G2 Computing K237 is a new subject in the 2027 Singapore-Cambridge Secondary Education Certificate (SEC). Its 135-minute Paper 2 is a computer-laboratory examination worth 80 marks and 50% of the final result. Candidates complete practical tasks in media software (approximately 30 marks), spreadsheets (approximately 25 marks) and programming (approximately 25 marks), then save the required files for marking within the examination time. This playbook trains the independent skills needed to build, test and save those outputs.
Paper 1 is a separate 90-minute electronic examination worth another 50%; it assesses all six computing modules and must not be neglected. For the lab, SEAB lists Microsoft Office Excel and PowerPoint and Scratch as relevant software, subject to the authorised examination environment. The most useful way to prepare is to follow a task specification, produce a correct artefact, check its behaviour and confirm that the appropriate final files exist. These are more important than memorising which icon to click in one practice project.
- Confirm the code: G2 Computing K237 is new for 2027; G3 Computing K349 has a different software and examination scheme.
- Paper 2: 2 hours 15 minutes, 80 marks, 50% overall, with three compulsory practical task areas.
- Media: visual graphics, slide and video skills, approximate 30-mark allocation.
- Spreadsheets: formulas, cell references, charts, functions and validation, approximate 25-mark allocation.
- Programming: Scratch variables, conditions, loops, lists and testing, approximate 25-mark allocation.
- Save: required softcopies must be saved correctly within the official lab time.
- Study: practise new specifications and debug independently, not just replay one prepared project.
Updated 12 November 2026. All example datasets and programming stories are original educational illustrations, not actual or predicted SEC 2027 exam tasks. The official K237 syllabus and examination-centre instructions are authoritative.
1. K237 is not merely the ability to use a computer
G2 Computing aims to develop conceptual understanding and practical creation. Students learn Computing Fundamentals, Networking, Impact of Computing, Spreadsheets, Media Software and Programming. Paper 1 examines all six modules, whereas Paper 2 draws on the final three. Knowing how to open a spreadsheet is not evidence of being able to construct a formula; knowing how to play a Scratch game does not mean the learner can debug one.
Under Full Subject-Based Banding, G2 is the level of this particular elective rather than a label for every subject a pupil takes. A student entering Secondary 3 should confirm that their school offers K237 and that the learner is actually registered at that level. Some schools may offer different computing electives or arrangements. Parents should not treat a generic coding class as automatically aligned with this national qualification.
2. The precise Paper 1 and Paper 2 weighting
The official K237 specification gives Paper 1 as a 90-minute e-examination, 70 marks, worth 50% overall. Section A contains twenty four-option multiple-choice questions worth 20 marks, and Section B contains short structured questions worth 50 marks. Both sections are compulsory and address the six modules.
Paper 2 is lab-based, 135 minutes, 80 marks, worth the remaining 50%. It contains three practical task areas: media software, spreadsheets and programming. The approximate mark allocations are thirty, twenty-five and twenty-five. The syllabus says the allotted time includes saving required work. A student who creates a perfect output but fails to produce a required saved file may be unable to demonstrate it for marking.
3. Do not copy G3 Python preparation into G2 Scratch
G2 K237 and G3 K349 are different Computing subjects. The G2 directory identifies Scratch for its programming environment, alongside approved Microsoft Office software. G3 uses Python and JupyterLab under its own assessment arrangements. A child may enjoy Python as enrichment, but that does not replace learning the Scratch blocks, project structure and game programming expected for K237.
The K237 syllabus also includes microcontrollers and MakeCode blocks in Paper 1 only. Parents sometimes misunderstand that sentence and buy hardware to prepare for a practical MakeCode lab task. The published Paper 2 modules include the relevant Scratch programming skills, not a required microcontroller build in the practical examination. Follow the actual K237 scope before spending on equipment.
4. Begin with a task specification, not a software icon
An examination task tells the candidate what result must be produced. For a spreadsheet it could specify calculations, formats, filters and a chart. For media it could require an image, presentation or video with stated properties. For Scratch it could require an interaction, scoring system and conditions for finishing. The student must translate instructions into a plan before using the application.
A useful practice page has three columns: requirement, planned tool and verification method. For example, if a fictional spreadsheet task asks to highlight values greater than 50, the planned tool is conditional formatting and the test is whether 49 remains unhighlighted while 51 receives the intended style. This is more reliable than clicking around until something looks similar.
