Loading…

Worked solutions

IB Computer Science HL · Paper 2

Computer Science, Higher Level

Session: May 2015
Marks: 260
Duration: 1 hour 20 min
Calculator: Not allowed
Questions: 18

Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Q15 Q16 Q17 Q18

Keep your copy of the paper open alongside. We don't reproduce the questions: each part below is labelled to match the May 2015 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 1–4 · 65 marks

Question 1

15 marks

Hotel booking database transactions, data versus information, isolation, and data sharing

Databases

(a)

2 marks

Distinguish raw data from processed information in the hotel booking context:
Data consists of unorganized, unprocessed values lacking context on their own, such as the isolated booking string '3EP9RE'.
Information is structured, interpreted data that carries meaning for the recipient, such as the follow-up confirmation email linking the code '3EP9RE' to the specific reservation details (double non-smoking room at the Overlook Hotel).

Mentor tip

Contrast raw symbols (the alphanumeric booking code alone) with contextualized output (the detailed confirmation message).

(b)

2 marks

Define a database transaction:
A database transaction is a single logical unit of work comprising a minimal sequence of read/write operations executed on a database.
It must execute completely or not at all (atomicity) to maintain database integrity, prevent conflicting room reservations, and allow recovery if a failure occurs.

Mentor tip

Emphasize that a transaction is an indivisible unit of work that preserves database consistency and supports rollback on failure.

(c)

3 marks

Describe how transactions enforce data consistency:
1. Atomicity and Persistence: All modifications within the booking transaction become permanent (committed) only if every step in the sequence terminates successfully.
2. Rollback on Failure: If any hardware, network, or logical failure interrupts the transaction prior to completion, a rollback operation undoes all partial changes.
3. State Preservation: Rolling back restores the database to its prior consistent state, ensuring rooms are neither double-booked nor marked unavailable without a valid payment and confirmation.

Mentor tip

Link transaction completion (commit) and failure recovery (rollback) directly to preserving a valid, consistent state.

(d)

2 marks

Explain transaction isolation in the room reservation scenario:
Isolation (concurrency control) ensures that intermediate states of an ongoing transaction are invisible to other concurrent operations until the transaction commits.
For example, while one customer is booking a specific double room, record locking prevents a second customer from simultaneously reserving the same room and causing a double-booking inconsistency.

Mentor tip

Include both the theoretical definition of isolation (concurrent operations cannot interfere) and a concrete example (locking a room record during booking).

(e)

6 marks

Explain the technical mechanism of data sharing:
1. Data sharing and replication expose selected portions of source hotel databases to authorized external systems or user groups.
2. Bringing relevant data closer to end users reduces redundant network transmission overhead and allows heterogeneous systems to query standardized views.
Explain benefits for Stakeholder 1 (Tourists / End Users):
3. Tourists can rapidly query and compare room availability, features, and pricing across numerous hotels from a single unified interface.
4. Real-time access prevents wasted effort attempting to book rooms that are already occupied.
Explain benefits for Stakeholder 2 (Best Fare Aggregation Service / Hotels):
5. Best Fare can dynamically aggregate offers from multiple hotels and scale its platform to new partner hotels without redesigning its core software.
6. Individual hotels gain broader market visibility and can update local room rates or availability in real time while maintaining control over their internal operational records.

Mentor tip

Cover the technical nature of database sharing/replication first, then provide two distinct points for each of two stakeholders (e.g., tourists and the comparison platform or hotel).

Question 2

10 marks

Hospital pharmacy database administration, data dictionaries, and user views

Databases

(a)

2 marks

Outline the responsibilities of the hospital's Database Administrator (DBA):
1. The DBA manages the operational performance, security, and structural integrity of the pharmacy database (e.g., configuring regular backups and recovery procedures).
2. The DBA controls access permissions, authorizing internal medical staff and partner hospitals to view or modify only the specific records appropriate to their clinical roles.

Mentor tip

Focus on administrative duties such as access control, backup management, schema maintenance, and security rather than day-to-day data entry.

(b)

2 marks

Define a data dictionary:
1. A data dictionary is a centralized repository that stores metadata (data about the data) for the database.
2. It defines field names, data types, validation rules, format constraints, and the relationships between tables.

Mentor tip

Mention 'metadata' and give concrete examples of what is stored (e.g., field types, constraints, and table relationships).

(c)

6 marks

Discuss why customized database views are required for different user groups:
1. Database Administrator (DBA): Requires an unrestricted view of the entire schema, data dictionary, and user tables to perform maintenance, structural updates, and security audits.
2. Internal Hospital Staff (Doctors, Pharmacists, Clerks): Require tailored read/write views depending on their clinical duties—doctors need access to patient prescription histories, while pharmacy clerks need views focused on stock quantities and supplier orders.
3. External Staff at Nearby Hospitals: Must be restricted to a read-only view showing only medicine stock availability (for emergency inter-hospital borrowing) while strictly hiding confidential patient medical records to comply with privacy laws.

Mentor tip

Address all three user groups mentioned in the prompt (DBA, internal staff, nearby hospital staff), stating both what they can see/do and why.

Question 3

20 marks

Shelving unit orders normalization from unnormalized form to 3NF

Databases

(a)

3 marks

Identify three reasons why the given table violates First Normal Form (1NF):
1. No primary key is defined to uniquely identify each row.
2. Repeating groups / blank cells exist: a single `OrderNr` and `TotalPrice` span multiple `ProductCode` rows, leaving empty attributes in subsequent rows.
3. Non-atomic values: the `Description` attribute combines two distinct pieces of data (material type and unit size, e.g., 'Wood, Small') in a single cell.

Mentor tip

Check the three classic 1NF requirements: atomic values in every cell, no repeating groups/blank multi-row cells, and a defined primary key.

(b)

3 marks

State three advantages of database normalization:
1. Minimizes data redundancy, reducing overall storage space requirements.
2. Eliminates update, insertion, and deletion anomalies, making data maintenance simpler and preventing inconsistencies.
3. Enforces atomic attributes and clear primary/foreign keys, which simplifies writing accurate queries and applying granular security controls.

Mentor tip

List three distinct benefits: reduced redundancy/storage, avoidance of modification anomalies, and simpler querying/security.

(c)

2 marks

Evaluate whether `TotalPrice` can serve as a primary key:
1. No, `TotalPrice` cannot be used alone as a primary key.
2. A primary key must be strictly unique for every record, whereas two different orders (Order 100 and Order 101) share the exact same `TotalPrice` of 190.00.

Mentor tip

Cite the specific duplicate value (`190.00` appearing for both Order 100 and Order 101) to prove why uniqueness fails.

(d)

2 marks

Propose a structural change to make `Description` atomic:
1. Decompose the `Description` column into two separate attributes.
2. Store the material type in one column (e.g., `Material`: Wood, Plastic, Extra-strong/Steel) and the dimension in a second column (e.g., `Size`: Small, Medium, Large).

Mentor tip

Explicitly name the two new columns that replace `Description`.

(e)

3 marks

Construct the 1NF relation(s) with atomic attributes and primary keys:
1. Split the repeating order items or flatten the table, and divide `Description` into `Material` (`Description1`) and `Size` (`Description2`).
2. Define the `ORDERS` table with primary key `OrderNr`: `ORDERS(OrderNr [PK], OrderDate, TotalPrice)` containing rows `(100, 10/10/2014, 190.00)`, `(101, 11/10/2014, 190.00)`, `(102, 11/10/2014, 600.00)`.
3. Define the `ORDER_ITEMS` table with composite primary key `(OrderNr, ProductCode)`: `ORDER_ITEMS(OrderNr [PK], ProductCode [PK], Material, Size, Quantity, UnitPrice)` containing all six individual item rows with `OrderNr` filled in for every row.

OrderNr (PK) 100101102 OrderDate 10/10/2014 11/10/2014 11/10/2014 TotalPrice 190.00 190.00 600.00 OrderNr (PK) 100100100101101102 ProductCode (PK) WS WL PS PS PM XX Material Wood Wood Plastic Plastic Plastic Extra-strong Size Small Large Small Small Medium Large Quantity 112123 UnitPrice 30.00 80.00 40.00 40.00 75.00 200.00

Mentor tip

Always clearly mark the Primary Key (`PK`) and Composite Primary Key (`OrderNr, ProductCode`) when constructing normalized tables.

(f)

2 marks

State the Second Normal Form (2NF) dependency condition:
1. Every non-key attribute in a table must be fully functionally dependent on the entire composite primary key.
2. No non-key attribute may have a partial dependency on only part of a composite primary key.

Mentor tip

State clearly that non-key attributes must depend on the whole composite key, not just a subset of it.

(g)

5 marks

Remove partial and transitive dependencies to reach 3NF:
1. Identify partial dependencies in `ORDER_ITEMS`: `Material`, `Size`, and `UnitPrice` depend solely on `ProductCode`, whereas `Quantity` depends on the full composite key `(OrderNr, ProductCode)`.
2. Separate the product details into a dedicated `PRODUCTS` table with primary key `ProductCode`, removing duplicate product rows (`PS` appears once).
3. Keep `ORDER_ITEMS` with composite primary key `(OrderNr, ProductCode)` and non-key attribute `Quantity`.
4. Keep `ORDERS` with primary key `OrderNr`, `OrderDate`, and `TotalPrice`.
5. Verify 3NF: in all three tables (`ORDERS`, `ORDER_ITEMS`, `PRODUCTS`), every non-key attribute depends solely on its table's primary key with no transitive dependencies between non-key fields.

OrderNr (PK) 100101102 OrderDate 10/10/2014 11/10/2014 11/10/2014 TotalPrice 190.00 190.00 600.00 OrderNr (PK, FK) 100100100101101102 ProductCode (PK, FK) WS WL PS PS PM XX Quantity 112123 ProductCode (PK) WS WL PS PM XX Material Wood Wood Plastic Plastic Extra-strong Size Small Large Small Medium Large UnitPrice 30.00 80.00 40.00 75.00 200.00

Mentor tip

Make sure to deduplicate the `PS` (Plastic, Small) entry in the `PRODUCTS` table so each `ProductCode` primary key is unique.

Question 4

20 marks

Tax office data warehousing, multi-dimensionality, deviation detection, and ETL processes

Databases

(a)

3 marks

Describe multi-dimensionality in a tax office data warehouse:
1. Multi-dimensionality organizes data into an OLAP hypercube (3D or higher-dimensional array) where each axis represents an analytical dimension.
2. One mandatory dimension in a data warehouse is `Time` (e.g., tax year, quarter).
3. Other dimensions categorize tax records by attributes such as `Taxpayer Type` (individual, corporation, trust) and `Geographic Region` (Region 1, Region 2), allowing analysts to slice and dice metrics across multiple axes.

Mentor tip

Explain the 3D/multi-axis data cube structure and name at least two specific dimensions alongside `Time`.

(b)

3 marks

Describe how data warehousing helps identify tax fraud:
1. Data Integration: The warehouse consolidates comprehensive historical taxpayer declarations with external datasets (such as property registries, vehicle registrations, bank interest reports, and census records).
2. Pattern & Discrepancy Analysis: Data mining queries cross-reference declared income against lifestyle indicators or industry benchmarks across time and region dimensions.
3. Fraud Profiling: For example, a taxpayer declaring minimal income while acquiring high-value real estate or reporting figures far below regional business peers is automatically flagged for audit.

