Loading…

Worked solutions

IB Computer Science SL · Paper 2

Computer Science, Standard Level

Session: November 2025
Marks: 180
Duration: 1 hour
Calculator: Not allowed
Questions: 12

Keep your copy of the paper open alongside. We don't reproduce the questions: each part below is labelled to match the November 2025 Global paper, which your school or the IB can provide.

Practice mode

Hide our working until you've had a go.

Option A: Databases · Option

Questions 1–3 · 45 marks

Question 1

15 marks

Relational Modeling, Electronic Health Records, Transaction States, and Crash Recovery

Databases

(a)

3 marks

Identify primary entities: `Clinic`, `Dentist`, and `Nurse`.
Define multiplicities with Clinic: A single clinic employs many dentists (1-to-many relationship from `Clinic` to `Dentist`), and the same clinic employs many nurses (1-to-many relationship from `Clinic` to `Nurse`).
Define the relationship between Dentist and Nurse: Because a dentist collaborates with multiple nurses and a nurse assists multiple dentists, their direct association is many-to-many (M:N). In a normalized relational implementation, this M:N relationship is resolved through an associative link table (such as `DutyAssignment` or `Appointment`) having two 1-to-many relationships pointing from `Dentist` and `Nurse`.

Mentor tip

Explicitly draw or state the entities with crow's foot notation or an associative intersection table to resolve many-to-many.

(b)

6 marks

Advantage 1 (Concurrent Multi-User Access and Mobility): Storing patient records in a centralized relational database allows dentists, dental hygienists, and reception staff to retrieve and update patient medical charts simultaneously across multiple clinical terminals. Unlike physical paper files that can only exist in one room at a time, digital records facilitate seamless cross-departmental coordination.
Advantage 2 (Automated Data Validation, Integrity, and Redundancy Prevention): Database management systems enforce domain constraints, required foreign keys, and standardized picklists during input. This eliminates handwriting ambiguity, prevents contradictory records, and drastically reduces clerical errors compared to paper charting.
Advantage 3 (Disaster Recovery and Longitudinal Data Mining): Electronic databases support automated off-site backups and replication, safeguarding patient histories against physical perils like fire or water damage. Furthermore, digitized records enable structured SQL analytics to monitor patient treatment progressions and track clinic-wide outcomes.

Mentor tip

Provide two comprehensive explanations, detailing the operational mechanism and its clinical benefit.

(c)

3 marks

Step 1 (Active State): The transaction begins execution, performing read and write operations in main memory buffers as patient billing data is updated.
Step 2 (Partially Committed State): Once the final database statement executes, the transaction enters the partially committed state, where updates reside in volatile memory awaiting physical persistence.
Step 3 (Committed or Aborted State): If all write operations and integrity constraints flush successfully to non-volatile disk storage, the transaction transitions to the Committed state. If a hardware fault or constraint violation occurs before completion, it transitions to the Failed state and is rolled back to the Aborted state, restoring data consistency.

Mentor tip

Trace the lifecycle: Active -> Partially Committed -> Committed (or Failed -> Aborted).

(d)

3 marks

Step 1 (Log Scan and Checkpoint Identification): Following a system failure, the recovery manager scans backward through the write-ahead transaction log to identify the most recent consistent checkpoint record.
Step 2 (Redo Phase): Scanning forward from the checkpoint, the recovery manager re-executes all operations for transactions that logged a successful `COMMIT` record prior to the failure, ensuring the ACID durability property.
Step 3 (Undo Phase): The recovery manager identifies all active transactions that were still uncommitted at the moment of the crash and reverses their operations in reverse chronological order using logged 'before-values', restoring full atomicity.

Mentor tip

Highlight the checkpoint reference, the REDO step for committed operations, and the UNDO step for uncommitted transactions.

Question 2

15 marks

Composite Primary Keys, Second Normal Form, SQL Filtering, and 3NF Relational Schema

Databases

(a)

2 marks

A patient can visit the clinic multiple times across different days, or receive multiple procedures from the same nurse on separate visits.
Without `DateTime`, the composite key `(PatientID, NurseID)` would prohibit a patient from ever being treated by the same nurse more than once due to primary key collisions. Incorporating `DateTime` ensures that every unique testing event remains uniquely identifiable.

Mentor tip

Point out that DateTime guarantees uniqueness across recurring appointments between the same patient and nurse.

(b)

2 marks

Second Normal Form (2NF) mandates that the table must already be in 1NF and that every non-key attribute must be fully functionally dependent on the complete primary key.
In the `TEST` table, `PatientName` is functionally dependent purely on `PatientID` alone rather than the composite key `(PatientID, NurseID, DateTime)`. This partial dependency violates 2NF.

Mentor tip

State the definition of 2NF and identify the partial functional dependency PatientName -> PatientID.

(c)

2 marks

Scan the `TEST` table for records matching `TestType = 'Oral X-ray'` and `NurseID = 'BM2385'`.
Matching Row 1: PatientID 5, NurseID BM2385, TestType Oral X-ray -> Simon Moore.
Matching Row 3: PatientID 78, NurseID BM2385, TestType Oral X-ray -> Scottie Ngatai.
Final projected output contains two patient names:
Simon Moore
Scottie Ngatai

Mentor tip

Verify both criteria against each record; ensure only the requested PatientName attribute is listed.

(d)

4 marks

Construct an SQL `SELECT` statement joining `TEST` and `NURSE` on `NurseID`.
Filter the nurse's full name to 'Anya Svetlana'.
Restrict `DateTime` to cover the entire day of 19 November 2024 using pattern matching or range comparison.
SQL Implementation:
SELECT TEST.PatientName FROM TEST INNER JOIN NURSE ON TEST.NurseID = NURSE.NurseID WHERE NURSE.NurseName = 'Anya Svetlana' AND TEST.DateTime LIKE '2024/11/19%';

SELECT TEST.PatientName FROM TEST INNER JOIN NURSE ON TEST.NurseID = NURSE.NurseID WHERE NURSE.NurseName = ’Anya Svetlana’ AND TEST.DateTime LIKE ’2024/11/19%’;

Mentor tip

Use an explicit INNER JOIN on the shared NurseID foreign key and filter date using LIKE '2024/11/19%'.

(e)

5 marks

Decompose the unnormalized appointment data into Third Normal Form (3NF) to eliminate partial and transitive dependencies.
Table 1: `PATIENT (PatientID, PatientName)` — isolates patient biographical information.
Table 2: `DENTIST (DentistID, DentistName, Specialties)` — removes dentist transitive dependencies.
Table 3: `TREATMENT (TreatmentID, TreatmentName, Cost)` — isolates treatment pricing from individual appointments.
Table 4: `APPOINTMENT (AppointmentID, PatientID, DentistID, TreatmentID, VisitDateTime)` — resolves appointment events with primary key `AppointmentID` and foreign keys referencing `PATIENT`, `DENTIST`, and `TREATMENT`.

PATIENT (\underline{PatientID}, PatientName) DENTIST (\underline{DentistID}, DentistName, Specialties) TREATMENT (\underline{TreatmentID}, TreatmentName, Cost) APPOINTMENT (\underline{AppointmentID}, PatientID^{*}, DentistID^{*}, TreatmentID^{*}, VisitDateTime)

Mentor tip

Ensure 3NF by isolating Dentist and Treatment entities into separate relations so non-key attributes depend only on the primary key.

Question 3

15 marks

Data Integrity Maintenance, Integrated Databases, Visual Querying, and Clinical Mining Ethics

Databases

(a)

2 marks

Strategy (Referential Integrity Constraints and Automated Backups): The database administrator configures database foreign key constraints (`ON DELETE RESTRICT` or `CASCADE`) and domain validation checks to prevent orphaned records or corrupt data entry.
Complementing this, the DBA schedules automated incremental off-site replication and periodic transaction log snapshots to guarantee rapid data restoration without corruption during unexpected hardware faults.

Mentor tip

Mention enforceable constraint rules (domain/referential integrity) or regularized automated replication.

(b)

2 marks

An integrated database system consolidates previously fragmented operational repositories (such as patient clinical records, billing, appointment scheduling, and inventory) into a unified, shared database structure.
This single source of truth eliminates redundant data entry across clinic departments, ensuring that any modification updates immediately across all clinic modules.

Mentor tip

Emphasize centralized unified data storage and elimination of departmental silos.

(c)

3 marks

Visual query tools (Query By Example - QBE) present the user with a graphical interface showing table names, field picklists, and criteria grids instead of requiring manual SQL coding.
The clinical user drags required fields (such as `PatientName` and `Treatment`) into the visual design grid and enters filtering conditions into designated constraint boxes (e.g. typing '> 200' under Cost or selecting 'Orthodontics' from a dropdown menu).
The underlying database software automatically converts these visual selections into a structured SQL query, executes it against the server, and formats the output into an interactive data view.

Mentor tip

Explain GUI table representations, setting criteria in visual rows, and automatic translation to executable queries.

(d)

3 marks

Data mining algorithms (such as classification decision trees and cluster analysis) process large volumes of historical patient diagnostic records, geographic demographic data, and seasonal visit frequencies.
The algorithms identify subtle statistical correlations and non-obvious disease patterns (for instance, abnormal spikes in specific oral bacterial infections clustered within a particular district or demographic group).
Healthcare providers use these predictive trend models to anticipate upcoming community outbreaks and deploy preventive health advisories before infections escalate.

Mentor tip

Describe pattern discovery across patient datasets and using predictive models for early public health intervention.