5. The first diagnostic is the first broken requirement
When a student loses marks in the lab, do not merely say that they need to learn Excel or Scratch. Identify what failed: misunderstanding the prompt, selecting a wrong cell range, using a relative reference where an absolute one was needed, building a loop that never ends, choosing inappropriate visual hierarchy or saving the wrong version of a file.
Keep one error log with requirement, first wrong decision, corrected method and independent retest. A copied formula or rebuilt game can appear correct without revealing whether the student can apply the method when the data or instructions change. Give a parallel task with altered ranges or conditions several days later. This checks transferable competence.
6. Spreadsheet data types affect results
A spreadsheet cell may contain a number, text, date, percentage, currency or a formula. A value stored as text can behave differently from a number when used in calculations. Students should be able to recognise which type is appropriate for the task and apply suitable display formats. The syllabus includes number, currency, percentage and specified date formats.
Imagine a fictional table of school-event spending where one column shows quantities and another unit prices. If the unit price is entered as text including an extra character rather than a numeric value, a calculation may fail. Fixing the display format alone is not necessarily enough; the underlying value must be valid. Ask the learner to check both how the cell looks and what it contains.
7. A formatted percentage is still a numerical value
In Excel, a stored decimal value of 0.25 can be displayed as 25% when the cell is formatted appropriately. That is different from entering the number 25 and applying percentage formatting, which represents 2,500%. A student who misunderstands this relation may produce totals that look impressive but are mathematically wrong.
Use an invented example where a discount rate is 10% and a listed price is $80. With a numeric rate of 0.10, a formula multiplying the price by the rate produces an $8 discount. Ask the learner to change the rate to 15% and see whether the result recalculates correctly. Data representation should remain consistent with mathematical meaning.
8. Relative references change when a formula is copied
Suppose B2 contains a price and C2 a quantity. A formula in D2 such as =B2*C2 calculates the line cost. When copied down one row, relative references become B3 and C3. This is useful when each row represents a different item, but a pupil must understand why Excel changes the references.
A student who hardcodes the first row’s numbers instead of referencing cells may produce a correct-looking D2 that fails when the dataset changes. The more reliable method builds a formula from the table’s structure, copies it appropriately and checks one other row against manual arithmetic. A changed dataset tests whether the learner understands the design.
9. Absolute references protect a fixed assumption
Imagine that F1 contains a uniform discount rate and column B contains item prices. A formula such as =B2*(1-$F$1) can be copied down while preserving the reference to F1. Without the dollar signs, the reference to F1 may shift as the formula is copied and point to the wrong cell.
Ask the student which values should vary by row and which should remain fixed. The dollar signs are not decoration; they communicate reference behaviour. A good test changes F1 after the formulas have been copied. All discounted prices should recalculate under the new rate while retaining their separate base prices.
10. An IF formula expresses a decision rule
The official K237 syllabus includes IF with up to one level of nested IFs. A fictional sheet might put a score in B2 and display a category in C2. A suitable basic example is =IF(B2>=50,”Met”,”Not met”), under the stated threshold. This is only a spreadsheet-training rule, not a real school grading scale.
The student should test the boundary deliberately. Values 49, 50 and 51 reveal whether the comparison includes the threshold. A formula that uses greater than instead of greater than or equal to can misclassify the boundary. The important skill is translating a verbal condition into an exact comparison, not memorising one expression.
11. AND, OR and NOT describe different logical conditions
A task may require an outcome when two conditions are both true or when at least one is true. AND returns true when all its arguments are true; OR is true when any listed argument is true; NOT reverses a logical value. Their meaning should be understood before combining them inside IF.
For an invented activity, a participant may qualify for a badge only if attendance and task completion both meet requirements. The student should identify whether the word and or the word or appears in the specification. Test cases where only one condition is true; these expose a formula that accidentally uses OR where AND was required.
12. One-level nested IF is not an invitation to build unlimited logic
K237 includes IF up to one level of nesting. A simple practice rule may classify a number as high, medium or low using a top-level IF and one nested IF. The learner should first write the decision tree in ordinary language, then map each branch to the formula. A complex construction may be harder to debug and outside what the task needs.