Mentor tip

Mention combining internal tax records with external sources (property, credit, census) and drilling down by dimension/profile to spot mismatches.

(c)

3 marks

Describe how deviation detection analyses tax evasion:
1. Statistical Baseline Modelling: The system establishes expected statistical norms or predictive models for tax returns within a given profession, income bracket, or historical profile.
2. Outlier Identification: Deviation detection algorithms scan new declarations to pinpoint anomalies (outliers) that diverge significantly from the expected model.
3. Concrete Example: If a retail business suddenly claims expense deductions three standard deviations above its 5-year historical mean or sector average, the record is flagged as a suspicious deviation.

Mentor tip

Define deviation detection as statistical outlier detection against an expected model, and give a specific tax-evasion example.

(d)

3 marks

Determine whether a network-model DBMS can replicate data warehouse functionality:
1. Verdict: No, a network-model DBMS is not equally suited for data warehousing analytics.
2. Transactional vs Analytical Orientation: Operational DBMS models (including network databases) are optimized for high-speed current CRUD transactions (OLTP), whereas data warehouses are subject-oriented and optimized for complex multi-dimensional analytical queries (OLAP).
3. Temporal Dimension: Operational DBMSs typically overwrite records with current states and lack the built-in multi-year historical time dimension essential to a data warehouse.

Mentor tip

Contrast OLTP (current operational transactions in a network DBMS) with OLAP (historical, time-variant, subject-oriented analysis in a data warehouse).

(e)(i)

2 marks

Explain why the Transformation (T) stage is necessary in ETL:
1. Data extracted from disparate external systems (banks, regional tax offices, employers) arrives in incompatible schemas, encodings, and naming conventions.
2. Transformation cleanses, validates, deduplicates, and standardizes these heterogeneous records into a single uniform schema before loading (e.g., converting different date formats or standardizing gender/status codes).

Mentor tip

Highlight that heterogeneous source databases use different formats/encodings that must be standardized into a uniform warehouse schema.

(e)(ii)

6 marks

Outline Application 1 — Consolidating Historical Data / Archiving (3 marks):
1. ETL extracts aged records (such as completed tax returns or dispatched orders older than a set retention period) from multiple live operational databases.
2. The transformation step filters out unnecessary operational fields, cleanses corrupt or duplicate entries, and formats the data for long-term analytical queries.
3. Loading the cleaned records into a separate historical archive offloads the transactional system and simplifies maintenance.
Outline Application 2 — System Data Migration or Data Mart Creation (3 marks):
4. When migrating from a legacy database to a modern platform (or building a departmental Data Mart slice, such as for a marketing or audit team), ETL extracts data from the source tables.
5. Transformation maps legacy field types, resolves structural incompatibilities, and establishes new relational links.
6. The converted data is loaded into the target database or specialized Data Mart so department users can run targeted business intelligence queries efficiently.

Mentor tip

Develop two distinct ETL use cases in full detail (3 marks each): e.g., (1) database migration between different systems, (2) historical data consolidation/archiving, or (3) generating departmental data marts.

Option B: Modelling and simulation · Option

Questions 5–8 · 65 marks

Question 5

15 marks

Lake fish population sampling, spreadsheet modelling, and model validation

Modelling and simulation

(a)

4 marks

Outline the steps to aggregate and record the five annual lake samples:
1. Extract the fish count and individual fish length measurements from each of the five random samples.
2. Sum the fish counts across all five samples and divide by 5 (or scale by lake area ratio) to compute the representative annual population estimate.
3. Sum the lengths of all captured fish across the five samples and divide by the total number of fish measured to calculate the estimated mean length for the year.
4. Append the year's identifier, estimated population total, and estimated mean length as a new annual summary record in the historical dataset.

Mentor tip

Show clearly how both metrics (representative population count and mean fish length) are computed from the 5 samples and stored by year.

(b)

2 marks

Identify the core variables required for the licensing decision model:
1. `Year` (or `Date`), to track the 5-year rolling window.
2. `EstimatedFishCount` (annual fish population estimate) and `AverageFishLength` (estimated mean fish length in cm).

Mentor tip

Include the temporal variable (Year) alongside both measured metrics (Population Count and Average Length).

(c)

3 marks

Explain how to construct and evaluate the model in spreadsheet software:
1. Data Layout: In a spreadsheet, organize `Year` in Column A, `EstimatedFishCount` in Column B, and `AverageFishLength` in Column C, with one row per year (e.g., rows 1 to 6 covering a 5-year span).
2. Percentage Change Formula: In Column D (e.g., cell `D6`), compute the 5-year percentage change in fish population using `= 100 * (B6 - B1) / B1`.
3. Decision Rule: In Column E, use a conditional formula such as `=IF(AND(D6 >= -5, C6 > 20), "Issue Licences", "Deny Licences")` to automatically determine whether both criteria are satisfied.

Percentage Change (D6) = \frac{B _{6} - B _{1}}{B _{1}} \times 100

Mentor tip

Reference specific spreadsheet columns, cell formulas, and the exact threshold conditions (`>= -5%` change and `> 20 cm` average length).

(d)

6 marks

Discuss how to test and refine the data collection and model rules:
1. Testing Data Collection Accuracy: Increase the sampling frequency (more than 5 times a year) and sample across a wider variety of lake zones/depths to verify whether five random samples accurately reflect the true fish population.
2. Refining Sampling Methodology: If parallel samples show high variance (e.g., due to seasonal breeding or fish congregating in specific temperatures), standardize the sampling locations and times of year.
3. Evaluating Model Thresholds: Compare historical licensing decisions against subsequent multi-year population trends to check if the 5% drop tolerance or 20 cm length threshold is overly strict (supporting anglers' complaints) or too lenient (supporting conservationists' complaints).
4. Adjusting Model Parameters: Modify the percentage tolerance or minimum length threshold incrementally and incorporate additional variables (such as catch limits per licence or water quality) to achieve stable fish populations without unnecessary bans.

Mentor tip

Address both potential sources of error: flawed input data (sampling methodology) and flawed decision rules (the 5% and 20 cm thresholds).

Question 6

15 marks

Weather pattern representation, climate displacement detection, and long-term simulation

Modelling and simulation

(a)

2 marks

Identify four measurable daily weather factors:
1. Temperature (minimum, maximum, or mean daily temperature) and Precipitation (daily rainfall/snowfall in mm).
2. Wind speed/direction, relative humidity, atmospheric pressure, or hours of sunshine.

Mentor tip

List any four standard meteorological variables (two per mark).

(b)

3 marks

Suggest how to visualize annual weather patterns:
1. Plot each meteorological factor on its own chart (or normalized multi-axis chart) across a 12-month horizontal time axis.
2. Use line graphs for continuous variables (such as daily or monthly average temperature) and bar charts for cumulative totals (such as monthly precipitation).
3. Group or smooth the daily values into weekly or monthly averages so seasonal cycles and trends over the year are immediately visible.

Mentor tip

Specify the chart type (line graph or bar chart), the axes (time over one year vs weather metric), and separate plotting for distinct units.

(c)

4 marks

Outline how a computer system detects seasonal displacement trends:
1. Long-term Archiving: Store daily or monthly weather records continuously across many decades.
2. Curve Fitting: Fit a mathematical curve or seasonal baseline profile to each year's annual weather cycle.
3. Time-Shift Comparison: Compare the curves across successive years by sliding (translating) the seasonal profile along the time axis or computing month-by-month anomalies.
4. Trend Identification: Measure whether peak summer temperatures or onset dates consistently shift later in the calendar year over time.

Mentor tip

Explain how multi-year historical data is mathematically compared (curve fitting and phase-shift/translation analysis) to spot delayed seasons.

(d)

6 marks

Discuss advantages of computer climate simulations:
1. Policy and Agricultural Planning: Advance warning of long-term warming or seasonal shifts allows governments and farmers to plan infrastructure, water management, and crop changes.
2. What-If Experimentation: Scientists can repeatedly run simulations with altered parameters (e.g., varying greenhouse gas emission rates) and validate models against historical climate records.
Discuss disadvantages and limitations:
3. Incomplete and Chaotic Variables: Climate systems involve complex, non-linear feedback loops and unknown external factors that cannot be perfectly modelled, leading to uncertainty.
4. Error Compounding: Small inaccuracies in initial conditions or cyclical patterns can amplify over long time horizons, producing unreliable predictions.
Formulate a balanced conclusion:
5. Conclusion: Provided models are continually calibrated against real-world observations, the ability to anticipate and mitigate severe environmental disasters outweighs the inherent uncertainties of long-range forecasting.

Mentor tip

Include concrete advantages, model limitations, and a reasoned conclusion weighing predictive value against uncertainty.

Question 7

15 marks

3D visualization, lighting and rendering, hardware/software requirements, and online gaming impacts

Modelling and simulation

(a)

2 marks

Define computer visualization:
1. Visualization is the computational process of converting abstract or binary numerical data stored in memory
2. into a graphical, pixel-based image displayed on a screen so humans can visually interpret it.

Mentor tip

State clearly that binary/numerical data in memory is transformed into a visual pixel representation.

(b)

5 marks

Outline how memory data is rendered into a realistic 3D screen image:
1. Geometric & Colour Representation: Objects are stored in memory as 3D coordinates (wireframe polygons/vectors) or voxel/pixel maps with base surface textures and colours.
2. Lighting & Viewpoint Calculation: The system identifies the position and intensity of virtual light sources relative to the object's surfaces and the camera's viewing angle.
3. Ray Tracing / Shading Equations: Mathematical equations calculate surface reflection, absorption, shadows, and occlusion for each polygon.
4. Pixel Shading: Each 2D screen pixel's colour and brightness value is transformed into the appropriate gradient or shade.
5. Final Rendering: The shaded pixels are written to the frame buffer, producing depth cues and realistic 3D perspective on the 2D display.

Mentor tip

Trace the pipeline from memory representation -> light source & camera angle math -> shading/reflection -> 2D pixel colour mapping (rendering).

(c)

3 marks

Outline essential hardware and software specifications for online 3D gaming:
1. Hardware (Graphics & Memory): A dedicated Graphics Processing Unit (GPU / fast video card) with high VRAM, supported by ample system RAM and a fast multi-core CPU, to render complex 3D frames in real time.
2. Hardware (Audio & Network): A high-speed broadband network interface (for low-latency server communication) and a capable sound card.
3. Software: An operating system capable of rapid real-time I/O scheduling, paired with graphics/multimedia APIs (such as DirectX or OpenGL) and up-to-date GPU drivers.

Mentor tip

Cover both hardware (GPU, RAM, network/sound) and software (OS with fast I/O, graphics APIs like DirectX/OpenGL).

(d)

5 marks

Discuss benefits of participating in online simulated gaming worlds:
1. Social & Cognitive Benefits: Players can relax, collaborate with international peers, build online communities, and develop strategic problem-solving and spatial skills.
2. Safe Exploration: Virtual environments allow users to experience scenarios and team roles that are impossible or dangerous in real life.
Discuss drawbacks and risks:
3. Addiction & Neglect: Excessive immersion can lead to gaming addiction, causing players to neglect academic, professional, or physical health responsibilities.
4. Social Isolation: Over-reliance on virtual interactions can isolate individuals from face-to-face relationships and blur the boundary between virtual and real-world norms.
Provide a balanced conclusion:
5. Conclusion: When gaming time is moderated and balanced with offline responsibilities, online virtual worlds offer an engaging, beneficial form of modern entertainment and social connection.

