Software Engineering › Notes › Exam day facts, key verbs and technique
Exam day facts, key verbs and technique
The exam in 10 facts (verified: NESA 2026 timetable + Quick Reference Guide)
- Date: Tuesday 27 October 2026. Software Engineering is the 9.50 am – 12.20 pm paper (9.50 is when reading time begins; check your personal timetable in Students Online).
- Length: 2 hours 30 minutes in total, including 10 minutes reading time. That leaves 2 h 20 min of writing for 80 marks, about 1.75 minutes per mark.
- Online: The exam is sat on a computer with wired headphones (NESA Quick Reference Guide; needed for video/audio stimulus, e.g. the 2025 Q27 human-bias video).
- Question types: Multiple choice, multi-select, drag-into-order, fill-in-the-blanks (dropdowns), drawing questions, pseudocode, Python, SQL, short answers and extended responses.
- Pace: In 2025, Q1–Q14 (16 marks) were objective items and Q15–Q27 (64 marks) were written, coded, drawn or extended. Do not spend more than about 1.75 min per mark; flag and return.
- Stimulus: Most 3–9 mark questions come with a scenario. Marking guidelines reward answers that use the scenario (the 2025 marker feedback says this again and again).
- Tools: The online exam has an HSC drawing tool and a pseudocode editor (named in the NESA/school exam-tips materials). Practise with the NESA familiarisation exam (https://fam.hsconline.nesa.nsw.edu.au/). UNVERIFIED: exactly how Python and SQL answers are entered or whether they can be run; check the familiarisation site before the exam.
- Code answers: 2025 marker feedback (code): call your functions in the main program after defining them, use correct indentation, put quotation marks around text in SQL.
- Syllabus: 2022 syllabus, first HSC in 2025. 2026 is the second HSC. Year 11 and Year 12 content are both examinable (the 2025 paper was drawn from Year 12 topics: SSA, web, automation, project).
- 2025 shape: Q1–14 objective (implementation methods, dev approaches, ML training, optimisation, API vulnerability, web protocols, CSS, user action controls, ORM v SQL, regulatory compliance, DevOps, logistic regression, SAST v DAST, data structures). Then 3-mark items (data dictionary, debugging tools, PWA v interactive site, load times, privacy by design, pseudocode, metadata/data mining, Python, session management), a 4-mark SQL, a 3-mark class diagram, Q23 food-delivery website (3+3+3), Q24 safety monitoring AI (5), Q25 neural network assistant (5), Q26 Python grid (5), Q27 human bias video (6).
NESA key verbs: what each one asks for
- Identify: Recognise and name. One word or phrase is enough (1 mark).
- Outline: Sketch in general terms; indicate the main features. A short statement per point, no deep detail.
- Describe: Provide characteristics and features. Say what it is and what it does, with a detail or example.
- Distinguish: Note the differences between things. Name both and say what is different.
- Compare: Show how things are similar AND different. Use linking words: both, whereas, however, in contrast.
- Explain: Relate cause and effect; make the relationship evident; give the why and/or how. Use because / therefore / which means.
- Discuss: Identify issues and give points for and/or against. Give both positives and negatives (2025 Q24 and Q25: both a positive AND a negative impact were needed for the top marks).
- Analyse: Identify components and the relationship between them; draw out and relate implications.
- Assess: Make a judgement of value, quality, outcomes, results or size. State a judgement, then support it.
- Evaluate: Make a judgement based on criteria; determine the value of something. Use criteria words (effectiveness, cost, security) and finish with a verdict.
- Justify: Support an argument or conclusion. Give the decision plus the reason.
- Construct / Design: Make or build a model, diagram, query or program (e.g. 'Construct a SQL query', 'Construct a class diagram').
Short-answer technique that matches the marking guidelines
- Count the marks: One clear point per mark. A 3-mark 'describe' needs about 3 developed points or 2 points with a scenario example. Do not write an essay for 3 marks.
- Use the stimulus: Pull a named detail out of the scenario (e.g. 'live location', 'weather warnings', 'cameras and wearables'). MG ladders give the top mark only when the answer is linked to the scenario.
- Terminology: Use syllabus words exactly (privacy by design, session management, metadata, weightings, hash). The marking feedback says 'communicate using appropriate course specific terms'.
- Answer the verb: 'Outline TWO ways' means exactly two, each with a short reason. 'Explain how X could' means give the mechanism, not just the name. 2025 Q16(b): markers said students listed ways to INVESTIGATE load times instead of ways to MINIMISE them.
- Pair it: When both a positive AND a negative are asked (2025 Q24/Q25), label them. When a comparison is asked, include one similarity and one difference.
- Code: Check your code runs: indentation, quotes, call the function after defining it, print or display the required output.
How the 2025 marking guidelines ladder a typical 3-mark and 5-mark item
- 3 marks: 3 = explains/describes the whole requirement with a link to the scenario. 2 = outlines features. 1 = identifies a feature.
- 5 marks (2025 Q25): 5 = explains how the neural network supports the virtual personal assistant AND includes one positive AND one negative impact. 4 = explains the network OR outlines the network plus a positive and a negative impact. 3 = outlines features of the network and/or impacts. 2 = outlines one feature or one impact. 1 = some relevant information.
- 6 marks (2025 Q27): 6 = EXPLAINS how human bias affects the usefulness of AI solutions and supports it with the video. 5 = DESCRIBES it with scenarios from the video. 4 = OUTLINES it and refers to the video. 2–3 = identifies sources/examples of bias, effects of bias, or outlines how AI learns. 1 = some relevant information.
- Code 5 marks: Correct program meeting ALL listed requirements (e.g. 2025 Q26: unique random numbers, processing the list, selection and loops, formatted output, built-in functions). Each missing requirement costs about one mark.