(e)

5 marks

Advantages for the Clinic: Direct access enables instant electronic verification of treatments, accelerating insurance claim pre-authorizations and cash flow reimbursements. This substantially reduces clerical administrative overhead, phone inquiries, and disputed claim delays.
Disadvantages for the Clinic: Exposing clinical records creates severe patient confidentiality and data security risks. Accidental leakage of protected health information could trigger substantial legal liabilities under data protection regulations (such as GDPR or HIPAA). Furthermore, third-party direct access creates potential security attack vectors if external credentials are compromised.
Evaluation & Conclusion: The clinic should not allow direct unrestricted database access. Instead, it should deploy a strictly scoped, read-only API gateway with field-level encryption, exposing only anonymized billing codes and procedure dates necessary for insurance claims.

Mentor tip

Provide a balanced appraisal covering administrative speed vs patient confidentiality, supported by a concluding synthesis.

Option B: Modelling and simulation · Option

Questions 4–6 · 45 marks

Question 4

17 marks

Permutation Cycle Strategy, Spreadsheet Limitations, and Monte Carlo Multi-Agent Simulation

Modelling and simulation

(a)(i)

1 mark

A single child opens 5 boxes randomly out of 10 total boxes without replacement.
Probability = 5 / 10 = 1 / 2 = 0.5 (or 50%).

P (Child 0 succeeds) = \frac{5}{10} = 0.5

Mentor tip

Express as a fraction (1/2), decimal (0.5), or percentage (50%).

(a)(ii)

1 mark

Under independent random box selection, each of the 10 children has a 1/2 chance of finding their number.
Joint probability = (1/2)^10 = 1 / 1024 ≈ 0.000977 (or ≈ 0.098%).

P (All 10 succeed) = (\frac{1}{2})^{10} = \frac{1}{1024} \approx 0.000977

Mentor tip

Multiply independent probabilities: (0.5)^10 = 1/1024.

(b)

1 mark

Follow the cycle sequence starting at box 3:
Step 1: Open Box 3 -> contains Paper 1.
Step 2: Open Box 1 -> contains Paper 8.
Step 3: Open Box 8 -> contains Paper 5.
Step 4: Open Box 5 -> contains Paper 6.
Step 5: Open Box 6 -> contains Paper 4.
Sequence of boxes opened: 3 -> 1 -> 8 -> 5 -> 6 (child does not locate number 3).

3 \to 1 \to 8 \to 5 \to 6

Mentor tip

Trace box indices: Box 3 has 1, Box 1 has 8, Box 8 has 5, Box 5 has 6.

(c)

4 marks

Reason 1 (Dynamic Conditional Pointer Traversal): Spreadsheets rely on static directed dependency formulas between cells. The follow-the-number strategy requires iterative pointer chasing where the next cell lookup dynamically depends on the integer returned by the previous step. Implementing variable-length loops that jump dynamically to `BOX[LOOK]` for up to 5 iterations requires complex circular formulas or external scripting.
Reason 2 (Shuffling Permutations without Replacement across Large Iterations): Simulating thousands of trials requires shuffling an array of 10 unique elements without replacement before each run. Generating distinct random permutations across tens of thousands of rows is unwieldy in standard spreadsheets compared to high-level procedural programming languages.

Mentor tip

Detail the challenge of dynamic pointer traversal (looping/state dependency) and generating random non-repeating permutations.

(d)

6 marks

Initialize the `CHILDREN` array of size 10 to `false`.
Shuffle the `BOX` array of 10 integers.
Iterate through each child `N` from 0 to 9.
For each child, initialize current box lookup `LOOK = N` and test up to 5 box attempts.
If `BOX[LOOK] == N`, mark `CHILDREN[N] = true` and exit inner loop; otherwise update `LOOK = BOX[LOOK]`.
Finally, loop from 0 to 9 and output the boolean state of `CHILDREN[N]`.

random.shuffle (BOX) loop N from 0 to 9 CHILDREN [N] = false LOOK = N loop ATTEMPT from 1 to 5 if BOX [LOOK] = N then CHILDREN [N] = true break else LOOK = BOX [LOOK] end if end loop end loop loop N from 0 to 9 output CHILDREN [N] end loop

Mentor tip

Set current pointer LOOK = N and update LOOK = BOX[LOOK] on each unmatching attempt.

(e)

4 marks

Initialize `WIN_COUNT = 0`.
Enclose the trial simulation inside a loop running 10000 iterations.
In each trial, execute the follow-the-number strategy across all 10 children.
Evaluate if all 10 elements in `CHILDREN` equal `true`. If so, increment `WIN_COUNT`.
After completing 10000 trials, calculate `PERCENTAGE = (WIN_COUNT / 10000) * 100` and display the result.

WIN_COUNT = 0 loop TRIAL from 1 to 10000 random.shuffle (BOX) ALL_WON = true loop N from 0 to 9 LOOK = N FOUND = false loop A from 1 to 5 if BOX [LOOK] = N then FOUND = true break else LOOK = BOX [LOOK] end if end loop if not FOUND then ALL_WON = false end loop if ALL_WON then WIN_COUNT = WIN_COUNT + 1 end loop PERCENT = (WIN_COUNT /10000) * 100 output PERCENT + "%"

Mentor tip

Verify all 10 children succeed within the same trial before incrementing the prize counter.

Question 5

16 marks

Autonomous Vehicle Modeling, Simulation Variables, Test Scenarios, and Real-World Limitations

Modelling and simulation

(a)

1 mark

Identify one system suitable for computer modeling: Urban traffic flow (or vehicle braking dynamics, climate systems, suspension mechanics, or aerodynamic airflow).

Mentor tip

State a well-defined physical or computational system.

(b)

1 mark

Identify one traffic condition variable: Traffic density (vehicles per kilometer), average traffic speed, intersection queue length, or pedestrian crosswalk frequency.

Mentor tip

Provide a measurable variable describing traffic state.

(c)

4 marks

Grouping 1 (Environmental and Atmospheric Conditions): Variables capturing external surroundings, such as precipitation level (mm/h), ambient road surface friction coefficient, fog visibility distance, and illumination level (lux).
Grouping 2 (Vehicle Sensor and Perception Telemetry): Variables measuring sensor hardware inputs, such as LiDAR point cloud reflection density, radar Doppler velocity signals, optical camera frame rate, and ultrasonic distance readings.

Mentor tip

Outline two distinct categories with relevant representative variables.

(d)

4 marks

Similarities: Both mathematical models and computer simulations use quantitative data and mathematical equations to predict vehicle dynamics, allowing engineers to test behaviors without endangering real passengers.
Differences (Theory vs Virtual Realization): A mathematical model represents abstract analytical equations (e.g. differential equations governing friction and braking distance). In contrast, a simulation applies these mathematical foundations inside a dynamic time-stepped virtual world, running iterative edge-case scenarios with multiple interacting agents.

Mentor tip

Contrast the theoretical mathematical foundation with the empirical, dynamic time-stepped simulation environment.

(e)

2 marks

Test-case scenario (LiDAR Occlusion at Highway Speed): The simulation systematically disables the forward LiDAR sensor or degrades its point cloud with simulated heavy snowfall while cruising at 100 km/h.
Evaluation metric: Test whether the autonomous vehicle immediately detects sensor failure, successfully switches to radar and stereoscopic vision redundancy, alerts the passenger, and smoothly navigates to a safe stop.

Mentor tip

Describe a specific failure condition and the vehicle's required automated fail-safe response.

(f)

4 marks

Identification: The reality gap and simulator overfitting.
Expansion: Real-world urban driving contains chaotic, unpredictable occurrences—such as jaywalking pedestrians, unusual road construction markers, and extreme weather reflections—that cannot be comprehensively anticipated in a synthetic virtual environment.
Consequence: Autonomous vehicle algorithms may become over-optimized for simulator physics, giving developers false confidence. When deployed onto actual roadways, unmodeled corner cases can cause sensor confusion, erratic braking, or fatal collisions.

Mentor tip

Structure your response around the reality gap, the difficulty of replicating chaotic edge cases, and safety risks.

Question 6

12 marks

Monte Carlo Pi Estimation via Spreadsheets and 3D Bounding Box Real-Time Computation

Modelling and simulation

(a)

2 marks

The Monte Carlo method for estimating π is based on a two-dimensional geometric proof: the ratio of the area of a circle quadrant to the enclosing unit square.
Because points are randomly placed using paired Cartesian coordinates (x, y), a 2D scatter plot directly represents the spatial distribution of samples and visualizes the circular arc boundary intuitively.

Mentor tip

Explain that the geometric area ratio is inherently two-dimensional, making 2D scatter plots an ideal visual match.

(b)(i)

1 mark

Apply the Pythagorean theorem in cell C2: `=SQRT(ColA2^2 + B2^2)`

C2: = SQRT (ColA2 ˆ 2 + B2 ˆ 2)

Mentor tip

Use the standard spreadsheet SQRT function with squared coordinates.

(b)(ii)

1 mark

Classify distance less than or equal to 1 as inside the circle: `=IF(C2<=1, "In", "Out")`

D2: = IF (C2 <= 1, "In", "Out")

Mentor tip

Use an IF conditional testing if distance <= 1.

(c)

3 marks

Count the number of points falling inside the circular quadrant across the 5000 generated coordinates (rows 2 to 5001): `COUNTIF(D2:D5001, "In")`.
Divide the count by the total sample size (5000) to find the fractional area of the quadrant.
Multiply by 4 because the area of the unit quadrant equals π / 4.
Spreadsheet Formula: `=COUNTIF(D2:D5001, "In") / 5000 * 4` (or `=COUNTIF(D2:D5001, "In") / COUNTA(D2:D5001) * 4`).