Mentor tip

Include two distinct advantages, two distinct disadvantages, and a concluding judgement.

Question 8

20 marks

Neural network hidden units and weights, OCR vs speech recognition, NLP approaches, and genetic algorithms for timetabling

Modelling and simulation

(a)

3 marks

Outline the function of hidden units and weights in an artificial neural network:
1. Hidden units (internal layer nodes between the input and output layers) receive numerical activations from the input nodes.
2. Each incoming connection value is multiplied by an adjustable numerical weight that represents the strength/importance of that feature.
3. Each hidden unit sums its weighted inputs, applies an activation function, and forwards the signal to the output layer (having more hidden nodes enables the network to model more complex non-linear patterns).

Mentor tip

Trace the signal flow: input nodes -> multiplied by connection weights -> summed/processed at hidden layer nodes -> sent to output layer.

(b)

4 marks

Compare similarities between neural networks in speech recognition and OCR (2 marks):
1. Both systems are trained using labelled training datasets of sample patterns.
2. Both iteratively adjust internal connection weights during training to minimize error against a target fitness/loss function.
Contrast differences between speech recognition and OCR (2 marks):
3. In Optical Character Recognition (OCR), printed characters conform closely to fixed geometric shapes and font templates (with variance mainly limited to handwriting).
4. In speech recognition, acoustic inputs exhibit far greater continuous variability across speakers due to pitch, speed, accent, intonation, and background noise, making pattern matching more sensitive to deviations from the training set.

Mentor tip

Provide two clear similarities (training sets, weight adjustment against a fitness criteria) and two clear differences (geometric uniformity of printed text vs high acoustic variability of human voices).

(c)

2 marks

Distinguish between speech synthesis and natural language generation (NLG):
1. Speech synthesis (text-to-speech) is the mechanical/phonetic conversion of existing written text into audible spoken sound waves.
2. Natural language generation (NLG) is the semantic and syntactic process of constructing meaningful human-language sentences from structured underlying data (such as composing a spoken answer from database query results).

Mentor tip

Speech synthesis = converting pre-existing text to audio (text-to-speech); NLG = deciding what words/sentences to compose from raw data/meaning.

(d)

4 marks

Discuss the heuristic vs cognitive approaches to automated natural language learning:
1. Heuristic Approach: Relies on statistical probabilities, rule-of-thumb shortcuts, and supervised learning corpora to map inputs to likely outputs.
2. Heuristic Limitation: Its accuracy is bounded by the breadth and quality of its pre-labelled training sets.
3. Cognitive Approach: Uses unsupervised learning and semantic modelling to infer underlying grammar rules, context, and conceptual meaning.
4. Cognitive Characteristic: Aims to acquire syntax and semantics organically in a manner analogous to how a human child learns language.

Mentor tip

Contrast heuristic learning (statistical probabilities + supervised training sets) with cognitive learning (unsupervised acquisition of syntax and semantics like a child).

(e)

7 marks

Outline step-by-step how a genetic algorithm generates a valid school timetable:
1. Define Inputs & Constraints: Encode all scheduling rules and resources (teacher maximum hours, student subject choices, room capacities, and no double-booked rooms).
2. Establish Fitness Function & Priorities: Assign weights/priorities to hard constraints (must be met, e.g., no room clashes) and soft constraints to calculate a percentage fitness score for any generated timetable.
3. Initialize Population: Randomly generate an initial population of trial school timetables.
4. Evaluate Fitness: Score each timetable in the current generation using the fitness function.
5. Selection: Select the highest-scoring (fittest) timetables to act as parents for the next generation.
6. Crossover & Mutation: Combine schedules from parent timetables (crossover) and introduce small random slot swaps (mutation) to produce a new generation of offspring timetables.
7. Termination Condition: Repeat the evaluation, selection, and breeding cycle iteratively until a timetable achieves an tolerable fitness threshold (all mandatory constraints satisfied).

Mentor tip

Walk through all stages of a genetic algorithm applied to timetabling: encoding constraints -> weighted fitness function -> random initial population -> selection of fittest -> crossover & mutation -> repeat until threshold met.

Option C: Web science · Option

Questions 9–14 · 65 marks

Question 9

17 marks

HTML vs HTTP, XML data structures, DNS resolution, and web protocols vs standards

Web science

(a)

2 marks

Distinguish between HTML and HTTP:
1. HTML (HyperText Markup Language) is a markup language used to structure and format content for display in a web browser.
2. HTTP (HyperText Transfer Protocol) is a communications protocol that defines the rules for transmitting requests and web files between a client and a server.

Mentor tip

Contrast a markup language for page structure (HTML) with a network communication protocol for data transfer (HTTP).

(b)

1 mark

Identify one structural similarity between HTML and XML:
Both are markup languages that use angled-bracket tags (`<tag>...</tag>`) to enclose and structure elements.

Mentor tip

Mention the shared use of markup tags.

(c)

2 marks

Outline how the given XML snippet demonstrates self-describing data:
1. XML uses descriptive, human-readable custom tag names that explicitly label the meaning of the enclosed data.
2. For example, `<species>Bubo bubo</species>` and `<habitat>Forests</habitat>` directly identify that 'Bubo bubo' is a biological species and 'Forests' is its habitat within a `<bird>` record.

Mentor tip

Quote specific tags from the snippet (such as `<bird>`, `<name>`, or `<species>`) to satisfy 'direct reference to the XML code above'.

(d)

2 marks

Describe how XML's hierarchical structure facilitates database updates:
1. The root/parent element (`<bird>`) represents a database record/entity, while the nested child tags (`<name>`, `<species>`, `<description>`, `<habitat>`) correspond to specific field/column names.
2. The server can parse the XML tree and map each tag's text value directly into the matching column of the target database table.

Mentor tip

Connect XML parent elements to database records and child tag names to database table columns/fields.

(e)

2 marks

Identify the two actions a Domain Name Server (DNS) can take upon receiving a URL:
1. If the domain name is in its local cache/database, it resolves the domain name and returns the corresponding IP address to the client browser.
2. If it does not hold the mapping, it forwards the lookup request to another more authoritative DNS server (or returns an `NXDOMAIN` error if the domain does not exist).

Mentor tip

Cover both the cache hit case (return matching IP address) and the cache miss case (query higher-level/authoritative DNS server).

(f)(i)

2 marks

Distinguish between a protocol and a standard:
1. A protocol is a formal set of procedural rules governing how data is transmitted and how devices communicate across a network.
2. A standard is an agreed-upon technical specification or format (maintained by bodies like W3C or IEEE) that ensures consistency and compatibility across different hardware and software implementations.

Mentor tip

Protocols govern communication procedures/rules in transit; standards define agreed technical specifications/formats.

(f)(ii)

6 marks

Discuss the consequences of lacking a specific web protocol (3 marks):
1. Named Protocol: TCP/IP or HTTP.
2. Technical Problem if Absent: Without TCP/IP, routers and end-hosts would have no agreed rules for packetizing, addressing, routing, or reassembling data across heterogeneous networks.
3. Concrete Consequence: Packets sent from a Linux server to a Windows client would be dropped or misinterpreted, making reliable global data transfer impossible.
Discuss the consequences of lacking a specific web standard (3 marks):
4. Named Standard: HTML, CSS, Unicode (UTF-8), or JPEG.
5. Technical Problem if Absent: Without a W3C HTML standard, every browser vendor and web developer would invent proprietary markup syntax.
6. Concrete Consequence: A webpage written for one browser (e.g., Chrome) would fail to render or display garbled code in another browser (e.g., Safari or Firefox), fragmenting the World Wide Web.

Mentor tip

Explicitly name one protocol (e.g., TCP or HTTP) and one standard (e.g., HTML or Unicode), and explain the specific technical breakdown that would occur without each.

Question 10

12 marks

PageRank algorithm, link weighting factors, meta-tag limitations, and black hat SEO

Web science

(a)

2 marks

Identify the specific purpose of Google's PageRank algorithm:
1. To calculate a numerical authority/importance score for each indexed web page
2. so that search results for a specific query can be ranked in order of relevance and importance.

Mentor tip

Mention both assigning a numerical importance rating to a page and ordering search results for a user's query.

(b)

2 marks

Identify two factors determining how much Page A's link boosts Page B's PageRank:
1. The existing PageRank (authority score) of Page A itself.
2. The total number of outbound links (`outlinks`) on Page A, since Page A's authority is divided among all pages it links to.

Mentor tip

Looking at the diagram, notice Page A has both incoming links (affecting Page A's own rank) and multiple outgoing links (diluting the share passed to Page B).

(c)

2 marks

Explain why meta-tags may return unexpected search results:
1. Polysemy / Ambiguity: Words placed in meta-tags often have multiple unrelated meanings (homonyms) without semantic context.
2. Example: A page tagged with `meta="star"` about astrophysics may be returned when a user searches for a celebrity movie 'star' (or webmasters may insert misleading/inaccurate keywords in meta-tags).

Mentor tip

Use a concrete example of an ambiguous keyword (like 'star', 'apple', or 'jaguar') to show why keyword matching alone can fail.

(d)

6 marks

Explain three black hat Search Engine Optimization (SEO) techniques:
1. Keyword Stuffing: Repetitively packing popular search terms into `<meta>` tags or page text so the indexing algorithm artificially scores the page as highly relevant to those queries.
2. Invisible / Hidden Text: Placing extra keywords on the page in the same colour as the background, in tiny font sizes, off-screen via CSS, or inside HTML comments so web crawlers index them while human visitors cannot see them.
3. Link Farms / Spamdexing: Creating networks of dummy websites or automated blog comments that all link back to the target page to artificially inflate its inbound link count and manipulate PageRank.

Mentor tip

Name and explain three distinct black-hat practices: keyword stuffing, hidden text/cloaking, and link farms.

Question 11

16 marks

Lossy compression file types, open standards vs interoperability, Web 2.0 social networking, and cloud security

Web science

(a)

2 marks

Identify two file types suitable for lossy compression:
1. Digital audio files (e.g., MP3, AAC).
2. Digital image/photograph files (e.g., JPEG) or video files (e.g., MP4).

Mentor tip

Choose multimedia formats (images, audio, video) where minor perceptual data loss is tolerable, never text or executable files.

(b)

4 marks

Outline the difference between open standards and interoperability with examples:
1. Open Standards: Publicly available, royalty-free technical specifications that anyone can inspect and implement without restriction.
2. Example of Open Standard: HTML, CSS, or HTTP published by the W3C/IETF.
3. Interoperability: The practical ability of diverse hardware devices, operating systems, and software applications from different vendors to exchange data and work together seamlessly.
4. Example of Interoperability: Because HTML is an open standard, a webpage hosted on a Linux Apache server can be rendered identically by Safari on iOS, Chrome on Android, and Edge on Windows.

Mentor tip

Define open standards as the publicly documented specifications, and interoperability as the resulting ability of different systems to work together.

