Cram sheet

SE HSC CRAM SHEET (one page) | Tue 27 Oct 2026, 9.50am Sydney time (10 min reading), 80 marks, online

VERBS identify (name) < outline (sketch main points) < describe (characteristics) < explain (how/why) < discuss/assess/evaluate (both sides + judgement). Link every answer to the SCENARIO. SECURE CIA + Authentication (who) + Authorisation (allowed) + Accountability (logs). Crypto (plain->cipher; symmetric=1 key, asymmetric=pair) + sandbox. Privacy by design = proactive not reactive, embed in design, respect user. SAST = code, not running, early. DAST = running app, no code. Pen test exploits; vuln assessment lists. Validate, sanitise, error-handle. API: excess data exposure. Session mgmt: random ID, timeout, HTTPS. XSS script in page; CSRF forged request; race conditions timing; redirects: allow-list. File attacks, side-channel (timing/power/EM). BCP vs DR (RTO time, RPO data). Privacy Act 1988 / GDPR. WEB Packet = header + payload, reassembled by sequence. IPv4 4x0-255. DNS name->IP. Ports: HTTP 80, HTTPS 443, FTP 21, SFTP 22, DNS 53, SMTP 25, POP3 110, IMAP 143. TLS cert = identity + encryption; hash = integrity; digital signature = origin. Data mining + metadata + streaming mgmt (CDN, load balance). W3C/WAI, WCAG 4.5:1 contrast. Front end HTML/CSS/JS; back end server+DB; SQL vs NoSQL. ORM = objects; SQL = control. Load time: minify, lazy load, cache, CDN, compress images, async/defer. PWA = service worker + manifest + HTTPS + offline. AUTOMATION ML subset of AI. Supervised=labels, unsupervised=none, semi=few, reinforcement=rewards. Linear=number, polynomial=curve, logistic=yes/no CLASSIFY, KNN=nearest K. NN: input/hidden/output, weights, activation, forward prop, loss, backprop, epochs. MLOps: design, model dev, operations. DevOps: feedback, automation, shared ownership. RPA tasks; BPA whole process. Bias: human + dataset + labelling; fix = diverse data, audits, oversight. PROJECT Direct (instant), phased (stages), parallel (both run, safest), pilot (small group). Waterfall sequential; Agile iterative sprints; WAgile hybrid. Gantt = tasks/time/dependencies; logbook = date, progress, issues. Test: unit, integration, system, acceptance, beta, volume; boundary, path, faulty data. Maintenance: corrective, adaptive, perfective. Stub (fake low module), driver (fake caller). Errors: syntax, logic, runtime. Debug: breakpoint, step, watch, print. YEAR 11 Binary/hex; two's complement = invert + 1. OOP: class, object, encapsulation, abstraction, inheritance, polymorphism. Closed loop = sensor feedback. Sensor in, actuator out, end effector grips. CODE input() is a string. s[:3]. random.randint(a,b). range(1,6)=1..5. with open(f) as x. try/except. class B(A): super().__init__(). sqlite: cur.execute("... WHERE a=?", (v,)). SQL SELECT f, COUNT(x) AS n FROM t WHERE c = 'v' GROUP BY f ORDER BY n DESC. Quote text. PSEUDO CAPS keywords, END pairs, indent, FOR..NEXT, WHILE..ENDWHILE, REPEAT..UNTIL, CASEWHERE..END CASE. DFD circle = process, store, external entity, labelled curved arrows. Level 0 = 1 process, no stores. Structure chart: empty circle data, filled flag, diamond decision, curved arrow loop. PACE Q1-14 objective (16 marks) then ~1.75 min/mark. Use scenario words. Write something for every question. EXTENDED (5-6 marks) Thesis using question words -> define concept -> 2-3 points (point + mechanism + SCENARIO evidence) -> positive AND negative -> judgement. Refer to the stimulus by name. 150-200 words. 2025 SHAPE Q1-14 objective (Pilot, Waterfall/Agile, supervised, authentication, race conditions, ORM v SQL, compliance, DevOps, logistic regression, SAST/DAST). Then data dictionary, debugging, PWA compare, load time, privacy by design, pseudocode, Python password, SQL, class diagram (inheritance), food-delivery website, safety AI, neural network VPA, Python 5x5 grid, human bias video (6 marks). WATCH OUT (school notes errors) SFTP is port 22 not 443. KNN is normally SUPERVISED, not unsupervised. Call your Python functions. Quote text in SQL. Exam wording beats notes wording. LAST 2 MIN Every question has something written. Re-read multi-select counts. Check units, > vs >=, and off-by-one in range().