E2: = COUNTIF (D2:D5001, "In") /5000 * 4

Mentor tip

Multiply the ratio of inside points to total points by 4.

(d)

5 marks

Processing Power Demands: Calculating 3D oriented bounding boxes in real time requires processing millions of raw LiDAR 3D point cloud coordinates and multi-camera video feeds per second. Convolutional neural networks and point-cloud clustering algorithms must execute at high frame rates (e.g. 30–60 Hz) with sub-10ms latencies to avoid collision hazards.
Memory Constraints: Storing high-density voxel grids, intermediate tensor feature maps, and multi-sensor temporal tracking buffers demands substantial high-bandwidth memory (HBM/VRAM) to prevent processing bottlenecks.
Evaluation & Synthesis: Standard automotive microcontrollers cannot handle these workloads. Autonomous vehicles require dedicated hardware accelerators (such as multi-core Tensor Processing Units or GPUs) paired with low-latency bus architectures to ensure real-time safety.

Mentor tip

Evaluate both processing throughput (GPU point-cloud clustering) and high-bandwidth memory storage, concluding with dedicated hardware necessity.

Option C: Web science · Option

Questions 7–9 · 45 marks

Question 7

15 marks

Web Protocols, Architecture, CGI Processing, and Server-Side PHP Execution

Web science

(a)(i)

1 mark

State the protocol in the URL `http://hardware4hotels.example.com`: `http` (Hypertext Transfer Protocol).

Mentor tip

State 'http' or 'Hypertext Transfer Protocol'.

(a)(ii)

1 mark

State the domain name in the URL: `hardware4hotels.example.com`.

Mentor tip

Omit the protocol prefix; state 'hardware4hotels.example.com'.

(b)

2 marks

Protocols establish standardized communication rules, data packet formats, and error-handling conventions across diverse hardware and operating systems.
This ensures interoperability between different web browsers and web servers, guaranteeing that requested web documents are transmitted reliably and rendered identically.

Mentor tip

Highlight standardization, interoperability, and reliable client-server communication.

(c)

2 marks

Identify two components of a webpage: HTML structural markup and CSS stylesheets (or client-side JavaScript, hyperlinks, navigation bar, multimedia elements).

Mentor tip

Name two structural or interactive elements.

(d)

2 marks

The Common Gateway Interface (CGI) serves as a standardized middleware gateway connecting the web server to backend executable programs and scripts.
When a client submits the feedback form, CGI relays the HTTP POST parameters to the backend script and forwards the dynamically generated response back to the client browser.

Mentor tip

Explain CGI as a gateway passing form data between the web server and backend scripts.

(e)(i)

4 marks

1. Parameter Extraction: The server captures the submitted `work_category` and `rating` values from the HTTP POST payload via `$_POST`.
2. Database Connection & Query: The script establishes a connection to the `sales` database and queries the existing average rating for that work category using `SELECT AVG(rating)`.
3. Record Insertion: The script inserts the newly submitted rating into the `ratings` table using an SQL `INSERT` statement.
4. Conditional Output & Cleanup: If the submitted rating is greater than or equal to the category average, it outputs a positive acknowledgment; otherwise, it outputs an apology. Finally, it closes the database connection.

Mentor tip

Detail four sequential steps: extraction, average calculation query, record insertion, and conditional echo output.

(e)(ii)

3 marks

Security and Credential Protection: Executing database scripts server-side prevents database credentials (` $u ser nam e ‘, ‘$ password`) from being exposed to the client browser, mitigating direct SQL injection or credential theft.
Furthermore, browsers cannot connect directly to secure SQL databases without exposing database ports publicly over the internet. Server-side processing maintains centralized access control and data integrity.

Mentor tip

Emphasize security, hiding database credentials, and preventing unauthorized direct database access.

Question 8

17 marks

Search Engine Results Estimation, PageRank, SEO Techniques, and Web Evolution

Web science

(a)

2 marks

The reported figure of 248,000 results is an algorithmic statistical estimate generated by the search engine to provide rapid response times, not an exact count of dedicated pages.
The index retrieves all pages matching any combination of the individual keywords ('one', 'seven', 'academy') across billions of web pages rather than exclusively retrieving pages solely focused on the school.

Mentor tip

Explain that the count is a rapid statistical estimate matching broad individual keywords across the index.

(b)

3 marks

Principle of Inbound Link Authority (Random Surfer Model): PageRank treats each inbound hyperlink to a webpage as a vote of recommendation.
A page achieves a higher PageRank if it receives inbound links from reputable, high-ranking pages rather than low-authority sites.
The link equity contributed by a pointing page is divided proportionally by the total number of outbound links on that page, damping vote dilution.

P R (A) = (1 - d) + d i = 1 \sum n \frac{P R ( T _{i} )}{C ( T _{i} )}

Mentor tip

Explain incoming links as votes of authority and note that links from high-ranking pages contribute greater equity.

(c)

2 marks

Way 1: Securing authoritative external backlinks from verified educational bodies and municipal school boards.
Way 2: Improving website technical performance, including mobile responsiveness and fast Core Web Vitals load times.

Mentor tip

State two legitimate ranking factors such as authoritative backlinks and mobile page speed optimization.

(d)

3 marks

Meta tags are placed in the `<head>` block of an HTML document, containing structured metadata (such as page description and keywords) that is not rendered visually to human users.
When a web crawler visits the page, it parses the HTML DOM, extracts the meta content, and indexes these descriptive tokens in its database to categorize the topic.
Search engines often use the meta description as the snippet displayed directly on search engine results pages.

Mentor tip

Describe placement in HTML head, parsing by web crawlers, and storage in the indexing catalog.

(e)

5 marks

White Hat SEO: Adheres strictly to search engine guidelines by providing high-quality original curriculum content, semantic HTML tags, fast loading speeds, and organic inbound links. It provides stable, sustained long-term visibility without penalties.
Black Hat SEO: Employs deceptive practices such as keyword stuffing, invisible text, cloaking, and private blog link networks. While it may provide short-term traffic spikes, modern search algorithms quickly detect manipulative patterns.
Consequences & Conclusion: Black hat tactics carry severe risks of search engine manual actions, algorithmic penalties, or complete domain de-indexing. For an educational institution like One2Seven Academy, preserving institutional credibility is paramount; the webmaster must strictly employ white hat SEO.

Mentor tip

Compare white hat vs black hat techniques, discussing sustainability vs penalties, and conclude decisively.

(f)

2 marks

Challenge 1 (Exponential Growth of Dynamic and AI-Generated Content): Crawling budgets are strained by billions of ephemeral JavaScript-rendered pages and synthetic content.
Challenge 2 (Combatting Disinformation and Misleading Content): Search engines must refine relevance algorithms to distinguish authoritative facts from automated web spam while maintaining user privacy.

Mentor tip

Identify challenges like crawling dynamic content at scale and filtering automated spam/disinformation.

Question 9

13 marks

Mobile vs Ubiquitous Computing, Web Decentralization, and Cloud Infrastructure Models

Web science

(a)

2 marks

Mobile Computing involves portable, battery-powered devices (such as smartphones and laptops) that users actively carry and consciously operate to access network services.
Ubiquitous Computing (pervasive computing) embeds computing intelligence invisibly into physical environments and everyday objects (e.g. smart sensors, connected lighting), operating autonomously with minimal active user engagement.

Mentor tip

Differentiate portable, user-operated mobile devices from embedded, invisible ubiquitous environments.

(b)

3 marks

Reason: Hardware Miniaturization and High-Speed Wireless Infrastructure.
Expansion: Advances in semiconductor fabrication have enabled powerful multi-core processors, low-power displays, and compact lithium batteries.
Consequence: Paired with ubiquitous 4G/5G mobile networks and public Wi-Fi, users can access high-bandwidth cloud services, streaming media, and enterprise software anywhere at affordable prices.

Mentor tip

Explain hardware miniaturization paired with high-speed cellular networks (4G/5G).

(c)

3 marks

Distributed networks and cloud platforms replace single centralized server architectures with thousands of geographically dispersed edge nodes and data centers.
Content Delivery Networks (CDNs) cache content and process microservices closer to end users, preventing central points of bottleneck or failure.
This democratizes web hosting and decentralizes data routing, enhancing system resilience against outages and reducing regional latency.

Mentor tip

Discuss moving from monolithic central servers to geographically dispersed cloud nodes and CDNs.

(d)

5 marks

Private Cloud: Provides dedicated on-premise or single-tenant infrastructure. Offers full control over security configurations, customizable data privacy governance, and predictable performance. However, it requires substantial upfront capital expenditure and ongoing IT administrative overhead.
Public Cloud: Multi-tenant infrastructure managed by external providers (e.g. AWS, Azure). Offers near-infinite scalability, pay-as-you-go pricing, and zero hardware maintenance. However, it introduces potential security and regulatory concerns because physical hardware is shared with other tenants.
Evaluation & Conclusion: A hybrid cloud model is often optimal for modern enterprises: keeping sensitive customer and proprietary financial data inside a secure private cloud while leveraging the elastic scalability of the public cloud for web applications.

Mentor tip

Evaluate security, cost, scalability, and maintenance for both models, recommending a hybrid approach.

Option D: Object-oriented programming · Option

Questions 10–12 · 45 marks

Question 10

15 marks

Object Modeling, Encapsulation, Array Searching, and Delayed Case Evaluation

Object-oriented programming

(a)

1 mark

A constructor is a special class method invoked during object instantiation that allocates memory and initializes instance variables with starting values.