(c)

4 marks

Describe two Web 2.0 technological features that enabled social networking:
1. User-Generated Content & Rich Interactivity (AJAX/DOM scripting): Dynamic client/server scripting allowed non-technical users to post status updates, comment on friends' walls, and personalize profiles in real time without writing HTML.
2. Seamless Multimedia Integration & Instant Messaging: High-speed browser media players, file-upload APIs, and asynchronous chat allowed users to share photos/videos and communicate instantaneously within the browser.

Mentor tip

State each Web 2.0 feature (e.g., dynamic user-generated content, multimedia sharing, real-time messaging) and explain how it fostered social interaction.

(d)

6 marks

Discuss security considerations of a third-party cloud service vs an in-house server:
1. Third-Party Cloud Advantages: Professional cloud providers employ dedicated security specialists, enterprise-grade firewalls, automated encryption, and redundant off-site backups that a small company could not afford on its own.
2. Third-Party Cloud Risks: The small company relinquishes direct physical control over customer data; internal cloud operations are opaque, and transmitting data over the internet introduces interception risks.
3. In-House Server Advantages: All customer records remain physically on-premises under the direct oversight of trusted company staff, avoiding third-party multi-tenancy exposure.
4. In-House Server Risks: A small business typically lacks a specialized 24/7 cybersecurity team, leaving local servers vulnerable to unpatched exploits, physical theft, or local disasters.
5. Balanced Conclusion: Unless regulatory rules mandate on-premises storage, a reputable private cloud provider generally delivers stronger, more up-to-date security than a small company's limited internal IT staff can maintain.

Mentor tip

Compare both options (Cloud storage vs Company in-house storage) in detail and provide a reasoned conclusion to earn full marks.

Question 12

12 marks

Directed web graphs, strongly connected core (SCC), bowtie IN/OUT sections, and recursive web crawling

Web science

(a)(i)

2 marks

Identify the web graph components in the diagram:
1. The nodes (vertices `A` through `F`) represent individual web pages.
2. The directed edges (arrows connecting the nodes) represent hyperlinks pointing from a source page to a target page.

Mentor tip

Explicitly state what both the nodes (web pages) and the directed arrows (hyperlinks and their direction) represent.

(a)(ii)

2 marks

Explain why these nodes cannot form a Strongly Connected Core (SCC):
1. In an SCC, every node must be reachable from every other node by following directed links.
2. In the given graph, Node `A` and Node `F` have no incoming links (in-degree = 0), so it is impossible to navigate back to Page `A` or Page `F` from any other page.

Mentor tip

Point out specific unreachable nodes (`A` or `F` have only outgoing arrows and zero incoming arrows) to prove mutual reachability fails.

(b)

4 marks

Distinguish between the IN and OUT sections of the bowtie web graph with examples:
1. IN Section Features: Contains web pages that have outbound links leading into the Strongly Connected Core (SCC), but cannot be reached from the SCC (only outlinks towards the core, no inlinks from the core).
2. IN Section Example: Newly published web pages or personal blogs that link to major websites before any external sites have linked back to them.
3. OUT Section Features: Contains web pages that can be reached via links from the SCC, but do not contain links leading back into the SCC (effectively dead-ends / inlinks only with respect to the core).
4. OUT Section Example: Static corporate PDF documents, reference archives, or government policy pages that receive links from the web but do not link back out.

Mentor tip

Contrast the link direction relative to the SCC for both IN and OUT, and give a concrete webpage example for each.

(c)

4 marks

Explain the recursive traversal algorithm used by web crawlers:
1. Fetch & Store: The crawler visits the current URL, downloads a copy of the web page for indexing, and marks the URL as visited.
2. Parse for Outlinks: The algorithm scans through the page's HTML source code sequentially looking for unvisited hyperlink URLs (`<a href="...">`).
3. Recursive Call (Depth-First Jump): When it encounters an unvisited URL, it invokes the crawl procedure recursively on that new target URL.
4. Base Case & Backtracking: When a page contains no further unvisited links (a dead end), the current call terminates and returns control to the previous caller URL to resume scanning remaining links until all reachable pages are visited.

Mentor tip

Clearly describe all four aspects: fetching the current page, scanning for links, jumping recursively to a found link, and backtracking to the parent page upon hitting a dead end.

Question 13

2 marks

Ontological systems for integrating heterogeneous publisher book data

Web science

2 marks

Describe how an ontology resolves heterogeneous field labels across book publishers:
1. An ontology defines a formal, shared vocabulary and semantic mapping of concepts and relationships within a domain.
2. When different publishers use conflicting tag names for the same attribute (e.g., `<Publisher>`, `<Publishing_House>`, or `<Imprint>`), the ontology maps all synonymous tags onto a single standardized schema concept so the bookshop's database can automatically merge records without manual reformatting.

Mentor tip

Define an ontology as a standardized semantic vocabulary/mapping and give a concrete example of resolving two different publisher tag names into one unified field.

Question 14

6 marks

Ambient intelligence technologies and workplace surveillance ethics

Web science

6 marks

Describe the ambient intelligence hardware and network infrastructure (3 marks):
1. Wearable/Embedded Identifiers: Employees carry ID badges or devices embedded with active RFID tags, Bluetooth Low Energy (BLE) beacons, or biometric sensors.
2. Distributed Sensor Network: Unobtrusive wireless readers/receivers installed throughout doorways, corridors, and offices detect the badges automatically without manual check-ins.
3. Central Real-Time Processing: The sensors transmit location coordinates wirelessly to a central server that logs and models employee positions in real time.
Discuss the ethical implications for employers and employees (3 marks):
4. Employer Perspective (Productivity & Safety): Management may argue that since staff are compensated during working hours, tracking location ensures accountability, optimizes building security, and locates personnel rapidly during emergency evacuations.
5. Employee Perspective (Privacy & Autonomy): Continuous spatial tracking infringes on personal privacy, creates a high-stress culture of micromanagement and distrust (e.g., monitoring washroom breaks), and exposes private workplace social interactions or union associations.

Mentor tip

Dedicate half your answer to the specific hardware/network mechanism (RFID/sensor badges + wireless room receivers + central tracking server) and half to weighing employer accountability/safety against employee privacy.

Option D: Object-oriented programming · Option

Questions 15–18 · 65 marks

Question 15

19 marks

College library OOP advantages, developer ethics, Student and Loan encapsulation, and static variables

Object-oriented programming

(a)

4 marks

Outline two advantages of using an OOP approach for the development team:
1. Parallel Modular Development: Because the team is divided into small sub-groups, each sub-group can independently develop and test distinct classes (such as `Student`, `Loan`, `shortTerm`) simultaneously, speeding up delivery.
2. Code Reusability & Easier Maintenance: Objects and methods developed by one sub-group can be reused across the project without rewriting code, and isolated class modules are faster to debug and unit-test.

Mentor tip

Link OOP features (modularity, reuse, independent testing) directly to the scenario of a team split into small sub-groups.

(b)

4 marks

Explain two ethical responsibilities of the software development team:
1. Rigorous Testing for Fairness and Accuracy: Developers have an ethical duty to thoroughly test fine calculations and suspension rules so students are not unfairly overcharged or wrongly barred from borrowing academic resources.
2. Data Privacy & Intellectual Property: Developers must protect confidential student borrowing records from unauthorized disclosure and properly acknowledge/license any external code libraries used to avoid plagiarism or copyright infringement.

Mentor tip

State each ethical issue (e.g., thorough testing of fine/suspension logic, protecting student privacy, or acknowledging code sources) and explain why it matters.

(c)

3 marks

Outline how encapsulation enhances security in the `Student` class:
1. In `Student`, attributes such as `private int studentID` and `private String studentName` are declared with the `private` access modifier.
2. This hides the internal state so external classes cannot directly read or accidentally corrupt student data.
3. External code can only interact with these fields through controlled `public` accessor/mutator methods (`getStudentID()`, `addLoan()`).

Mentor tip

Explicitly reference the `private` keyword on attributes like `studentID` or `booksBorrowed` in the `Student` class.

(d)

2 marks

Construct the Java statement to instantiate the `Student` object:
1. Declare a `Student` reference variable and invoke the `new` operator.
2. Pass the `int` ID `93003` and `String` name `"Smith"` to the `Student` constructor:

Student newStudent = new Student(93003, "Smith");

Mentor tip

Match the exact parameter order of the constructor `public Student(int studentID, String studentName)`.

(e)

3 marks

Explain the purpose of `static` in `static int numBooksLoaned = 0;`:
1. A `static` variable belongs to the `Loan` class itself rather than to any individual `Loan` instance, so only a single shared copy exists in memory.
2. It is initialized once when the class loads, not reset to zero each time a new `Loan` object is instantiated.
3. Because the constructor increments `numBooksLoaned` every time a `Loan` is created, it tracks the cumulative total of all books borrowed across the entire library.

Mentor tip

Contrast a class-level variable (one shared copy across all instances) with an instance variable, and explain its role as a global loan counter.

(f)

3 marks

Construct the Java code to add the specified `Loan` to `Student` object `ST`:
1. Instantiate a new `Loan` object with `new Loan(212000, "The Stars")`.
2. Pass that newly created `Loan` instance into `ST.addLoan(...)`:

ST.addLoan(new Loan(212000, "The Stars"));

Mentor tip

You can either instantiate the `Loan` inline inside `ST.addLoan(...)` or assign it to a temporary `Loan` variable first and pass that variable to `ST.addLoan(...)`.

Question 16

15 marks

Tracing Student/Loan arrays, constructing showDetails(), and addressing missing requirements

Object-oriented programming

(a)(i)

1 mark

Trace `borrowers[93001].getStudentName()`:
`borrowers[93001]` references the `Student` initialized with ID `93001` and name `"Jones"`. Output: `Jones`.

Mentor tip

Look up index 93001 in the `borrowers` array.

(a)(ii)

1 mark

Trace `borrowers[93001].getLoan(1).getBookTitle()`:
Student `93001` had `"The Sky"` added at loan index `0`, and `"The Spooks"` added second at loan index `1`. Output: `The Spooks`.

Mentor tip

Remember Java arrays are 0-indexed: `getLoan(0)` is the first loan (`The Sky`), so `getLoan(1)` is the second loan (`The Spooks`).

(a)(iii)

1 mark

Trace `borrowers[3012].getLoan(0).getBookTitle()`:
Student `3012` (`Zang`) had `"The Animals"` added as their first loan (index `0`). Output: `The Animals`.

Mentor tip

Index `0` retrieves the first `Loan` added to `borrowers[3012]`.