Use boundary values to test every branch. If the cut-offs are 80 and 50, test values below, at and above each threshold. This reveals whether equality and the order of checks are correct. The goal is to meet the requirement with dependable logic, not to create the longest formula possible.
13. SUM and SUMIF answer different questions
SUM adds the values in a specified range. SUMIF adds values that meet a condition. Suppose A2:A6 contains fictional activity categories and B2:B6 contains expenses. A formula such as =SUM(B2:B6) gives the total of all recorded expenses, whereas =SUMIF(A2:A6,”Transport”,B2:B6) adds only rows labelled Transport.
A pupil who uses the wrong criteria range may get an incorrect result without a visible error message. Check the sizes of the criteria and sum ranges and test against a manually identified subset. The student should be able to explain what each argument represents before relying on the output.
14. AVERAGE and AVERAGEIF require valid observations
AVERAGE computes the mean of numerical values in a range, while AVERAGEIF limits the calculation to values meeting a condition. An invented dataset might record travel time for bus and train users. A candidate asked for the average time among bus users should select only those relevant records.
Check whether empty cells, zeroes and text entries are being handled as intended. A missing travel time should not automatically be interpreted as an observed journey lasting zero minutes. Averages are meaningful only when the included observations match the specified question. A correct-looking number can still be based on a wrong subset.
15. COUNT, COUNTA and COUNTBLANK are not interchangeable
COUNT counts numeric entries in a range, while COUNTA counts non-empty entries and COUNTBLANK counts blank cells. In a fictional attendance sheet, a student may need to know how many scores are recorded, how many rows contain identifiers and how many entries remain incomplete. These can produce three different results from the same columns.
Use a miniature practice table with text names, numerical scores and a few genuinely blank cells. Ask the learner to predict each count before writing a function. When the prediction and spreadsheet differ, examine what Excel considers empty or numerical in those cells. That builds understanding of data types rather than relying entirely on trial and error.
16. COUNTIF checks a criterion, not every row
COUNTIF counts cells satisfying a condition. A fictional spreadsheet could count how many participants chose the word “Science” in an interest column or how many numeric values exceed a specified threshold. The criterion must refer to the correct range and match the requested relationship.
A useful test adds one new row that meets the criterion and one that does not. The count should change only as expected. This exposes range omissions and incorrectly typed criteria. Students should also understand that criteria based on the text inside a cell differ from criteria based on its formatting or colour.
17. Sorting changes order; filtering changes visibility
Sorting places records into a chosen order, such as increasing scores or alphabetical names. Filtering shows only records meeting selected criteria while leaving other records available in the underlying table. A student who manually deletes rows to simulate filtering risks losing source information and may produce the wrong final artefact.
A practice task can ask for a list of all participants above a score threshold, followed by a descending sort. Verify that the same records remain and that the ordering is correct. Ask whether a filtered row has been deleted or merely hidden. This conceptual distinction is important when interpreting totals and preserving a dataset.
18. Data validation prevents inappropriate input
The K237 syllabus includes setting data-validation criteria and custom error messages. A worksheet collecting a score between zero and one hundred might reject values outside that range. The purpose is to improve consistency and reduce input mistakes, not guarantee that every accepted entry is truthful.
Ask the learner to test the boundaries: negative one, zero, one hundred and one hundred one. A custom message should help users understand the permitted values. The candidate should be able to explain why validation is useful and which invalid cases it cannot detect. A plausible but incorrect score of eighty could still satisfy the rule.
19. Conditional formatting should respond to data
Conditional formatting can highlight cells when their values meet a rule. The published K237 scope limits relevant comparisons to greater than, less than and equal to. A candidate should not colour every desired cell manually if the task specifies a rule that must update when the data change.
Use a fictional spreadsheet highlighting scores greater than fifty. Test cells containing forty-nine, fifty and fifty-one. The boundary is important: greater than fifty does not include exactly fifty. Then change a value and confirm the visual style updates automatically. That is evidence of a rule-based solution rather than a static imitation.
20. Choose a chart because of the data and question
A bar or column chart compares categories; a line chart can show a progression such as time; a pie chart communicates parts of a whole where the data suit that representation. The syllabus also includes combination charts limited to a line-and-column combination. These formats answer different questions, and one style should not be chosen merely because it looks attractive.