Mentor tip

Define constructor as an initialization method invoked upon instantiation.

(b)

2 marks

The method inspects the `judgmentGiven` character parameter.
If `judgmentGiven` equals `'F'` (favorable), it updates `caseStatus` to `'won'`; if `judgmentGiven` equals `'A'` (against), it updates `caseStatus` to `'lost'` (either directly or via `setCaseStatus`).

Mentor tip

Explain testing judgmentGiven and updating caseStatus to 'won' or 'lost'.

(c)

4 marks

Loop through the `allLawyers` array of size 20.
Check that the lawyer element is not null.
Verify that `lawyerType` equals 'civil' (case-insensitively).
Verify that `casesWon` is strictly greater than 30.
Print the matching lawyer's name to the console.

Mentor tip

Check lawyerType.equalsIgnoreCase('civil') and casesWon > 30 before printing.

(d)(i)

1 mark

A return statement terminates method execution and passes a specified result value back to the calling method, returning program control.

Mentor tip

State that return sends a value back and returns control to the caller.

(d)(ii)

2 marks

Method signature requires return type `int`, method name `countDelayedCases`, and an integer parameter for lawyer ID:
`public static int countDelayedCases(int lawyerID)` (or `public int countDelayedCases(int id)`).

Mentor tip

Include int return type, method name countDelayedCases, and int parameter.

(d)(iii)

5 marks

Initialize `count = 0`.
Loop through `allLawyers[]` to locate the lawyer whose ID matches `lawyerID`.
Upon finding the matching lawyer, retrieve their `lawyerCases[]` array.
Loop through each case in `lawyerCases[]`, ensuring non-null references.
Check if `daysFromStart > 45` and `judgmentGiven == 'N'`. If both hold, increment `count`.
Return `count` after loop termination.

public static int countDelayedCases (int lawyerID) {int count = 0; for (int i = 0; i < 20; i + +) {if (allLawyers [i] \neq = null && allLawyers [i] . getLawyerID () == lawyerID) {Case [] cases = allLawyers [i] . getLawyerCases (); for (int j = 0; j < 15; j + +) {if (cases [j] \neq = null && cases [j] . getDaysFromStart () > 45 && cases [j] . getJudgmentGiven () == ’N’) {count + +;}}}} return count;}

Mentor tip

Verify cases[j] != null before checking getDaysFromStart() > 45 and getJudgmentGiven() == 'N'.

Question 11

15 marks

Subclass Inheritance, Aggregation, and UML Class Diagram Modeling

Object-oriented programming

(a)(i)

2 marks

Identify suitable data types for TraineeLawyer attributes:
- joining date: `String` (or `Date` / `int`)
- exam status: `boolean` (or `String` / `char`)

Mentor tip

Specify String/Date for joining date and boolean/String for exam status.

(a)(ii)

4 marks

Construct detailed UML class diagram with three-part compartments (name, attributes, methods):
1. Superclass `Lawyer`:
Attributes: `- lawyerID: int`, `- lawyerName: String`, `- lawyerType: String`, `- lawyerRank: int`, `- casesWon: int`, `- lawyerCases: Case[]`
Methods: `+ Lawyer(...)`, `+ getLawyerName(): String`, `+ getLawyerType(): String`, `+ getLawyerRank(): int`, `+ setLawyerRank(int): void`
2. Subclass `TraineeLawyer`:
Attributes: `- joiningDate: String`, `- examStatus: boolean`
Methods: `+ TraineeLawyer(...)`, `+ getJoiningDate(): String`, `+ getExamStatus(): boolean`
3. Relationship: Draw an inheritance arrow (solid line with open triangular arrowhead) pointing from `TraineeLawyer` up to `Lawyer`.

Mentor tip

Draw three compartments per class and use an open triangular arrow for inheritance.

(b)(i)

1 mark

State the relationship between `TraineeLawyer` and `Lawyer`: Inheritance ('is-a' relationship; `TraineeLawyer` is a subclass of `Lawyer`).

Mentor tip

State inheritance or 'is-a' relationship.

(b)(ii)

1 mark

State the relationship between `Lawyer` and `Case`: Aggregation / Composition ('has-a' relationship; `Lawyer` contains an array of `Case` objects).

Mentor tip

State aggregation or 'has-a' relationship.

(b)(iii)

4 marks

Difference 1 (Inheritance vs Containment): Inheritance ('is-a') creates a specialized class that inherits attributes and behaviors from a superclass, supporting method overriding. Aggregation ('has-a') represents object containment where one class stores references to independent objects without inheriting their class interface.
Difference 2 (Polymorphism and Coupling): Inheritance enables polymorphic substitution (a `TraineeLawyer` instance can be assigned to a `Lawyer` reference). Aggregation does not support polymorphism between container and contained objects; `Case` objects remain distinct entities.

Mentor tip

Contrast 'is-a' specialization and polymorphism against 'has-a' containment and reference lifecycle.

(c)

2 marks

Encapsulation is the object-oriented principle of packaging data attributes and the methods that operate on them within a single class.
Internal instance variables are kept private to protect internal state, exposing controlled public getters and setters for controlled access.

Mentor tip

Explain packaging variables and methods in a single unit and restricting direct access via private fields.

(d)

1 mark

A parent object (or superclass) is a base class in an inheritance hierarchy whose attributes, methods, and behaviors are inherited and extended by child subclasses.

Mentor tip

Define parent class as a superclass providing inheritable members to subclasses.

Question 12

15 marks

Loop Logic Debugging, Multi-Entity Array Traversal, and Open Source Software Impact

Object-oriented programming

(a)(i)

3 marks

1. Faulty Loop Condition: The condition `while (x > 20)` evaluates to false immediately since `x = 0`, so the loop body never executes (should be `x < 20`).
2. Missing Loop Counter Increment: The index variable `x` is never incremented inside the loop body (`x++`), which would cause an infinite loop if entered.
3. Undeclared Variable: The variable `countB` is used inside the loop (`countB++`) without prior declaration or type definition, triggering a compilation error.

Mentor tip

Identify false while condition (x > 20), missing x++ increment, and undeclared countB variable.

(a)(ii)

7 marks

Construct `displayClientNames()` method:
1. Traverse `allLawyers` array of length 20.
2. Check that the lawyer element is not null and `lawyerType` equals 'company'.
3. Retrieve the lawyer's `lawyerCases[]` array.
4. Iterate through the 15 cases, checking for non-null elements.
5. Check if `judgmentGiven == 'A'`. If matched, retrieve `caseClient` and print to the console.

public static void displayClientNames () {for (int i = 0; i < 20; i + +) {if (allLawyers [i] \neq = null && allLawyers [i] . getLawyerType () . equalsIgnoreCase ("company")) {Case [] cases = allLawyers [i] . getLawyerCases (); for (int j = 0; j < 15; j + +) {if (cases [j] \neq = null && cases [j] . getJudgmentGiven () == ’A’) {System.out.println (cases [j] . getCaseClient ());}}}}}

Mentor tip

Check lawyerType is 'company', cases[j] != null, and judgmentGiven == 'A' before printing getCaseClient().

(b)

5 marks

Benefits of Open Source: Reduces software development costs by providing mature, community-vetted frameworks and libraries without licensing fees. Public source visibility encourages rapid peer review, active vulnerability patching, and global collaboration.
Drawbacks and Risks: Dependency on unmaintained external libraries can introduce software supply chain vulnerabilities. Additionally, strict copyleft licenses (such as GPL) can impose legal requirements to open-source proprietary code if integrated improperly.
Conclusion: Open source software fosters innovation and cost efficiency, but organizations must implement robust dependency scanning and license governance policies to safeguard proprietary code.

Mentor tip

Discuss cost benefits and community collaboration versus security supply chain and licensing risks, followed by a reasoned conclusion.

Our analysis

Where the marks are

Each part sorted by the topic it mostly tests. Databases carries the most marks on this paper.

Databases45 marks · 25%
Q1, Q2, Q3
Modelling and simulation45 marks · 25%
Q4, Q5, Q6
Web science45 marks · 25%
Q7, Q8, Q9
Object-oriented programming45 marks · 25%
Q10, Q11, Q12

Mentor insights

Traps to watch for

01
Partial Key Dependency in 2NF Normalization · Q2
In composite primary key tables like TEST(PatientID, NurseID, DateTime), forgetting that PatientName depends solely on PatientID leads to misidentifying the table as 2NF. Second normal form requires full functional dependency on the entire composite key.
02
Permutation Cycle Structures vs Independent Random Guessing · Q4
Assuming the children open boxes independently at random yields a joint win probability of (1/2)^10 ≈ 0.098%. The follow-the-number strategy exploits permutation cycle structures, raising team success to over 31%. Confusing independent random selection with directed cycle traversal distorts the simulation logic.
03
Null Reference Exceptions in Fixed-Length Object Arrays · Q10
Iterating through lawyerCases[] up to index 14 without verifying cases[j] != null causes a NullPointerException at runtime whenever a lawyer handles fewer than 15 active cases. Always verify that object references are non-null before invoking getters.
04
Undeclared Variables and Unincremented While Loops · Q12
The loop while (x > 20) fails to execute because x=0, but even if corrected to x < 20, neglecting to increment x++ creates an infinite loop, and referencing countB without an explicit type declaration triggers a compilation failure.

Solutions and tips are GeorgeTheTutor's own; the questions are not reproduced here. Not affiliated with or endorsed by the International Baccalaureate Organization.

Want someone to go through your paper with you?

Our maths mentors took these exams recently. One-to-one, they'll find where your marks are going and fix it before May.