(b)

7 marks

Construct the `showDetails(int studentID)` method:
1. Define a `void` method taking `int studentID` as a parameter.
2. Retrieve the `Student` object from `borrowers[studentID]` and output `getStudentName()`.
3. Loop through the `booksBorrowed` array from index `0` to `9` (`i < 10`).
4. Retrieve each `Loan` via `s.getLoan(i)`, check that it is not `null`, and print `getBookTitle()` and `getDate()`.
```java public void showDetails(int studentID) { Student s = borrowers[studentID]; System.out.println(s.getStudentName()); for (int i = 0; i < 10; i++) { Loan currentLoan = s.getLoan(i); if (currentLoan != null) { System.out.println(currentLoan.getBookTitle()); System.out.println(currentLoan.getDate()); } } } ```

Mentor tip

Check `if (currentLoan != null)` before calling `.getBookTitle()` and `.getDate()` because a student may have borrowed fewer than 10 books.

(c)

3 marks

State three user requirements missing from the current class design:
1. No distinction between short-term (5-day) and long-term (30-day) book loans.
2. No attributes or methods to calculate overdue fines ( $1.50/ d a y f or s h or t - t er m v s$ 0.10/day for long-term).
3. No flag or validation logic to suspend a student's borrowing rights when a short-term loan is overdue by more than 4 days (nor a boundary check preventing `addLoan` from exceeding 10 books).

Mentor tip

Compare the bulleted user requirements at the start of Option D against the attributes/methods actually present in `Student` and `Loan`.

(d)

2 marks

Describe a modification to the `Student` class to address one missing requirement:
1. Add a `private boolean canBorrow = true;` (or `borrowingSuspended`) attribute to `Student` along with getter/setter methods.
2. Alternatively, modify `addLoan(Loan book)` to check `if (numBooks < 10 && canBorrow)` before inserting a new loan into `booksBorrowed`.

Mentor tip

Specify both the new attribute/check added to `Student` and how its value changes or governs borrowing.

Question 17

11 marks

Inheritance hierarchy, shortTerm/longTerm subclasses, UML class diagram, and static counters

Object-oriented programming

(a)

2 marks

Describe one benefit of inheritance in OOP:
1. Inheritance promotes code reuse and eliminates duplication
2. by allowing subclasses (`shortTerm` and `longTerm`) to automatically inherit common attributes (`bookID`, `bookTitle`, `d`) and methods from the `Loan` superclass.

Mentor tip

Connect reduced code duplication to subclasses inheriting shared fields and methods from their superclass.

(b)

2 marks

State two additional attributes required in the `shortTerm` subclass:
1. `private double dailyFineRate;` (or `fine`, e.g., `$1.50`).
2. `private int maxLoanDays;` (borrowing limit of 5 days) or `private boolean causesSuspension;` / `maxOverdueDays` (4 days).

Mentor tip

Pick any two attributes specific to short-term loans: daily fine rate, loan duration limit, or overdue suspension threshold.

(c)

4 marks

Construct the UML class diagram showing relationships among `Student`, `Loan`, `shortTerm`, and `longTerm`:
1. Draw four class boxes labelled `Student`, `Loan`, `shortTerm`, and `longTerm`.
2. Draw a `has-a` association/aggregation line between `Student` and `Loan` (a `Student` holds an array of `Loan` objects).
3. Draw `is-a` inheritance arrows (hollow triangle arrowheads pointing to the superclass) from both `shortTerm` and `longTerm` up to `Loan`.
``` [Student] <>---- (has-a) ----> [Loan] ^ | (is-a / inherits) +---------+---------+ | | [shortTerm] [longTerm] ```

Mentor tip

Clearly distinguish the 'has-a' relationship (`Student` to `Loan`) from the 'is-a' inheritance relationships (`shortTerm` and `longTerm` pointing to `Loan`).

(d)

3 marks

Outline how to display the total number of books on loan without traversing `borrowers`:
1. Declare a `static` counter variable (or use `Loan.numBooksLoaned`, or `static` counters in `shortTerm` and `longTerm`) that increments whenever a loan is instantiated and decrements when a book is returned.
2. Provide a `public static` accessor method in `Loan` (or sum the subclass static counters).
3. Call this `static` method directly on the class to output the total count in $O (1)$ time without looping through `borrowers`.

Mentor tip

Use `static` class variables and a `static` method so the count is maintained at the class level rather than by summing array elements.

Question 18

20 marks

Searching borrowed books, binary search trees for fast lookup, and replacing arrays with linked lists

Object-oriented programming

(a)

4 marks

Explain why locating who borrowed a specific book is slow in the current design:
1. Encapsulated Unindexed Arrays: Borrowed `Loan` objects are scattered across individual `booksBorrowed` arrays inside separate `Student` objects within the `borrowers` array.
2. Sparse Array Traversal: The `borrowers` array has 100,000 slots (most of which are `null`), and each student's `booksBorrowed` array has 10 slots (many of which are also `null`).
3. Exhaustive Linear Search: Because there is no index keyed by `bookID`, the system must perform a nested linear search across every non-null `Student` and every non-null `Loan` until a matching `bookID` is found.
4. High Time Complexity: This exhaustive scan results in poor $O (N \times M)$ performance.

Mentor tip

Explain that loans are distributed inside individual `Student` objects with no `bookID` index, forcing a slow linear search across sparse arrays.

(b)(i)

2 marks

Describe a suitable course data structure for fast book lookup:
1. A Binary Search Tree (BST) of nodes representing active book loans,
2. ordered/indexed by the unique `bookID` key, where each node stores both the `bookID` and the borrowing student's `studentID`.

Mentor tip

Name a Binary Search Tree (BST) and specify that `bookID` is used as the ordering key.

(b)(ii)

4 marks

Outline the steps to populate and search the Binary Search Tree:
1. Insertion on Loan: Whenever a book is borrowed, insert a new node containing `(bookID, studentID)` into the BST ordered by `bookID`.
2. Binary Search Traversal: To find who borrowed a target `bookID`, start at the root node and compare the target `bookID` with the current node's `bookID`, branching left if smaller or right if larger ( $O (lo g n)$ average comparisons).
3. Retrieve Student ID: Once the matching `bookID` node is reached, extract the stored `studentID`.
4. Direct Array Access: Use `borrowers[studentID]` to immediately access the corresponding `Student` object and output their details.

Mentor tip

Cover both how nodes `(bookID, studentID)` are inserted when loaned and how the tree is searched by `bookID` to retrieve `studentID`.

(c)

4 marks

Describe one advantage of replacing the `borrowers` array with a linked list (2 marks):
1. Dynamic Memory Efficiency: A linked list allocates memory nodes dynamically only for currently enrolled students.
2. Eliminates Wasted Space: Avoids reserving a fixed 100,000-element `borrowers` array where the vast majority of indices remain `null`.
Describe one disadvantage of using a linked list (2 marks):
3. Loss of $O (1)$ Direct Indexing: You can no longer access a student directly via `borrowers[studentID]`.
4. Slower Sequential Lookup: Finding a specific student requires traversing the linked list sequentially from the `head` pointer via a linear search ( $O (n)$ time).

Mentor tip

Give one clear advantage with elaboration (dynamic sizing avoids wasted memory on empty array slots) and one clear disadvantage with elaboration (sequential traversal replaces instant index lookup).

(d)