Consider monthly event attendance and a percentage change. A student may decide that columns display attendance counts while a line shows a separate rate, but they must first inspect the task and axis scales. A well-labelled chart can communicate two related measures, whereas an unlabelled dual-scale graphic may confuse the reader.
21. A correct chart needs title, labels and source ranges
The syllabus expects students to customise chart titles, data labels, axes, axis titles and legends. An incorrect selection of cells can make a chart misleading even when every formatting option looks polished. If a range includes an unrelated totals row, the chart might double count or distort categories.
The first check is whether plotted values correspond to the intended data. The next is whether the reader can identify units, categories and series. Then change one source cell and verify whether the chart updates appropriately. Such checks connect output appearance to data accuracy, rather than treating chart creation as a decorative software exercise.
22. LOOKUP functions require exact-match discipline
K237 includes HLOOKUP and VLOOKUP for exact matching, as well as INDEX and MATCH. An invented table of item codes and prices can be used to practise retrieving the price for an exact item code. With VLOOKUP, a formula such as =VLOOKUP(E2,A2:C10,3,FALSE) returns the third-column value for the code in E2 when the lookup table is arranged appropriately.
Students must check that the lookup value exists, the first column contains the identifiers and the returned column index matches the requested field. An approximate match where the task requires exact identification can quietly produce misleading results. Test an existing code, a different existing code and one nonexistent code.
23. INDEX and MATCH make lookup logic visible
INDEX retrieves a value from a specified position, while MATCH can find a position within a range. Under a fictional vertical data table, =INDEX(C2:C10,MATCH(E2,A2:A10,0)) retrieves a value from column C corresponding to an exact code in column A. The zero requests an exact match in MATCH.
The learner should explain how the two functions work together: one finds the matching row position and the other retrieves the value from the corresponding list. A wrong row count or misaligned range can return an unrelated item even if the formula’s syntax is accepted. The independent check is to compare output with the original data record.
24. Text functions solve a different family of problems
The syllabus includes LEFT, MID, RIGHT, LEN, CONCAT, FIND and SEARCH. These can extract parts of text, measure length, combine strings and locate characters. A fictional identifier such as AB-204 can be analysed for its prefix or numeric portion, but students must first determine the exact data format and which characters the task requests.
FIND is case-sensitive while SEARCH is not, a distinction that matters for some inputs. If a string might contain spaces or varying separators, blindly using a fixed number of characters can create errors. Test the function against records with different but permitted contents. Text extraction is about accurate data handling, not merely remembering an argument order.
25. Dates are data, not just words on screen
The K237 spreadsheet scope includes TODAY, NOW and DAYS. A genuine spreadsheet date value may be formatted visually in several ways, while an ordinary text string resembling a date may behave differently. A task asking for the number of days between two dates requires the program to interpret actual dates and their order.
Practice with a fictional school event starting on 10 March and a registration deadline on 5 March in the same year. The interval is five days when calculated in the relevant direction. Students should verify the year, stored date format and whether the question requests signed or positive differences. Unclear date representation can produce errors even with a valid function name.
26. Why a spreadsheet that looks right can still be wrong
A candidate might manually type totals, highlight a fixed set of cells or draw a chart from the wrong range while producing a screenshot that looks correct. But the file’s internal formulas, relationships and validation may not meet the task. A practical assessment requires an artefact that behaves correctly when input data change.
Use a changed-input test: alter one quantity, one price and one condition, then see which formulas and charts update. Restore the original data when required by the final task. This reveals hardcoded values, missed references and conditional rules that only imitate the intended output. Working behaviour matters as much as surface appearance.
27. Vector graphics and raster images are different
Vector graphics describe shapes using nodes and paths and can generally be resized without the same kind of pixelation seen when enlarging a small raster image. Raster graphics consist of coloured pixels. The K237 media syllabus expects students to understand both and manipulate appropriate objects and effects.
A practice poster might use a vector icon that can scale cleanly alongside a photographic raster background. If the photograph is enlarged excessively from a low-resolution source, it may appear blurred. The student should choose resources appropriate to the output and understand why the underlying representation affects quality, not simply keep adding decorative effects.
28. Layers, grouping and alignment support clean design
Media software tasks may require objects to be moved, resized, rotated, flipped, grouped and arranged from front to back. A well-designed image or slide needs content order. If a caption is hidden behind a shape, the intended communication fails regardless of how attractive the colours are.