Find your maths mentor More past papers

Worked solutions

IB Computer Science SL · Paper 2

Computer Science, Standard Level

Session: November 2025
Marks: 180
Duration: 1 hour
Calculator: Not allowed
Questions: 12

Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12

Keep your copy of the paper open alongside. We don't reproduce the questions: each part below is labelled to match the November 2025 Global paper, which your school or the IB can provide.

Practice mode

Hide our working until you've had a go.

Option A: Databases · Option

Questions 1–3 · 45 marks

Question 1

15 marks

Relational Modeling, Electronic Health Records, Transaction States, and Crash Recovery

Databases

(a)

3 marks

Identify primary entities: `Clinic`, `Dentist`, and `Nurse`.
Define multiplicities with Clinic: A single clinic employs many dentists (1-to-many relationship from `Clinic` to `Dentist`), and the same clinic employs many nurses (1-to-many relationship from `Clinic` to `Nurse`).
Define the relationship between Dentist and Nurse: Because a dentist collaborates with multiple nurses and a nurse assists multiple dentists, their direct association is many-to-many (M:N). In a normalized relational implementation, this M:N relationship is resolved through an associative link table (such as `DutyAssignment` or `Appointment`) having two 1-to-many relationships pointing from `Dentist` and `Nurse`.

Mentor tip

Explicitly draw or state the entities with crow's foot notation or an associative intersection table to resolve many-to-many.

(b)

6 marks

Advantage 1 (Concurrent Multi-User Access and Mobility): Storing patient records in a centralized relational database allows dentists, dental hygienists, and reception staff to retrieve and update patient medical charts simultaneously across multiple clinical terminals. Unlike physical paper files that can only exist in one room at a time, digital records facilitate seamless cross-departmental coordination.
Advantage 2 (Automated Data Validation, Integrity, and Redundancy Prevention): Database management systems enforce domain constraints, required foreign keys, and standardized picklists during input. This eliminates handwriting ambiguity, prevents contradictory records, and drastically reduces clerical errors compared to paper charting.
Advantage 3 (Disaster Recovery and Longitudinal Data Mining): Electronic databases support automated off-site backups and replication, safeguarding patient histories against physical perils like fire or water damage. Furthermore, digitized records enable structured SQL analytics to monitor patient treatment progressions and track clinic-wide outcomes.

Mentor tip

Provide two comprehensive explanations, detailing the operational mechanism and its clinical benefit.

(c)

3 marks

Step 1 (Active State): The transaction begins execution, performing read and write operations in main memory buffers as patient billing data is updated.
Step 2 (Partially Committed State): Once the final database statement executes, the transaction enters the partially committed state, where updates reside in volatile memory awaiting physical persistence.
Step 3 (Committed or Aborted State): If all write operations and integrity constraints flush successfully to non-volatile disk storage, the transaction transitions to the Committed state. If a hardware fault or constraint violation occurs before completion, it transitions to the Failed state and is rolled back to the Aborted state, restoring data consistency.

Mentor tip

Trace the lifecycle: Active -> Partially Committed -> Committed (or Failed -> Aborted).

(d)

3 marks

Step 1 (Log Scan and Checkpoint Identification): Following a system failure, the recovery manager scans backward through the write-ahead transaction log to identify the most recent consistent checkpoint record.
Step 2 (Redo Phase): Scanning forward from the checkpoint, the recovery manager re-executes all operations for transactions that logged a successful `COMMIT` record prior to the failure, ensuring the ACID durability property.
Step 3 (Undo Phase): The recovery manager identifies all active transactions that were still uncommitted at the moment of the crash and reverses their operations in reverse chronological order using logged 'before-values', restoring full atomicity.

Mentor tip

Highlight the checkpoint reference, the REDO step for committed operations, and the UNDO step for uncommitted transactions.

Question 2

15 marks

Composite Primary Keys, Second Normal Form, SQL Filtering, and 3NF Relational Schema

Databases

(a)

2 marks

A patient can visit the clinic multiple times across different days, or receive multiple procedures from the same nurse on separate visits.
Without `DateTime`, the composite key `(PatientID, NurseID)` would prohibit a patient from ever being treated by the same nurse more than once due to primary key collisions. Incorporating `DateTime` ensures that every unique testing event remains uniquely identifiable.

Mentor tip

Point out that DateTime guarantees uniqueness across recurring appointments between the same patient and nurse.

(b)

2 marks

Second Normal Form (2NF) mandates that the table must already be in 1NF and that every non-key attribute must be fully functionally dependent on the complete primary key.
In the `TEST` table, `PatientName` is functionally dependent purely on `PatientID` alone rather than the composite key `(PatientID, NurseID, DateTime)`. This partial dependency violates 2NF.

Mentor tip

State the definition of 2NF and identify the partial functional dependency PatientName -> PatientID.

(c)

2 marks

Scan the `TEST` table for records matching `TestType = 'Oral X-ray'` and `NurseID = 'BM2385'`.
Matching Row 1: PatientID 5, NurseID BM2385, TestType Oral X-ray -> Simon Moore.
Matching Row 3: PatientID 78, NurseID BM2385, TestType Oral X-ray -> Scottie Ngatai.
Final projected output contains two patient names:
Simon Moore
Scottie Ngatai

Mentor tip

Verify both criteria against each record; ensure only the requested PatientName attribute is listed.

(d)

4 marks

Construct an SQL `SELECT` statement joining `TEST` and `NURSE` on `NurseID`.
Filter the nurse's full name to 'Anya Svetlana'.
Restrict `DateTime` to cover the entire day of 19 November 2024 using pattern matching or range comparison.
SQL Implementation:
SELECT TEST.PatientName FROM TEST INNER JOIN NURSE ON TEST.NurseID = NURSE.NurseID WHERE NURSE.NurseName = 'Anya Svetlana' AND TEST.DateTime LIKE '2024/11/19%';

SELECT TEST.PatientName FROM TEST INNER JOIN NURSE ON TEST.NurseID = NURSE.NurseID WHERE NURSE.NurseName = ’Anya Svetlana’ AND TEST.DateTime LIKE ’2024/11/19%’;

Mentor tip

Use an explicit INNER JOIN on the shared NurseID foreign key and filter date using LIKE '2024/11/19%'.

(e)

5 marks

Decompose the unnormalized appointment data into Third Normal Form (3NF) to eliminate partial and transitive dependencies.
Table 1: `PATIENT (PatientID, PatientName)` — isolates patient biographical information.
Table 2: `DENTIST (DentistID, DentistName, Specialties)` — removes dentist transitive dependencies.
Table 3: `TREATMENT (TreatmentID, TreatmentName, Cost)` — isolates treatment pricing from individual appointments.
Table 4: `APPOINTMENT (AppointmentID, PatientID, DentistID, TreatmentID, VisitDateTime)` — resolves appointment events with primary key `AppointmentID` and foreign keys referencing `PATIENT`, `DENTIST`, and `TREATMENT`.

PATIENT (\underline{PatientID}, PatientName) DENTIST (\underline{DentistID}, DentistName, Specialties) TREATMENT (\underline{TreatmentID}, TreatmentName, Cost) APPOINTMENT (\underline{AppointmentID}, PatientID^{*}, DentistID^{*}, TreatmentID^{*}, VisitDateTime)

Mentor tip

Ensure 3NF by isolating Dentist and Treatment entities into separate relations so non-key attributes depend only on the primary key.

Question 3

15 marks

Data Integrity Maintenance, Integrated Databases, Visual Querying, and Clinical Mining Ethics

Databases

(a)

2 marks

Strategy (Referential Integrity Constraints and Automated Backups): The database administrator configures database foreign key constraints (`ON DELETE RESTRICT` or `CASCADE`) and domain validation checks to prevent orphaned records or corrupt data entry.
Complementing this, the DBA schedules automated incremental off-site replication and periodic transaction log snapshots to guarantee rapid data restoration without corruption during unexpected hardware faults.

Mentor tip

Mention enforceable constraint rules (domain/referential integrity) or regularized automated replication.

(b)

2 marks

An integrated database system consolidates previously fragmented operational repositories (such as patient clinical records, billing, appointment scheduling, and inventory) into a unified, shared database structure.
This single source of truth eliminates redundant data entry across clinic departments, ensuring that any modification updates immediately across all clinic modules.

Mentor tip

Emphasize centralized unified data storage and elimination of departmental silos.

(c)

3 marks

Visual query tools (Query By Example - QBE) present the user with a graphical interface showing table names, field picklists, and criteria grids instead of requiring manual SQL coding.
The clinical user drags required fields (such as `PatientName` and `Treatment`) into the visual design grid and enters filtering conditions into designated constraint boxes (e.g. typing '> 200' under Cost or selecting 'Orthodontics' from a dropdown menu).
The underlying database software automatically converts these visual selections into a structured SQL query, executes it against the server, and formats the output into an interactive data view.

Mentor tip

Explain GUI table representations, setting criteria in visual rows, and automatic translation to executable queries.

(d)

3 marks

Data mining algorithms (such as classification decision trees and cluster analysis) process large volumes of historical patient diagnostic records, geographic demographic data, and seasonal visit frequencies.
The algorithms identify subtle statistical correlations and non-obvious disease patterns (for instance, abnormal spikes in specific oral bacterial infections clustered within a particular district or demographic group).
Healthcare providers use these predictive trend models to anticipate upcoming community outbreaks and deploy preventive health advisories before infections escalate.

Mentor tip

Describe pattern discovery across patient datasets and using predictive models for early public health intervention.