6 marks

Construct the `addToEnd()` algorithm to append a `shortTerm` loan node to the end of a linked list:
1. Instantiate or receive the `shortTerm` node to be inserted and set its `next` pointer to `null`.
2. Obtain the `head` reference of the linked list.
3. Handle the empty-list edge case: if `head == null`, set `head` to the new `shortTerm` node.
4. Otherwise, initialize a temporary traversal pointer `temp = head`.
5. Loop `while (temp.getNext() != null)` to advance `temp = temp.getNext()` until the last node is reached.
6. Link the last node's `next` pointer to the new `shortTerm` object (`temp.setNext(newBook)`).
```java public void addToEnd(shortTerm newBook) { newBook.setNext(null); if (head == null) { head = newBook; } else { Node temp = head; while (temp.getNext() != null) { temp = temp.getNext(); } temp.setNext(newBook); } } ```

Mentor tip

Include initializing `temp` to `head`, iterating until the last node (`temp.next == null`), setting the last node's `next` pointer to the new `shortTerm` object, and ensuring the new node's pointer is `null`.

Our analysis

Where the marks are

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

Databases65 marks · 25%
Q1, Q2, Q3, Q4
Modelling and simulation65 marks · 25%
Q5, Q6, Q7, Q8
Web science65 marks · 25%
Q9, Q10, Q11, Q12, Q13, Q14
Object-oriented programming65 marks · 25%
Q15, Q16, Q17, Q18

Mentor insights

Traps to watch for

01
Databases: Shelving unit orders normalization from unnormalized form to 3NF · Q3
Check the three classic 1NF requirements: atomic values in every cell, no repeating groups/blank multi-row cells, and a defined primary key.
02
Modelling and simulation: Neural network hidden units and weights, OCR vs speech recognitio · Q8
Trace the signal flow: input nodes -> multiplied by connection weights -> summed/processed at hidden layer nodes -> sent to output layer.
03
Web science: HTML vs HTTP, XML data structures, DNS resolution, and web protoc · Q9
Contrast a markup language for page structure (HTML) with a network communication protocol for data transfer (HTTP).
04
Object-oriented programming: Searching borrowed books, binary search trees for fast lookup, an · Q18
Explain that loans are distributed inside individual `Student` objects with no `bookID` index, forcing a slow linear search across sparse arrays.

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 HL · Paper 2

Computer Science, Higher Level

Session: May 2015
Marks: 260
Duration: 1 hour 20 min
Calculator: Not allowed
Questions: 18

Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Q15 Q16 Q17 Q18

Keep your copy of the paper open alongside. We don't reproduce the questions: each part below is labelled to match the May 2015 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 1–4 · 65 marks

Question 1

15 marks

Hotel booking database transactions, data versus information, isolation, and data sharing

Databases

(a)

2 marks

Distinguish raw data from processed information in the hotel booking context:
Data consists of unorganized, unprocessed values lacking context on their own, such as the isolated booking string '3EP9RE'.
Information is structured, interpreted data that carries meaning for the recipient, such as the follow-up confirmation email linking the code '3EP9RE' to the specific reservation details (double non-smoking room at the Overlook Hotel).

Mentor tip

Contrast raw symbols (the alphanumeric booking code alone) with contextualized output (the detailed confirmation message).

(b)

2 marks

Define a database transaction:
A database transaction is a single logical unit of work comprising a minimal sequence of read/write operations executed on a database.
It must execute completely or not at all (atomicity) to maintain database integrity, prevent conflicting room reservations, and allow recovery if a failure occurs.

Mentor tip

Emphasize that a transaction is an indivisible unit of work that preserves database consistency and supports rollback on failure.

(c)

3 marks

Describe how transactions enforce data consistency:
1. Atomicity and Persistence: All modifications within the booking transaction become permanent (committed) only if every step in the sequence terminates successfully.
2. Rollback on Failure: If any hardware, network, or logical failure interrupts the transaction prior to completion, a rollback operation undoes all partial changes.
3. State Preservation: Rolling back restores the database to its prior consistent state, ensuring rooms are neither double-booked nor marked unavailable without a valid payment and confirmation.

Mentor tip

Link transaction completion (commit) and failure recovery (rollback) directly to preserving a valid, consistent state.

(d)

2 marks

Explain transaction isolation in the room reservation scenario:
Isolation (concurrency control) ensures that intermediate states of an ongoing transaction are invisible to other concurrent operations until the transaction commits.
For example, while one customer is booking a specific double room, record locking prevents a second customer from simultaneously reserving the same room and causing a double-booking inconsistency.

Mentor tip

Include both the theoretical definition of isolation (concurrent operations cannot interfere) and a concrete example (locking a room record during booking).

(e)

6 marks

Explain the technical mechanism of data sharing:
1. Data sharing and replication expose selected portions of source hotel databases to authorized external systems or user groups.
2. Bringing relevant data closer to end users reduces redundant network transmission overhead and allows heterogeneous systems to query standardized views.
Explain benefits for Stakeholder 1 (Tourists / End Users):
3. Tourists can rapidly query and compare room availability, features, and pricing across numerous hotels from a single unified interface.
4. Real-time access prevents wasted effort attempting to book rooms that are already occupied.
Explain benefits for Stakeholder 2 (Best Fare Aggregation Service / Hotels):
5. Best Fare can dynamically aggregate offers from multiple hotels and scale its platform to new partner hotels without redesigning its core software.
6. Individual hotels gain broader market visibility and can update local room rates or availability in real time while maintaining control over their internal operational records.

Mentor tip

Cover the technical nature of database sharing/replication first, then provide two distinct points for each of two stakeholders (e.g., tourists and the comparison platform or hotel).

Question 2

10 marks

Hospital pharmacy database administration, data dictionaries, and user views

Databases

(a)

2 marks

Outline the responsibilities of the hospital's Database Administrator (DBA):
1. The DBA manages the operational performance, security, and structural integrity of the pharmacy database (e.g., configuring regular backups and recovery procedures).
2. The DBA controls access permissions, authorizing internal medical staff and partner hospitals to view or modify only the specific records appropriate to their clinical roles.

Mentor tip

Focus on administrative duties such as access control, backup management, schema maintenance, and security rather than day-to-day data entry.

(b)

2 marks

Define a data dictionary:
1. A data dictionary is a centralized repository that stores metadata (data about the data) for the database.
2. It defines field names, data types, validation rules, format constraints, and the relationships between tables.

Mentor tip

Mention 'metadata' and give concrete examples of what is stored (e.g., field types, constraints, and table relationships).

(c)

6 marks

Discuss why customized database views are required for different user groups:
1. Database Administrator (DBA): Requires an unrestricted view of the entire schema, data dictionary, and user tables to perform maintenance, structural updates, and security audits.
2. Internal Hospital Staff (Doctors, Pharmacists, Clerks): Require tailored read/write views depending on their clinical duties—doctors need access to patient prescription histories, while pharmacy clerks need views focused on stock quantities and supplier orders.
3. External Staff at Nearby Hospitals: Must be restricted to a read-only view showing only medicine stock availability (for emergency inter-hospital borrowing) while strictly hiding confidential patient medical records to comply with privacy laws.

Mentor tip

Address all three user groups mentioned in the prompt (DBA, internal staff, nearby hospital staff), stating both what they can see/do and why.

Question 3

20 marks

Shelving unit orders normalization from unnormalized form to 3NF

Databases

(a)

3 marks

Identify three reasons why the given table violates First Normal Form (1NF):
1. No primary key is defined to uniquely identify each row.
2. Repeating groups / blank cells exist: a single `OrderNr` and `TotalPrice` span multiple `ProductCode` rows, leaving empty attributes in subsequent rows.
3. Non-atomic values: the `Description` attribute combines two distinct pieces of data (material type and unit size, e.g., 'Wood, Small') in a single cell.

Mentor tip

Check the three classic 1NF requirements: atomic values in every cell, no repeating groups/blank multi-row cells, and a defined primary key.

(b)

3 marks

State three advantages of database normalization:
1. Minimizes data redundancy, reducing overall storage space requirements.
2. Eliminates update, insertion, and deletion anomalies, making data maintenance simpler and preventing inconsistencies.
3. Enforces atomic attributes and clear primary/foreign keys, which simplifies writing accurate queries and applying granular security controls.

Mentor tip

List three distinct benefits: reduced redundancy/storage, avoidance of modification anomalies, and simpler querying/security.

(c)

2 marks

Evaluate whether `TotalPrice` can serve as a primary key:
1. No, `TotalPrice` cannot be used alone as a primary key.
2. A primary key must be strictly unique for every record, whereas two different orders (Order 100 and Order 101) share the exact same `TotalPrice` of 190.00.

Mentor tip

Cite the specific duplicate value (`190.00` appearing for both Order 100 and Order 101) to prove why uniqueness fails.

(d)

2 marks

Propose a structural change to make `Description` atomic:
1. Decompose the `Description` column into two separate attributes.
2. Store the material type in one column (e.g., `Material`: Wood, Plastic, Extra-strong/Steel) and the dimension in a second column (e.g., `Size`: Small, Medium, Large).

Mentor tip

Explicitly name the two new columns that replace `Description`.

(e)

3 marks

Construct the 1NF relation(s) with atomic attributes and primary keys:
1. Split the repeating order items or flatten the table, and divide `Description` into `Material` (`Description1`) and `Size` (`Description2`).
2. Define the `ORDERS` table with primary key `OrderNr`: `ORDERS(OrderNr [PK], OrderDate, TotalPrice)` containing rows `(100, 10/10/2014, 190.00)`, `(101, 11/10/2014, 190.00)`, `(102, 11/10/2014, 600.00)`.
3. Define the `ORDER_ITEMS` table with composite primary key `(OrderNr, ProductCode)`: `ORDER_ITEMS(OrderNr [PK], ProductCode [PK], Material, Size, Quantity, UnitPrice)` containing all six individual item rows with `OrderNr` filled in for every row.

OrderNr (PK) 100101102 OrderDate 10/10/2014 11/10/2014 11/10/2014 TotalPrice 190.00 190.00 600.00 OrderNr (PK) 100100100101101102 ProductCode (PK) WS WL PS PS PM XX Material Wood Wood Plastic Plastic Plastic Extra-strong Size Small Large Small Small Medium Large Quantity 112123 UnitPrice 30.00 80.00 40.00 40.00 75.00 200.00

Mentor tip

Always clearly mark the Primary Key (`PK`) and Composite Primary Key (`OrderNr, ProductCode`) when constructing normalized tables.

(f)

2 marks

State the Second Normal Form (2NF) dependency condition:
1. Every non-key attribute in a table must be fully functionally dependent on the entire composite primary key.
2. No non-key attribute may have a partial dependency on only part of a composite primary key.

Mentor tip

State clearly that non-key attributes must depend on the whole composite key, not just a subset of it.

(g)

5 marks

Remove partial and transitive dependencies to reach 3NF:
1. Identify partial dependencies in `ORDER_ITEMS`: `Material`, `Size`, and `UnitPrice` depend solely on `ProductCode`, whereas `Quantity` depends on the full composite key `(OrderNr, ProductCode)`.
2. Separate the product details into a dedicated `PRODUCTS` table with primary key `ProductCode`, removing duplicate product rows (`PS` appears once).
3. Keep `ORDER_ITEMS` with composite primary key `(OrderNr, ProductCode)` and non-key attribute `Quantity`.
4. Keep `ORDERS` with primary key `OrderNr`, `OrderDate`, and `TotalPrice`.
5. Verify 3NF: in all three tables (`ORDERS`, `ORDER_ITEMS`, `PRODUCTS`), every non-key attribute depends solely on its table's primary key with no transitive dependencies between non-key fields.

OrderNr (PK) 100101102 OrderDate 10/10/2014 11/10/2014 11/10/2014 TotalPrice 190.00 190.00 600.00 OrderNr (PK, FK) 100100100101101102 ProductCode (PK, FK) WS WL PS PS PM XX Quantity 112123 ProductCode (PK) WS WL PS PM XX Material Wood Wood Plastic Plastic Extra-strong Size Small Large Small Medium Large UnitPrice 30.00 80.00 40.00 75.00 200.00

Mentor tip

Make sure to deduplicate the `PS` (Plastic, Small) entry in the `PRODUCTS` table so each `ProductCode` primary key is unique.

Question 4

20 marks

Tax office data warehousing, multi-dimensionality, deviation detection, and ETL processes

Databases

(a)

3 marks

Describe multi-dimensionality in a tax office data warehouse:
1. Multi-dimensionality organizes data into an OLAP hypercube (3D or higher-dimensional array) where each axis represents an analytical dimension.
2. One mandatory dimension in a data warehouse is `Time` (e.g., tax year, quarter).
3. Other dimensions categorize tax records by attributes such as `Taxpayer Type` (individual, corporation, trust) and `Geographic Region` (Region 1, Region 2), allowing analysts to slice and dice metrics across multiple axes.

Mentor tip

Explain the 3D/multi-axis data cube structure and name at least two specific dimensions alongside `Time`.

(b)

3 marks

Describe how data warehousing helps identify tax fraud:
1. Data Integration: The warehouse consolidates comprehensive historical taxpayer declarations with external datasets (such as property registries, vehicle registrations, bank interest reports, and census records).
