to make a computer perform an action once. It is the ability to understand a problem, design a reliable process, test it and recognise how the result affects real people. That is the capability worth building for K349 and beyond.