(e)

5 marks

Advantages for the Clinic: Direct access enables instant electronic verification of treatments, accelerating insurance claim pre-authorizations and cash flow reimbursements. This substantially reduces clerical administrative overhead, phone inquiries, and disputed claim delays.
Disadvantages for the Clinic: Exposing clinical records creates severe patient confidentiality and data security risks. Accidental leakage of protected health information could trigger substantial legal liabilities under data protection regulations (such as GDPR or HIPAA). Furthermore, third-party direct access creates potential security attack vectors if external credentials are compromised.
Evaluation & Conclusion: The clinic should not allow direct unrestricted database access. Instead, it should deploy a strictly scoped, read-only API gateway with field-level encryption, exposing only anonymized billing codes and procedure dates necessary for insurance claims.

Mentor tip

Provide a balanced appraisal covering administrative speed vs patient confidentiality, supported by a concluding synthesis.

Option B: Modelling and simulation · Option

Questions 4–6 · 45 marks

Question 4

17 marks

Permutation Cycle Strategy, Spreadsheet Limitations, and Monte Carlo Multi-Agent Simulation

Modelling and simulation

(a)(i)

1 mark

A single child opens 5 boxes randomly out of 10 total boxes without replacement.
Probability = 5 / 10 = 1 / 2 = 0.5 (or 50%).

P (Child 0 succeeds) = \frac{5}{10} = 0.5

Mentor tip

Express as a fraction (1/2), decimal (0.5), or percentage (50%).

(a)(ii)

1 mark

Under independent random box selection, each of the 10 children has a 1/2 chance of finding their number.
Joint probability = (1/2)^10 = 1 / 1024 ≈ 0.000977 (or ≈ 0.098%).

P (All 10 succeed) = (\frac{1}{2})^{10} = \frac{1}{1024} \approx 0.000977

Mentor tip

Multiply independent probabilities: (0.5)^10 = 1/1024.

(b)

1 mark

Follow the cycle sequence starting at box 3:
Step 1: Open Box 3 -> contains Paper 1.
Step 2: Open Box 1 -> contains Paper 8.
Step 3: Open Box 8 -> contains Paper 5.
Step 4: Open Box 5 -> contains Paper 6.
Step 5: Open Box 6 -> contains Paper 4.
Sequence of boxes opened: 3 -> 1 -> 8 -> 5 -> 6 (child does not locate number 3).

3 \to 1 \to 8 \to 5 \to 6

Mentor tip

Trace box indices: Box 3 has 1, Box 1 has 8, Box 8 has 5, Box 5 has 6.

(c)

4 marks

Reason 1 (Dynamic Conditional Pointer Traversal): Spreadsheets rely on static directed dependency formulas between cells. The follow-the-number strategy requires iterative pointer chasing where the next cell lookup dynamically depends on the integer returned by the previous step. Implementing variable-length loops that jump dynamically to `BOX[LOOK]` for up to 5 iterations requires complex circular formulas or external scripting.
Reason 2 (Shuffling Permutations without Replacement across Large Iterations): Simulating thousands of trials requires shuffling an array of 10 unique elements without replacement before each run. Generating distinct random permutations across tens of thousands of rows is unwieldy in standard spreadsheets compared to high-level procedural programming languages.

Mentor tip

Detail the challenge of dynamic pointer traversal (looping/state dependency) and generating random non-repeating permutations.

(d)

6 marks

Initialize the `CHILDREN` array of size 10 to `false`.
Shuffle the `BOX` array of 10 integers.
Iterate through each child `N` from 0 to 9.
For each child, initialize current box lookup `LOOK = N` and test up to 5 box attempts.
If `BOX[LOOK] == N`, mark `CHILDREN[N] = true` and exit inner loop; otherwise update `LOOK = BOX[LOOK]`.
Finally, loop from 0 to 9 and output the boolean state of `CHILDREN[N]`.

random.shuffle (BOX) loop N from 0 to 9 CHILDREN [N] = false LOOK = N loop ATTEMPT from 1 to 5 if BOX [LOOK] = N then CHILDREN [N] = true break else LOOK = BOX [LOOK] end if end loop end loop loop N from 0 to 9 output CHILDREN [N] end loop

Mentor tip

Set current pointer LOOK = N and update LOOK = BOX[LOOK] on each unmatching attempt.

(e)

4 marks

Initialize `WIN_COUNT = 0`.
Enclose the trial simulation inside a loop running 10000 iterations.
In each trial, execute the follow-the-number strategy across all 10 children.
Evaluate if all 10 elements in `CHILDREN` equal `true`. If so, increment `WIN_COUNT`.
After completing 10000 trials, calculate `PERCENTAGE = (WIN_COUNT / 10000) * 100` and display the result.

WIN_COUNT = 0 loop TRIAL from 1 to 10000 random.shuffle (BOX) ALL_WON = true loop N from 0 to 9 LOOK = N FOUND = false loop A from 1 to 5 if BOX [LOOK] = N then FOUND = true break else LOOK = BOX [LOOK] end if end loop if not FOUND then ALL_WON = false end loop if ALL_WON then WIN_COUNT = WIN_COUNT + 1 end loop PERCENT = (WIN_COUNT /10000) * 100 output PERCENT + "%"

Mentor tip

Verify all 10 children succeed within the same trial before incrementing the prize counter.

Question 5

16 marks

Autonomous Vehicle Modeling, Simulation Variables, Test Scenarios, and Real-World Limitations

Modelling and simulation

(a)

1 mark

Identify one system suitable for computer modeling: Urban traffic flow (or vehicle braking dynamics, climate systems, suspension mechanics, or aerodynamic airflow).

Mentor tip

State a well-defined physical or computational system.

(b)

1 mark

Identify one traffic condition variable: Traffic density (vehicles per kilometer), average traffic speed, intersection queue length, or pedestrian crosswalk frequency.

Mentor tip

Provide a measurable variable describing traffic state.

(c)

4 marks

Grouping 1 (Environmental and Atmospheric Conditions): Variables capturing external surroundings, such as precipitation level (mm/h), ambient road surface friction coefficient, fog visibility distance, and illumination level (lux).
Grouping 2 (Vehicle Sensor and Perception Telemetry): Variables measuring sensor hardware inputs, such as LiDAR point cloud reflection density, radar Doppler velocity signals, optical camera frame rate, and ultrasonic distance readings.

Mentor tip

Outline two distinct categories with relevant representative variables.

(d)

4 marks

Similarities: Both mathematical models and computer simulations use quantitative data and mathematical equations to predict vehicle dynamics, allowing engineers to test behaviors without endangering real passengers.
Differences (Theory vs Virtual Realization): A mathematical model represents abstract analytical equations (e.g. differential equations governing friction and braking distance). In contrast, a simulation applies these mathematical foundations inside a dynamic time-stepped virtual world, running iterative edge-case scenarios with multiple interacting agents.

Mentor tip

Contrast the theoretical mathematical foundation with the empirical, dynamic time-stepped simulation environment.

(e)

2 marks

Test-case scenario (LiDAR Occlusion at Highway Speed): The simulation systematically disables the forward LiDAR sensor or degrades its point cloud with simulated heavy snowfall while cruising at 100 km/h.
Evaluation metric: Test whether the autonomous vehicle immediately detects sensor failure, successfully switches to radar and stereoscopic vision redundancy, alerts the passenger, and smoothly navigates to a safe stop.

Mentor tip

Describe a specific failure condition and the vehicle's required automated fail-safe response.

(f)

4 marks

Identification: The reality gap and simulator overfitting.
Expansion: Real-world urban driving contains chaotic, unpredictable occurrences—such as jaywalking pedestrians, unusual road construction markers, and extreme weather reflections—that cannot be comprehensively anticipated in a synthetic virtual environment.
Consequence: Autonomous vehicle algorithms may become over-optimized for simulator physics, giving developers false confidence. When deployed onto actual roadways, unmodeled corner cases can cause sensor confusion, erratic braking, or fatal collisions.

Mentor tip

Structure your response around the reality gap, the difficulty of replicating chaotic edge cases, and safety risks.

Question 6

12 marks

Monte Carlo Pi Estimation via Spreadsheets and 3D Bounding Box Real-Time Computation

Modelling and simulation

(a)

2 marks

The Monte Carlo method for estimating π is based on a two-dimensional geometric proof: the ratio of the area of a circle quadrant to the enclosing unit square.
Because points are randomly placed using paired Cartesian coordinates (x, y), a 2D scatter plot directly represents the spatial distribution of samples and visualizes the circular arc boundary intuitively.

Mentor tip

Explain that the geometric area ratio is inherently two-dimensional, making 2D scatter plots an ideal visual match.

(b)(i)

1 mark

Apply the Pythagorean theorem in cell C2: `=SQRT(ColA2^2 + B2^2)`

C2: = SQRT (ColA2 ˆ 2 + B2 ˆ 2)

Mentor tip

Use the standard spreadsheet SQRT function with squared coordinates.

(b)(ii)

1 mark

Classify distance less than or equal to 1 as inside the circle: `=IF(C2<=1, "In", "Out")`

D2: = IF (C2 <= 1, "In", "Out")

Mentor tip

Use an IF conditional testing if distance <= 1.

(c)

3 marks

Count the number of points falling inside the circular quadrant across the 5000 generated coordinates (rows 2 to 5001): `COUNTIF(D2:D5001, "In")`.
Divide the count by the total sample size (5000) to find the fractional area of the quadrant.
Multiply by 4 because the area of the unit quadrant equals π / 4.
Spreadsheet Formula: `=COUNTIF(D2:D5001, "In") / 5000 * 4` (or `=COUNTIF(D2:D5001, "In") / COUNTA(D2:D5001) * 4`).