A student can practise building a fictional information graphic with three labelled elements, a title and a supporting illustration. Verify that text is readable, objects are aligned and the requested grouping or layer order is correct. Then change a requirement and adapt the layout without rebuilding everything from scratch. That tests tool competence and design reasoning.
29. Cropping and resizing have different consequences
Resizing changes the dimensions of an image, while cropping removes portions outside a selected region. A student who crops a figure when the task asks for preserving the whole subject may lose important information. Enlarging a low-resolution raster image may cause visible pixelation. The choice should be linked to the final product’s purpose.
Imagine a fictional school-event poster requiring an accessible route map. Cropping away the legend could make the map unusable, even though the cropped composition looks cleaner. A strong candidate checks whether essential content remains legible and whether the visual meets the stated requirement before exporting.
30. Presentation design should follow a storyboard
K237 includes producing slide presentations from storyboards and using slide-master features for consistent style and layout. A student who adds random animations and text boxes may produce an attractive deck without meeting the planned communication sequence. Begin with the message, audience and required slide content.
A fictional presentation explaining a community recycling programme could follow problem, process, participation instructions and final reminder. A consistent slide master can standardise fonts and placement, reducing manual inconsistencies. The candidate should check both the visual sequence and the actual information on each slide. A coherent storyboard saves time and helps avoid missing a compulsory element.
31. Video editing requires attention to sequence and sound
The syllabus includes creating videos from still images, clips, text, transitions and sound. A candidate should understand that videos are sequences of frames and that frame rate affects motion smoothness and potential file size. The actual lab task may specify content and export requirements; these should govern decisions rather than an arbitrary social-media preset.
An original practice video could explain a school library’s borrowing steps in four short scenes. The student should place the scenes in a logical order, use readable labels and check that sound does not obscure the message. Excessive transitions can distract from the required content. Test the rendered result rather than assume the editing timeline guarantees the final file plays correctly.
32. Design quality is a communication decision
The purpose of a practical media product is usually to convey information, tell a story or present a useful visual. A candidate may achieve technical operations but still make a slide impossible to read by using low-contrast text, cramped spacing or unnecessary visual effects. A clear hierarchy—headline, supporting message and relevant detail—often improves the result.
Before finishing, view the artefact at its intended size. Are names, dates and units legible? Does the picture support the text? Is the audience obvious? These checks are part of producing a purposeful digital product, not merely personal taste. Students should respect the actual task specification when making design choices.
33. Scratch programs need a clear initial state
The K237 programming module uses Scratch and includes variables, loops, conditions, lists and game interactions. A common bug occurs when a game works once but behaves strangely on the second run because the score, position or visibility was not reset. An explicit initial state prevents old values from carrying over unintentionally.
For a fictional collection game, initialise score to zero, place the player at the starting location and set appropriate sprite visibility when the green flag is clicked. Then test the game twice in a row. If the second run begins with an old score, the initialization requirement has not been met. This is a useful diagnostic even without advanced programming.
34. Variables store state, not intentions
A variable might represent score, number of attempts or remaining time. Students must understand what value it contains, when that value changes and which scripts read it. A variable named score does not increase automatically when a sprite touches an object; the program requires an actual instruction.
An original practice condition could add one to score whenever a player collects a star. Ask the learner where the collision is checked and where the update occurs. A score that rises continuously while sprites remain touching may indicate that the collision event is being counted more often than intended. Testing behaviour matters more than a variable’s label.
35. If and if-else blocks encode decisions
Conditional instructions make a program choose actions based on a true or false test. In an invented Scratch game, the player might gain a point after touching a target and lose a life after touching a hazard. The learner must check which condition is evaluated and whether competing cases can both happen.
A common bug is putting the wrong action inside an else branch or using a comparison that excludes the boundary value. Trace an example with no contact, target contact and hazard contact. Then change a condition and ask the learner to predict the output before running the code. This develops algorithmic reasoning rather than random block rearrangement.
36. Loops can repeat the wrong action forever
Scratch has loops such as repeat, forever and repeat until. The difference between fixed repetitions and condition-controlled repetition matters. A forever loop without a suitable state change or exit control can keep updating a score or moving a sprite in an unintended way.