2. Pattern & Discrepancy Analysis: Data mining queries cross-reference declared income against lifestyle indicators or industry benchmarks across time and region dimensions.
3. Fraud Profiling: For example, a taxpayer declaring minimal income while acquiring high-value real estate or reporting figures far below regional business peers is automatically flagged for audit.

Mentor tip

Mention combining internal tax records with external sources (property, credit, census) and drilling down by dimension/profile to spot mismatches.

(c)

3 marks

Describe how deviation detection analyses tax evasion:
1. Statistical Baseline Modelling: The system establishes expected statistical norms or predictive models for tax returns within a given profession, income bracket, or historical profile.
2. Outlier Identification: Deviation detection algorithms scan new declarations to pinpoint anomalies (outliers) that diverge significantly from the expected model.
3. Concrete Example: If a retail business suddenly claims expense deductions three standard deviations above its 5-year historical mean or sector average, the record is flagged as a suspicious deviation.

Mentor tip

Define deviation detection as statistical outlier detection against an expected model, and give a specific tax-evasion example.

(d)

3 marks

Determine whether a network-model DBMS can replicate data warehouse functionality:
1. Verdict: No, a network-model DBMS is not equally suited for data warehousing analytics.
2. Transactional vs Analytical Orientation: Operational DBMS models (including network databases) are optimized for high-speed current CRUD transactions (OLTP), whereas data warehouses are subject-oriented and optimized for complex multi-dimensional analytical queries (OLAP).
3. Temporal Dimension: Operational DBMSs typically overwrite records with current states and lack the built-in multi-year historical time dimension essential to a data warehouse.

Mentor tip

Contrast OLTP (current operational transactions in a network DBMS) with OLAP (historical, time-variant, subject-oriented analysis in a data warehouse).

(e)(i)

2 marks

Explain why the Transformation (T) stage is necessary in ETL:
1. Data extracted from disparate external systems (banks, regional tax offices, employers) arrives in incompatible schemas, encodings, and naming conventions.
2. Transformation cleanses, validates, deduplicates, and standardizes these heterogeneous records into a single uniform schema before loading (e.g., converting different date formats or standardizing gender/status codes).

Mentor tip

Highlight that heterogeneous source databases use different formats/encodings that must be standardized into a uniform warehouse schema.

(e)(ii)

6 marks

Outline Application 1 — Consolidating Historical Data / Archiving (3 marks):
1. ETL extracts aged records (such as completed tax returns or dispatched orders older than a set retention period) from multiple live operational databases.
2. The transformation step filters out unnecessary operational fields, cleanses corrupt or duplicate entries, and formats the data for long-term analytical queries.
3. Loading the cleaned records into a separate historical archive offloads the transactional system and simplifies maintenance.
Outline Application 2 — System Data Migration or Data Mart Creation (3 marks):
4. When migrating from a legacy database to a modern platform (or building a departmental Data Mart slice, such as for a marketing or audit team), ETL extracts data from the source tables.
5. Transformation maps legacy field types, resolves structural incompatibilities, and establishes new relational links.
6. The converted data is loaded into the target database or specialized Data Mart so department users can run targeted business intelligence queries efficiently.

Mentor tip

Option B: Modelling and simulation · Option

Questions 5–8 · 65 marks

Question 5

15 marks

Lake fish population sampling, spreadsheet modelling, and model validation

Modelling and simulation

(a)

4 marks

Outline the steps to aggregate and record the five annual lake samples:
1. Extract the fish count and individual fish length measurements from each of the five random samples.
2. Sum the fish counts across all five samples and divide by 5 (or scale by lake area ratio) to compute the representative annual population estimate.
3. Sum the lengths of all captured fish across the five samples and divide by the total number of fish measured to calculate the estimated mean length for the year.
4. Append the year's identifier, estimated population total, and estimated mean length as a new annual summary record in the historical dataset.

Mentor tip

Show clearly how both metrics (representative population count and mean fish length) are computed from the 5 samples and stored by year.

(b)

2 marks

Identify the core variables required for the licensing decision model:
1. `Year` (or `Date`), to track the 5-year rolling window.
2. `EstimatedFishCount` (annual fish population estimate) and `AverageFishLength` (estimated mean fish length in cm).

Mentor tip

Include the temporal variable (Year) alongside both measured metrics (Population Count and Average Length).

(c)

3 marks

Explain how to construct and evaluate the model in spreadsheet software:
1. Data Layout: In a spreadsheet, organize `Year` in Column A, `EstimatedFishCount` in Column B, and `AverageFishLength` in Column C, with one row per year (e.g., rows 1 to 6 covering a 5-year span).
2. Percentage Change Formula: In Column D (e.g., cell `D6`), compute the 5-year percentage change in fish population using `= 100 * (B6 - B1) / B1`.
3. Decision Rule: In Column E, use a conditional formula such as `=IF(AND(D6 >= -5, C6 > 20), "Issue Licences", "Deny Licences")` to automatically determine whether both criteria are satisfied.

Percentage Change (D6) = \frac{B _{6} - B _{1}}{B _{1}} \times 100

Mentor tip

Reference specific spreadsheet columns, cell formulas, and the exact threshold conditions (`>= -5%` change and `> 20 cm` average length).

(d)

6 marks

Discuss how to test and refine the data collection and model rules:
1. Testing Data Collection Accuracy: Increase the sampling frequency (more than 5 times a year) and sample across a wider variety of lake zones/depths to verify whether five random samples accurately reflect the true fish population.
2. Refining Sampling Methodology: If parallel samples show high variance (e.g., due to seasonal breeding or fish congregating in specific temperatures), standardize the sampling locations and times of year.
3. Evaluating Model Thresholds: Compare historical licensing decisions against subsequent multi-year population trends to check if the 5% drop tolerance or 20 cm length threshold is overly strict (supporting anglers' complaints) or too lenient (supporting conservationists' complaints).
4. Adjusting Model Parameters: Modify the percentage tolerance or minimum length threshold incrementally and incorporate additional variables (such as catch limits per licence or water quality) to achieve stable fish populations without unnecessary bans.

Mentor tip

Address both potential sources of error: flawed input data (sampling methodology) and flawed decision rules (the 5% and 20 cm thresholds).

Question 6

15 marks

Weather pattern representation, climate displacement detection, and long-term simulation

Modelling and simulation

(a)

2 marks

Identify four measurable daily weather factors:
1. Temperature (minimum, maximum, or mean daily temperature) and Precipitation (daily rainfall/snowfall in mm).
2. Wind speed/direction, relative humidity, atmospheric pressure, or hours of sunshine.

Mentor tip

List any four standard meteorological variables (two per mark).

(b)

3 marks

Suggest how to visualize annual weather patterns:
1. Plot each meteorological factor on its own chart (or normalized multi-axis chart) across a 12-month horizontal time axis.
2. Use line graphs for continuous variables (such as daily or monthly average temperature) and bar charts for cumulative totals (such as monthly precipitation).
3. Group or smooth the daily values into weekly or monthly averages so seasonal cycles and trends over the year are immediately visible.

Mentor tip

Specify the chart type (line graph or bar chart), the axes (time over one year vs weather metric), and separate plotting for distinct units.

(c)

4 marks

Outline how a computer system detects seasonal displacement trends:
1. Long-term Archiving: Store daily or monthly weather records continuously across many decades.
2. Curve Fitting: Fit a mathematical curve or seasonal baseline profile to each year's annual weather cycle.
3. Time-Shift Comparison: Compare the curves across successive years by sliding (translating) the seasonal profile along the time axis or computing month-by-month anomalies.
4. Trend Identification: Measure whether peak summer temperatures or onset dates consistently shift later in the calendar year over time.

Mentor tip

Explain how multi-year historical data is mathematically compared (curve fitting and phase-shift/translation analysis) to spot delayed seasons.

(d)

6 marks

Discuss advantages of computer climate simulations:
1. Policy and Agricultural Planning: Advance warning of long-term warming or seasonal shifts allows governments and farmers to plan infrastructure, water management, and crop changes.
2. What-If Experimentation: Scientists can repeatedly run simulations with altered parameters (e.g., varying greenhouse gas emission rates) and validate models against historical climate records.
Discuss disadvantages and limitations:
3. Incomplete and Chaotic Variables: Climate systems involve complex, non-linear feedback loops and unknown external factors that cannot be perfectly modelled, leading to uncertainty.
4. Error Compounding: Small inaccuracies in initial conditions or cyclical patterns can amplify over long time horizons, producing unreliable predictions.
Formulate a balanced conclusion:
5. Conclusion: Provided models are continually calibrated against real-world observations, the ability to anticipate and mitigate severe environmental disasters outweighs the inherent uncertainties of long-range forecasting.

Mentor tip

Include concrete advantages, model limitations, and a reasoned conclusion weighing predictive value against uncertainty.

Question 7

15 marks

3D visualization, lighting and rendering, hardware/software requirements, and online gaming impacts

Modelling and simulation

(a)

2 marks

Define computer visualization:
1. Visualization is the computational process of converting abstract or binary numerical data stored in memory
2. into a graphical, pixel-based image displayed on a screen so humans can visually interpret it.

Mentor tip

State clearly that binary/numerical data in memory is transformed into a visual pixel representation.

(b)

5 marks

Outline how memory data is rendered into a realistic 3D screen image:
1. Geometric & Colour Representation: Objects are stored in memory as 3D coordinates (wireframe polygons/vectors) or voxel/pixel maps with base surface textures and colours.
2. Lighting & Viewpoint Calculation: The system identifies the position and intensity of virtual light sources relative to the object's surfaces and the camera's viewing angle.
3. Ray Tracing / Shading Equations: Mathematical equations calculate surface reflection, absorption, shadows, and occlusion for each polygon.
4. Pixel Shading: Each 2D screen pixel's colour and brightness value is transformed into the appropriate gradient or shade.
5. Final Rendering: The shaded pixels are written to the frame buffer, producing depth cues and realistic 3D perspective on the 2D display.

Mentor tip

Trace the pipeline from memory representation -> light source & camera angle math -> shading/reflection -> 2D pixel colour mapping (rendering).

(c)

3 marks

Outline essential hardware and software specifications for online 3D gaming:
1. Hardware (Graphics & Memory): A dedicated Graphics Processing Unit (GPU / fast video card) with high VRAM, supported by ample system RAM and a fast multi-core CPU, to render complex 3D frames in real time.
2. Hardware (Audio & Network): A high-speed broadband network interface (for low-latency server communication) and a capable sound card.
3. Software: An operating system capable of rapid real-time I/O scheduling, paired with graphics/multimedia APIs (such as DirectX or OpenGL) and up-to-date GPU drivers.

Mentor tip

Cover both hardware (GPU, RAM, network/sound) and software (OS with fast I/O, graphics APIs like DirectX/OpenGL).

(d)

5 marks

Discuss benefits of participating in online simulated gaming worlds:
1. Social & Cognitive Benefits: Players can relax, collaborate with international peers, build online communities, and develop strategic problem-solving and spatial skills.
2. Safe Exploration: Virtual environments allow users to experience scenarios and team roles that are impossible or dangerous in real life.
Discuss drawbacks and risks:
3. Addiction & Neglect: Excessive immersion can lead to gaming addiction, causing players to neglect academic, professional, or physical health responsibilities.
4. Social Isolation: Over-reliance on virtual interactions can isolate individuals from face-to-face relationships and blur the boundary between virtual and real-world norms.
Provide a balanced conclusion:
5. Conclusion: When gaming time is moderated and balanced with offline responsibilities, online virtual worlds offer an engaging, beneficial form of modern entertainment and social connection.

Mentor tip

Include two distinct advantages, two distinct disadvantages, and a concluding judgement.

Question 8

20 marks

Neural network hidden units and weights, OCR vs speech recognition, NLP approaches, and genetic algorithms for timetabling

Modelling and simulation

(a)

3 marks

Outline the function of hidden units and weights in an artificial neural network:
1. Hidden units (internal layer nodes between the input and output layers) receive numerical activations from the input nodes.
2. Each incoming connection value is multiplied by an adjustable numerical weight that represents the strength/importance of that feature.
3. Each hidden unit sums its weighted inputs, applies an activation function, and forwards the signal to the output layer (having more hidden nodes enables the network to model more complex non-linear patterns).

Mentor tip

Trace the signal flow: input nodes -> multiplied by connection weights -> summed/processed at hidden layer nodes -> sent to output layer.