E2: = COUNTIF (D2:D5001, "In") /5000 * 4

Mentor tip

Multiply the ratio of inside points to total points by 4.

(d)

5 marks

Processing Power Demands: Calculating 3D oriented bounding boxes in real time requires processing millions of raw LiDAR 3D point cloud coordinates and multi-camera video feeds per second. Convolutional neural networks and point-cloud clustering algorithms must execute at high frame rates (e.g. 30–60 Hz) with sub-10ms latencies to avoid collision hazards.
Memory Constraints: Storing high-density voxel grids, intermediate tensor feature maps, and multi-sensor temporal tracking buffers demands substantial high-bandwidth memory (HBM/VRAM) to prevent processing bottlenecks.
Evaluation & Synthesis: Standard automotive microcontrollers cannot handle these workloads. Autonomous vehicles require dedicated hardware accelerators (such as multi-core Tensor Processing Units or GPUs) paired with low-latency bus architectures to ensure real-time safety.

Mentor tip

Evaluate both processing throughput (GPU point-cloud clustering) and high-bandwidth memory storage, concluding with dedicated hardware necessity.

Option C: Web science · Option

Questions 7–9 · 45 marks

Question 7

15 marks

Web Protocols, Architecture, CGI Processing, and Server-Side PHP Execution

Web science

(a)(i)

1 mark

State the protocol in the URL `http://hardware4hotels.example.com`: `http` (Hypertext Transfer Protocol).

Mentor tip

State 'http' or 'Hypertext Transfer Protocol'.

(a)(ii)

1 mark

State the domain name in the URL: `hardware4hotels.example.com`.

Mentor tip

Omit the protocol prefix; state 'hardware4hotels.example.com'.

(b)

2 marks

Protocols establish standardized communication rules, data packet formats, and error-handling conventions across diverse hardware and operating systems.
This ensures interoperability between different web browsers and web servers, guaranteeing that requested web documents are transmitted reliably and rendered identically.

Mentor tip

Highlight standardization, interoperability, and reliable client-server communication.

(c)

2 marks

Identify two components of a webpage: HTML structural markup and CSS stylesheets (or client-side JavaScript, hyperlinks, navigation bar, multimedia elements).

Mentor tip

Name two structural or interactive elements.

(d)

2 marks

The Common Gateway Interface (CGI) serves as a standardized middleware gateway connecting the web server to backend executable programs and scripts.
When a client submits the feedback form, CGI relays the HTTP POST parameters to the backend script and forwards the dynamically generated response back to the client browser.

Mentor tip

Explain CGI as a gateway passing form data between the web server and backend scripts.

(e)(i)

4 marks

1. Parameter Extraction: The server captures the submitted `work_category` and `rating` values from the HTTP POST payload via `$_POST`.
2. Database Connection & Query: The script establishes a connection to the `sales` database and queries the existing average rating for that work category using `SELECT AVG(rating)`.
3. Record Insertion: The script inserts the newly submitted rating into the `ratings` table using an SQL `INSERT` statement.
4. Conditional Output & Cleanup: If the submitted rating is greater than or equal to the category average, it outputs a positive acknowledgment; otherwise, it outputs an apology. Finally, it closes the database connection.

Mentor tip

Detail four sequential steps: extraction, average calculation query, record insertion, and conditional echo output.

(e)(ii)

3 marks

Security and Credential Protection: Executing database scripts server-side prevents database credentials (` $u ser nam e ‘, ‘$ password`) from being exposed to the client browser, mitigating direct SQL injection or credential theft.
Furthermore, browsers cannot connect directly to secure SQL databases without exposing database ports publicly over the internet. Server-side processing maintains centralized access control and data integrity.

Mentor tip

Emphasize security, hiding database credentials, and preventing unauthorized direct database access.

Question 8

17 marks

Search Engine Results Estimation, PageRank, SEO Techniques, and Web Evolution

Web science

(a)

2 marks

The reported figure of 248,000 results is an algorithmic statistical estimate generated by the search engine to provide rapid response times, not an exact count of dedicated pages.
The index retrieves all pages matching any combination of the individual keywords ('one', 'seven', 'academy') across billions of web pages rather than exclusively retrieving pages solely focused on the school.

Mentor tip

Explain that the count is a rapid statistical estimate matching broad individual keywords across the index.

(b)

3 marks

Principle of Inbound Link Authority (Random Surfer Model): PageRank treats each inbound hyperlink to a webpage as a vote of recommendation.
A page achieves a higher PageRank if it receives inbound links from reputable, high-ranking pages rather than low-authority sites.
The link equity contributed by a pointing page is divided proportionally by the total number of outbound links on that page, damping vote dilution.

P R (A) = (1 - d) + d i = 1 \sum n \frac{P R ( T _{i} )}{C ( T _{i} )}

Mentor tip

Explain incoming links as votes of authority and note that links from high-ranking pages contribute greater equity.

(c)

2 marks

Way 1: Securing authoritative external backlinks from verified educational bodies and municipal school boards.
Way 2: Improving website technical performance, including mobile responsiveness and fast Core Web Vitals load times.

Mentor tip

State two legitimate ranking factors such as authoritative backlinks and mobile page speed optimization.

(d)

3 marks

Meta tags are placed in the `<head>` block of an HTML document, containing structured metadata (such as page description and keywords) that is not rendered visually to human users.
When a web crawler visits the page, it parses the HTML DOM, extracts the meta content, and indexes these descriptive tokens in its database to categorize the topic.
Search engines often use the meta description as the snippet displayed directly on search engine results pages.

Mentor tip

Describe placement in HTML head, parsing by web crawlers, and storage in the indexing catalog.

(e)

5 marks

White Hat SEO: Adheres strictly to search engine guidelines by providing high-quality original curriculum content, semantic HTML tags, fast loading speeds, and organic inbound links. It provides stable, sustained long-term visibility without penalties.
Black Hat SEO: Employs deceptive practices such as keyword stuffing, invisible text, cloaking, and private blog link networks. While it may provide short-term traffic spikes, modern search algorithms quickly detect manipulative patterns.
Consequences & Conclusion: Black hat tactics carry severe risks of search engine manual actions, algorithmic penalties, or complete domain de-indexing. For an educational institution like One2Seven Academy, preserving institutional credibility is paramount; the webmaster must strictly employ white hat SEO.

Mentor tip

Compare white hat vs black hat techniques, discussing sustainability vs penalties, and conclude decisively.

(f)

2 marks

Challenge 1 (Exponential Growth of Dynamic and AI-Generated Content): Crawling budgets are strained by billions of ephemeral JavaScript-rendered pages and synthetic content.
Challenge 2 (Combatting Disinformation and Misleading Content): Search engines must refine relevance algorithms to distinguish authoritative facts from automated web spam while maintaining user privacy.

Mentor tip

Identify challenges like crawling dynamic content at scale and filtering automated spam/disinformation.

Question 9

13 marks

Mobile vs Ubiquitous Computing, Web Decentralization, and Cloud Infrastructure Models

Web science

(a)

2 marks

Mobile Computing involves portable, battery-powered devices (such as smartphones and laptops) that users actively carry and consciously operate to access network services.
Ubiquitous Computing (pervasive computing) embeds computing intelligence invisibly into physical environments and everyday objects (e.g. smart sensors, connected lighting), operating autonomously with minimal active user engagement.

Mentor tip

Differentiate portable, user-operated mobile devices from embedded, invisible ubiquitous environments.

(b)

3 marks

Reason: Hardware Miniaturization and High-Speed Wireless Infrastructure.
Expansion: Advances in semiconductor fabrication have enabled powerful multi-core processors, low-power displays, and compact lithium batteries.
Consequence: Paired with ubiquitous 4G/5G mobile networks and public Wi-Fi, users can access high-bandwidth cloud services, streaming media, and enterprise software anywhere at affordable prices.

Mentor tip

Explain hardware miniaturization paired with high-speed cellular networks (4G/5G).

(c)

3 marks

Distributed networks and cloud platforms replace single centralized server architectures with thousands of geographically dispersed edge nodes and data centers.
Content Delivery Networks (CDNs) cache content and process microservices closer to end users, preventing central points of bottleneck or failure.
This democratizes web hosting and decentralizes data routing, enhancing system resilience against outages and reducing regional latency.

Mentor tip

Discuss moving from monolithic central servers to geographically dispersed cloud nodes and CDNs.

(d)

5 marks

Private Cloud: Provides dedicated on-premise or single-tenant infrastructure. Offers full control over security configurations, customizable data privacy governance, and predictable performance. However, it requires substantial upfront capital expenditure and ongoing IT administrative overhead.
Public Cloud: Multi-tenant infrastructure managed by external providers (e.g. AWS, Azure). Offers near-infinite scalability, pay-as-you-go pricing, and zero hardware maintenance. However, it introduces potential security and regulatory concerns because physical hardware is shared with other tenants.
Evaluation & Conclusion: A hybrid cloud model is often optimal for modern enterprises: keeping sensitive customer and proprietary financial data inside a secure private cloud while leveraging the elastic scalability of the public cloud for web applications.

Mentor tip

Evaluate security, cost, scalability, and maintenance for both models, recommending a hybrid approach.

Option D: Object-oriented programming · Option

Questions 10–12 · 45 marks

Question 10

15 marks

Object Modeling, Encapsulation, Array Searching, and Delayed Case Evaluation

Object-oriented programming

(a)

1 mark

A constructor is a special class method invoked during object instantiation that allocates memory and initializes instance variables with starting values.