For an invented timer, a loop might decrement a remaining-seconds variable until it reaches zero. Ask what initial value is required, when the condition is checked and whether the loop can continue past zero. A step-by-step trace helps the pupil find off-by-one errors. The student should test both normal and edge cases rather than assume one successful playthrough proves correctness.
37. Sprites have positions and orientations
Scratch represents positions on a stage using x and y coordinates. A pupil may move a sprite by changing x when they intended vertical movement, or rotate it to the wrong orientation. The program may technically run while violating the task. A good practice exercise uses a coordinate sketch and tests the sprite’s expected final position.
A changed starting coordinate exposes whether the learner has hardcoded a path that only works from one location. The student should understand which coordinate controls horizontal versus vertical position and how orientation affects movement. A visual language still requires precise logic and spatial reasoning.
38. Broadcast messages coordinate multiple scripts
A game may have several sprites that need to react to the same event. Scratch’s broadcast-and-receive mechanism can coordinate behaviour, such as displaying a result screen after the game ends. A student who individually copies instructions into every sprite can create inconsistent behaviour and make debugging harder.
In a fictional game, reaching a target score might broadcast a game-over message. One sprite hides, another displays a result and a sound plays. Test that all components respond as intended and that no sprite continues collecting points after the game has ended. The skill is coordinating events, not simply using more blocks.
39. Lists are more than several individual variables
The syllabus includes creating lists, adding and removing items, reading or writing an indexed item, checking membership and joining lists. Lists can represent collections such as items found or names of completed stages. A student should understand which position an item occupies and whether the intended operation changes the list correctly.
An invented Scratch project might store three collected objects in a list. A test adds a fourth, removes the second and checks whether the list contains a requested item. The learner should predict the expected list before executing each operation. Index mistakes are easier to repair when the current state is written out explicitly.
40. Custom blocks help separate complicated actions
K237 includes custom blocks or functions built from multiple instructions. A complex animation can become easier to understand if a repeated action is represented by a named custom block. The name should describe what the block actually does rather than conceal unrelated operations.
A fictional game may use a reset-player custom block to return a sprite to its starting position and appropriate costume. Test that calling it from different scripts produces the same intended state. This teaches abstraction: the learner can reason about the block’s purpose while preserving a reusable implementation. The task specification determines when such design is appropriate.
41. Debugging begins with a reproducible failure
Randomly dragging blocks until a program seems to work is an inefficient debugging strategy. First describe the expected behaviour, reproduce the failure and identify which variable, condition, event or loop differs from expectation. Then change one relevant part and run the test again.
Imagine a fictional Scratch game where points increase twice for each collected item. The student should inspect whether two scripts both update the score or whether a continuous collision repeatedly triggers the increment. A targeted fix is stronger than deleting unrelated blocks. Record the original cause and test another contact situation to confirm the error is resolved.
42. Use a small test matrix instead of one successful run
A program can work for the most common input while failing at boundaries. Test at least a normal condition, a boundary and an unexpected but relevant state. For a timer, try a typical countdown, a starting value of one and the end condition. For a score threshold, test just below, equal to and above the required value.
Spreadsheets benefit from the same approach: test IF criteria at both sides of a cut-off and lookups with existing and missing identifiers. Media artefacts should be checked at their actual presentation size and in the exported format. One clear test matrix can expose more problems than repeating the identical successful example.
43. File names and saving locations are part of the output
The official lab duration includes time to save required softcopies. A completed PowerPoint, Excel workbook or Scratch project is useful for marking only when the candidate saves the required artefact as instructed. Students should learn to distinguish opening a file, saving its current state and exporting another format. The exact folder, naming convention and submission process must come from the authorised examination instructions.
During home practice, use an invented task that asks for three named files. Save the workbook, presentation and Scratch project, close the applications, then reopen the files to verify their contents. This confirms that they exist and are usable, not merely visible on the screen before the session ended.
44. Exporting is not the same as saving an editable project
A raster image exported from vector work may display correctly but not retain the editable shapes of the original. A rendered video may play properly but not allow editing of individual timeline elements. A spreadsheet screenshot may show the right totals but not contain the formulas or validation specified by the task.
Before submitting any practical artefact, check exactly which file types are requested. Where both an editable source and an exported result are required, create both. Do not assume that an image is a substitute for a working spreadsheet or Scratch program. School instructors can provide the authorised file formats for rehearsal.