Mentor tip

Define constructor as an initialization method invoked upon instantiation.

(b)

2 marks

The method inspects the `judgmentGiven` character parameter.
If `judgmentGiven` equals `'F'` (favorable), it updates `caseStatus` to `'won'`; if `judgmentGiven` equals `'A'` (against), it updates `caseStatus` to `'lost'` (either directly or via `setCaseStatus`).

Mentor tip

Explain testing judgmentGiven and updating caseStatus to 'won' or 'lost'.

(c)

4 marks

Loop through the `allLawyers` array of size 20.
Check that the lawyer element is not null.
Verify that `lawyerType` equals 'civil' (case-insensitively).
Verify that `casesWon` is strictly greater than 30.
Print the matching lawyer's name to the console.

Mentor tip

Check lawyerType.equalsIgnoreCase('civil') and casesWon > 30 before printing.

(d)(i)

1 mark

A return statement terminates method execution and passes a specified result value back to the calling method, returning program control.

Mentor tip

State that return sends a value back and returns control to the caller.

(d)(ii)

2 marks

Method signature requires return type `int`, method name `countDelayedCases`, and an integer parameter for lawyer ID:
`public static int countDelayedCases(int lawyerID)` (or `public int countDelayedCases(int id)`).

Mentor tip

Include int return type, method name countDelayedCases, and int parameter.

(d)(iii)

5 marks

Initialize `count = 0`.
Loop through `allLawyers[]` to locate the lawyer whose ID matches `lawyerID`.
Upon finding the matching lawyer, retrieve their `lawyerCases[]` array.
Loop through each case in `lawyerCases[]`, ensuring non-null references.
Check if `daysFromStart > 45` and `judgmentGiven == 'N'`. If both hold, increment `count`.
Return `count` after loop termination.

public static int countDelayedCases (int lawyerID) {int count = 0; for (int i = 0; i < 20; i + +) {if (allLawyers [i] \neq = null && allLawyers [i] . getLawyerID () == lawyerID) {Case [] cases = allLawyers [i] . getLawyerCases (); for (int j = 0; j < 15; j + +) {if (cases [j] \neq = null && cases [j] . getDaysFromStart () > 45 && cases [j] . getJudgmentGiven () == ’N’) {count + +;}}}} return count;}

Mentor tip

Verify cases[j] != null before checking getDaysFromStart() > 45 and getJudgmentGiven() == 'N'.

Question 11

15 marks

Subclass Inheritance, Aggregation, and UML Class Diagram Modeling

Object-oriented programming

(a)(i)

2 marks

Identify suitable data types for TraineeLawyer attributes:
- joining date: `String` (or `Date` / `int`)
- exam status: `boolean` (or `String` / `char`)

Mentor tip

Specify String/Date for joining date and boolean/String for exam status.

(a)(ii)

4 marks

Construct detailed UML class diagram with three-part compartments (name, attributes, methods):
1. Superclass `Lawyer`:
Attributes: `- lawyerID: int`, `- lawyerName: String`, `- lawyerType: String`, `- lawyerRank: int`, `- casesWon: int`, `- lawyerCases: Case[]`
Methods: `+ Lawyer(...)`, `+ getLawyerName(): String`, `+ getLawyerType(): String`, `+ getLawyerRank(): int`, `+ setLawyerRank(int): void`
2. Subclass `TraineeLawyer`:
Attributes: `- joiningDate: String`, `- examStatus: boolean`
Methods: `+ TraineeLawyer(...)`, `+ getJoiningDate(): String`, `+ getExamStatus(): boolean`
3. Relationship: Draw an inheritance arrow (solid line with open triangular arrowhead) pointing from `TraineeLawyer` up to `Lawyer`.

Mentor tip

Draw three compartments per class and use an open triangular arrow for inheritance.

(b)(i)

1 mark

State the relationship between `TraineeLawyer` and `Lawyer`: Inheritance ('is-a' relationship; `TraineeLawyer` is a subclass of `Lawyer`).

Mentor tip

State inheritance or 'is-a' relationship.

(b)(ii)

1 mark

State the relationship between `Lawyer` and `Case`: Aggregation / Composition ('has-a' relationship; `Lawyer` contains an array of `Case` objects).

Mentor tip

State aggregation or 'has-a' relationship.

(b)(iii)

4 marks

Difference 1 (Inheritance vs Containment): Inheritance ('is-a') creates a specialized class that inherits attributes and behaviors from a superclass, supporting method overriding. Aggregation ('has-a') represents object containment where one class stores references to independent objects without inheriting their class interface.
Difference 2 (Polymorphism and Coupling): Inheritance enables polymorphic substitution (a `TraineeLawyer` instance can be assigned to a `Lawyer` reference). Aggregation does not support polymorphism between container and contained objects; `Case` objects remain distinct entities.

Mentor tip

Contrast 'is-a' specialization and polymorphism against 'has-a' containment and reference lifecycle.

(c)

2 marks

Encapsulation is the object-oriented principle of packaging data attributes and the methods that operate on them within a single class.
Internal instance variables are kept private to protect internal state, exposing controlled public getters and setters for controlled access.

Mentor tip

Explain packaging variables and methods in a single unit and restricting direct access via private fields.

(d)

1 mark

A parent object (or superclass) is a base class in an inheritance hierarchy whose attributes, methods, and behaviors are inherited and extended by child subclasses.

Mentor tip

Define parent class as a superclass providing inheritable members to subclasses.

Question 12

15 marks

Loop Logic Debugging, Multi-Entity Array Traversal, and Open Source Software Impact

Object-oriented programming

(a)(i)

3 marks

1. Faulty Loop Condition: The condition `while (x > 20)` evaluates to false immediately since `x = 0`, so the loop body never executes (should be `x < 20`).
2. Missing Loop Counter Increment: The index variable `x` is never incremented inside the loop body (`x++`), which would cause an infinite loop if entered.
3. Undeclared Variable: The variable `countB` is used inside the loop (`countB++`) without prior declaration or type definition, triggering a compilation error.

Mentor tip

Identify false while condition (x > 20), missing x++ increment, and undeclared countB variable.

(a)(ii)

7 marks

Construct `displayClientNames()` method:
1. Traverse `allLawyers` array of length 20.
2. Check that the lawyer element is not null and `lawyerType` equals 'company'.
3. Retrieve the lawyer's `lawyerCases[]` array.
4. Iterate through the 15 cases, checking for non-null elements.
5. Check if `judgmentGiven == 'A'`. If matched, retrieve `caseClient` and print to the console.

public static void displayClientNames () {for (int i = 0; i < 20; i + +) {if (allLawyers [i] \neq = null && allLawyers [i] . getLawyerType () . equalsIgnoreCase ("company")) {Case [] cases = allLawyers [i] . getLawyerCases (); for (int j = 0; j < 15; j + +) {if (cases [j] \neq = null && cases [j] . getJudgmentGiven () == ’A’) {System.out.println (cases [j] . getCaseClient ());}}}}}

Mentor tip

Check lawyerType is 'company', cases[j] != null, and judgmentGiven == 'A' before printing getCaseClient().

(b)

5 marks

Benefits of Open Source: Reduces software development costs by providing mature, community-vetted frameworks and libraries without licensing fees. Public source visibility encourages rapid peer review, active vulnerability patching, and global collaboration.
Drawbacks and Risks: Dependency on unmaintained external libraries can introduce software supply chain vulnerabilities. Additionally, strict copyleft licenses (such as GPL) can impose legal requirements to open-source proprietary code if integrated improperly.
Conclusion: Open source software fosters innovation and cost efficiency, but organizations must implement robust dependency scanning and license governance policies to safeguard proprietary code.

Mentor tip

Discuss cost benefits and community collaboration versus security supply chain and licensing risks, followed by a reasoned conclusion.

Our analysis

Where the marks are

Each part sorted by the topic it mostly tests. Databases carries the most marks on this paper.

Databases45 marks · 25%
Q1, Q2, Q3
Modelling and simulation45 marks · 25%
Q4, Q5, Q6
Web science45 marks · 25%
Q7, Q8, Q9
Object-oriented programming45 marks · 25%
Q10, Q11, Q12

Mentor insights

Traps to watch for

01
Partial Key Dependency in 2NF Normalization · Q2
In composite primary key tables like TEST(PatientID, NurseID, DateTime), forgetting that PatientName depends solely on PatientID leads to misidentifying the table as 2NF. Second normal form requires full functional dependency on the entire composite key.
02
Permutation Cycle Structures vs Independent Random Guessing · Q4
Assuming the children open boxes independently at random yields a joint win probability of (1/2)^10 ≈ 0.098%. The follow-the-number strategy exploits permutation cycle structures, raising team success to over 31%. Confusing independent random selection with directed cycle traversal distorts the simulation logic.
03
Null Reference Exceptions in Fixed-Length Object Arrays · Q10
Iterating through lawyerCases[] up to index 14 without verifying cases[j] != null causes a NullPointerException at runtime whenever a lawyer handles fewer than 15 active cases. Always verify that object references are non-null before invoking getters.
04
Undeclared Variables and Unincremented While Loops · Q12
The loop while (x > 20) fails to execute because x=0, but even if corrected to x < 20, neglecting to increment x++ creates an infinite loop, and referencing countB without an explicit type declaration triggers a compilation failure.

Solutions and tips are GeorgeTheTutor's own; the questions are not reproduced here. Not affiliated with or endorsed by the International Baccalaureate Organization.

Want someone to go through your paper with you?

Our maths mentors took these exams recently. One-to-one, they'll find where your marks are going and fix it before May.

Find your maths mentor More past papers