45. A practice lab should include time for verification
Because Paper 2 is 135 minutes, students need a realistic time plan covering media, spreadsheets, programming and saving. The syllabus provides approximate marks but does not prescribe a fixed minute allocation for each task. A learner may spend too long polishing one slide while leaving a broken formula or incomplete Scratch program.
Use timed rehearsal after techniques are secure. Reserve a final verification phase to open and inspect every required output. If a spreadsheet task is unfinished, identify which feature consumes the time and practise that skill separately. A clock is useful for diagnosing bottlenecks, not as permission to skip required saving.
46. Paper 1 still assesses cybersecurity and computing concepts
K237 Paper 1 examines all six modules, including computer components, networking and impact of computing. A learner who spends the whole term making impressive media products may still struggle with questions about CPU, GPU, storage, LAN versus WAN, MAC versus IP addressing, cloud services, phishing, malware and backups.
Teach these concepts through accurate everyday examples, but do not confuse a consumer technology opinion with a definition. A cloud service can provide remote access while depending on network availability; local storage offers different benefits and risks. Strong Paper 1 answers explain the relevant distinction instead of using marketing slogans.
47. Responsible computing includes copyright and privacy
Students must understand appropriate use of copyrighted material and consider privacy settings and disclosure of personal information. A practical media task should use authorised assets under the instructions, not images copied from an unknown site without permission. A Scratch project should not expose a classmate’s real personal data as sample input.
Use invented names and fabricated datasets for practice. The learner should understand why screenshots of identifiable accounts, passwords or private messages do not belong in a school project without proper authority. Responsible file handling is part of digital literacy and can be taught alongside design quality.
48. An eight-week K237 practical recovery programme
Weeks 1–2: diagnose spreadsheet formulas and reference errors, then practise IF, SUMIF, COUNTIF and formatted data. Weeks 3–4: develop charts, exact lookups, validation and a small self-checking workbook. Change at least one source value and verify that formulas and visuals update.
Weeks 5–6: create a media product from a storyboard and a Scratch game with variables, conditions and loops. Practise debugging and edge cases. Weeks 7–8: attempt a realistic supervised or school-authorised three-task lab rehearsal, with deliberate time reserved for saving and reopening outputs. Maintain Paper 1 concept practice throughout. The schedule is editorial teaching guidance rather than an official exam requirement.
49. Parents should ask to see a working file
A parent need not be a software engineer to ask useful questions. Can the student explain why a formula uses an absolute reference? Does a chart update when a source value changes? Can the Scratch game restart correctly? Can the pupil open the actual saved files without relying on an adult to locate them?
These are observable behaviours and reveal understanding more clearly than a promise that the student spent four hours at a computer. Parents should not create assessed coursework or examinations on the child’s behalf. An effective tutor diagnoses an error, models a repair and lets the pupil complete a different task independently.
50. Frequently asked questions about G2 Computing Paper 2
Is Computing K237 new for 2027? Yes, SEAB marks it New with no previous national subject code. How long is Paper 2? Two hours fifteen minutes. How many marks and what weighting? Eighty marks, 50%. Which three practical tasks? Media, spreadsheets and programming, approximately 30, 25 and 25 marks. Is Paper 1 also compulsory? Yes, 90 minutes and 70 marks, worth the other half.
Is Scratch used? Yes, in the specified G2 programming scope. Is G3 Python identical? No. Are MakeCode microcontrollers assessed in practical Paper 2? The syllabus notes microcontroller outcomes are Paper 1 only. Do spreadsheets require formulas, charts and lookup functions? Yes, within the listed outcomes. Must files be saved for marking? Yes, required softcopies must be saved within the allotted exam time. Can a tutor guarantee a distinction? No. What should the pupil do first? Build one small specified artefact, test edge cases and confirm the required final files exist.
51. Official sources and eduKate next reading
The SEAB 2027 G2 subject directory links the exact official K237 Computing syllabus and Paper 1/Paper 2 specimen papers. The SEAB G2 private-candidate directory confirms the approved software family. The actual school or examination centre controls installed applications, access arrangements and saving instructions.
For the existing subject overview, visit the G2 Computing K237 overview playbook, the official syllabus owner’s detailed exam map, K237 versus G3 K349 for parents, and the immutable Clementi three-person tutorial reference. This article’s distinct purpose is independent practical creation, debugging and reliable file submission.