(b)

4 marks

Compare similarities between neural networks in speech recognition and OCR (2 marks):
1. Both systems are trained using labelled training datasets of sample patterns.
2. Both iteratively adjust internal connection weights during training to minimize error against a target fitness/loss function.
Contrast differences between speech recognition and OCR (2 marks):
3. In Optical Character Recognition (OCR), printed characters conform closely to fixed geometric shapes and font templates (with variance mainly limited to handwriting).
4. In speech recognition, acoustic inputs exhibit far greater continuous variability across speakers due to pitch, speed, accent, intonation, and background noise, making pattern matching more sensitive to deviations from the training set.

Mentor tip

(c)

2 marks

Distinguish between speech synthesis and natural language generation (NLG):
1. Speech synthesis (text-to-speech) is the mechanical/phonetic conversion of existing written text into audible spoken sound waves.
2. Natural language generation (NLG) is the semantic and syntactic process of constructing meaningful human-language sentences from structured underlying data (such as composing a spoken answer from database query results).

Mentor tip

Speech synthesis = converting pre-existing text to audio (text-to-speech); NLG = deciding what words/sentences to compose from raw data/meaning.

(d)

4 marks

Discuss the heuristic vs cognitive approaches to automated natural language learning:
1. Heuristic Approach: Relies on statistical probabilities, rule-of-thumb shortcuts, and supervised learning corpora to map inputs to likely outputs.
2. Heuristic Limitation: Its accuracy is bounded by the breadth and quality of its pre-labelled training sets.
3. Cognitive Approach: Uses unsupervised learning and semantic modelling to infer underlying grammar rules, context, and conceptual meaning.
4. Cognitive Characteristic: Aims to acquire syntax and semantics organically in a manner analogous to how a human child learns language.

Mentor tip

Contrast heuristic learning (statistical probabilities + supervised training sets) with cognitive learning (unsupervised acquisition of syntax and semantics like a child).

(e)

7 marks

Outline step-by-step how a genetic algorithm generates a valid school timetable:
1. Define Inputs & Constraints: Encode all scheduling rules and resources (teacher maximum hours, student subject choices, room capacities, and no double-booked rooms).
2. Establish Fitness Function & Priorities: Assign weights/priorities to hard constraints (must be met, e.g., no room clashes) and soft constraints to calculate a percentage fitness score for any generated timetable.
3. Initialize Population: Randomly generate an initial population of trial school timetables.
4. Evaluate Fitness: Score each timetable in the current generation using the fitness function.
5. Selection: Select the highest-scoring (fittest) timetables to act as parents for the next generation.
6. Crossover & Mutation: Combine schedules from parent timetables (crossover) and introduce small random slot swaps (mutation) to produce a new generation of offspring timetables.
7. Termination Condition: Repeat the evaluation, selection, and breeding cycle iteratively until a timetable achieves an tolerable fitness threshold (all mandatory constraints satisfied).

Mentor tip

Option C: Web science · Option

Questions 9–14 · 65 marks

Question 9

17 marks

HTML vs HTTP, XML data structures, DNS resolution, and web protocols vs standards

Web science

(a)

2 marks

Distinguish between HTML and HTTP:
1. HTML (HyperText Markup Language) is a markup language used to structure and format content for display in a web browser.
2. HTTP (HyperText Transfer Protocol) is a communications protocol that defines the rules for transmitting requests and web files between a client and a server.

Mentor tip

Contrast a markup language for page structure (HTML) with a network communication protocol for data transfer (HTTP).

(b)

1 mark

Identify one structural similarity between HTML and XML:
Both are markup languages that use angled-bracket tags (`<tag>...</tag>`) to enclose and structure elements.

Mentor tip

Mention the shared use of markup tags.

(c)

2 marks

Outline how the given XML snippet demonstrates self-describing data:
1. XML uses descriptive, human-readable custom tag names that explicitly label the meaning of the enclosed data.
2. For example, `<species>Bubo bubo</species>` and `<habitat>Forests</habitat>` directly identify that 'Bubo bubo' is a biological species and 'Forests' is its habitat within a `<bird>` record.

Mentor tip

Quote specific tags from the snippet (such as `<bird>`, `<name>`, or `<species>`) to satisfy 'direct reference to the XML code above'.

(d)

2 marks

Describe how XML's hierarchical structure facilitates database updates:
1. The root/parent element (`<bird>`) represents a database record/entity, while the nested child tags (`<name>`, `<species>`, `<description>`, `<habitat>`) correspond to specific field/column names.
2. The server can parse the XML tree and map each tag's text value directly into the matching column of the target database table.

Mentor tip

Connect XML parent elements to database records and child tag names to database table columns/fields.

(e)

2 marks

Identify the two actions a Domain Name Server (DNS) can take upon receiving a URL:
1. If the domain name is in its local cache/database, it resolves the domain name and returns the corresponding IP address to the client browser.
2. If it does not hold the mapping, it forwards the lookup request to another more authoritative DNS server (or returns an `NXDOMAIN` error if the domain does not exist).

Mentor tip

Cover both the cache hit case (return matching IP address) and the cache miss case (query higher-level/authoritative DNS server).

(f)(i)

2 marks

Distinguish between a protocol and a standard:
1. A protocol is a formal set of procedural rules governing how data is transmitted and how devices communicate across a network.
2. A standard is an agreed-upon technical specification or format (maintained by bodies like W3C or IEEE) that ensures consistency and compatibility across different hardware and software implementations.

Mentor tip

Protocols govern communication procedures/rules in transit; standards define agreed technical specifications/formats.

(f)(ii)

6 marks

Discuss the consequences of lacking a specific web protocol (3 marks):
1. Named Protocol: TCP/IP or HTTP.
2. Technical Problem if Absent: Without TCP/IP, routers and end-hosts would have no agreed rules for packetizing, addressing, routing, or reassembling data across heterogeneous networks.
3. Concrete Consequence: Packets sent from a Linux server to a Windows client would be dropped or misinterpreted, making reliable global data transfer impossible.
Discuss the consequences of lacking a specific web standard (3 marks):
4. Named Standard: HTML, CSS, Unicode (UTF-8), or JPEG.
5. Technical Problem if Absent: Without a W3C HTML standard, every browser vendor and web developer would invent proprietary markup syntax.
6. Concrete Consequence: A webpage written for one browser (e.g., Chrome) would fail to render or display garbled code in another browser (e.g., Safari or Firefox), fragmenting the World Wide Web.

Mentor tip

Explicitly name one protocol (e.g., TCP or HTTP) and one standard (e.g., HTML or Unicode), and explain the specific technical breakdown that would occur without each.

Question 10

12 marks

PageRank algorithm, link weighting factors, meta-tag limitations, and black hat SEO

Web science

(a)

2 marks

Identify the specific purpose of Google's PageRank algorithm:
1. To calculate a numerical authority/importance score for each indexed web page
2. so that search results for a specific query can be ranked in order of relevance and importance.

Mentor tip

Mention both assigning a numerical importance rating to a page and ordering search results for a user's query.

(b)

2 marks

Identify two factors determining how much Page A's link boosts Page B's PageRank:
1. The existing PageRank (authority score) of Page A itself.
2. The total number of outbound links (`outlinks`) on Page A, since Page A's authority is divided among all pages it links to.

Mentor tip

Looking at the diagram, notice Page A has both incoming links (affecting Page A's own rank) and multiple outgoing links (diluting the share passed to Page B).

(c)

2 marks

Explain why meta-tags may return unexpected search results:
1. Polysemy / Ambiguity: Words placed in meta-tags often have multiple unrelated meanings (homonyms) without semantic context.
2. Example: A page tagged with `meta="star"` about astrophysics may be returned when a user searches for a celebrity movie 'star' (or webmasters may insert misleading/inaccurate keywords in meta-tags).

Mentor tip

Use a concrete example of an ambiguous keyword (like 'star', 'apple', or 'jaguar') to show why keyword matching alone can fail.

(d)

6 marks

Explain three black hat Search Engine Optimization (SEO) techniques:
1. Keyword Stuffing: Repetitively packing popular search terms into `<meta>` tags or page text so the indexing algorithm artificially scores the page as highly relevant to those queries.
2. Invisible / Hidden Text: Placing extra keywords on the page in the same colour as the background, in tiny font sizes, off-screen via CSS, or inside HTML comments so web crawlers index them while human visitors cannot see them.
3. Link Farms / Spamdexing: Creating networks of dummy websites or automated blog comments that all link back to the target page to artificially inflate its inbound link count and manipulate PageRank.

Mentor tip

Name and explain three distinct black-hat practices: keyword stuffing, hidden text/cloaking, and link farms.

Question 11

16 marks

Lossy compression file types, open standards vs interoperability, Web 2.0 social networking, and cloud security

Web science

(a)

2 marks

Identify two file types suitable for lossy compression:
1. Digital audio files (e.g., MP3, AAC).
2. Digital image/photograph files (e.g., JPEG) or video files (e.g., MP4).

Mentor tip

Choose multimedia formats (images, audio, video) where minor perceptual data loss is tolerable, never text or executable files.

(b)

4 marks

Outline the difference between open standards and interoperability with examples:
1. Open Standards: Publicly available, royalty-free technical specifications that anyone can inspect and implement without restriction.
2. Example of Open Standard: HTML, CSS, or HTTP published by the W3C/IETF.
3. Interoperability: The practical ability of diverse hardware devices, operating systems, and software applications from different vendors to exchange data and work together seamlessly.
4. Example of Interoperability: Because HTML is an open standard, a webpage hosted on a Linux Apache server can be rendered identically by Safari on iOS, Chrome on Android, and Edge on Windows.

Mentor tip

Define open standards as the publicly documented specifications, and interoperability as the resulting ability of different systems to work together.

(c)

4 marks

Describe two Web 2.0 technological features that enabled social networking:
1. User-Generated Content & Rich Interactivity (AJAX/DOM scripting): Dynamic client/server scripting allowed non-technical users to post status updates, comment on friends' walls, and personalize profiles in real time without writing HTML.
2. Seamless Multimedia Integration & Instant Messaging: High-speed browser media players, file-upload APIs, and asynchronous chat allowed users to share photos/videos and communicate instantaneously within the browser.

Mentor tip

State each Web 2.0 feature (e.g., dynamic user-generated content, multimedia sharing, real-time messaging) and explain how it fostered social interaction.

(d)

6 marks

Discuss security considerations of a third-party cloud service vs an in-house server:
1. Third-Party Cloud Advantages: Professional cloud providers employ dedicated security specialists, enterprise-grade firewalls, automated encryption, and redundant off-site backups that a small company could not afford on its own.
2. Third-Party Cloud Risks: The small company relinquishes direct physical control over customer data; internal cloud operations are opaque, and transmitting data over the internet introduces interception risks.
3. In-House Server Advantages: All customer records remain physically on-premises under the direct oversight of trusted company staff, avoiding third-party multi-tenancy exposure.
4. In-House Server Risks: A small business typically lacks a specialized 24/7 cybersecurity team, leaving local servers vulnerable to unpatched exploits, physical theft, or local disasters.
5. Balanced Conclusion: Unless regulatory rules mandate on-premises storage, a reputable private cloud provider generally delivers stronger, more up-to-date security than a small company's limited internal IT staff can maintain.

Mentor tip

Compare both options (Cloud storage vs Company in-house storage) in detail and provide a reasoned conclusion to earn full marks.

Question 12

12 marks

Directed web graphs, strongly connected core (SCC), bowtie IN/OUT sections, and recursive web crawling

Web science

(a)(i)

2 marks

Identify the web graph components in the diagram:
1. The nodes (vertices `A` through `F`) represent individual web pages.
2. The directed edges (arrows connecting the nodes) represent hyperlinks pointing from a source page to a target page.

Mentor tip

Explicitly state what both the nodes (web pages) and the directed arrows (hyperlinks and their direction) represent.

(a)(ii)

2 marks

Explain why these nodes cannot form a Strongly Connected Core (SCC):
1. In an SCC, every node must be reachable from every other node by following directed links.
2. In the given graph, Node `A` and Node `F` have no incoming links (in-degree = 0), so it is impossible to navigate back to Page `A` or Page `F` from any other page.

Mentor tip

Point out specific unreachable nodes (`A` or `F` have only outgoing arrows and zero incoming arrows) to prove mutual reachability fails.

(b)

4 marks

Distinguish between the IN and OUT sections of the bowtie web graph with examples:
1. IN Section Features: Contains web pages that have outbound links leading into the Strongly Connected Core (SCC), but cannot be reached from the SCC (only outlinks towards the core, no inlinks from the core).
2. IN Section Example: Newly published web pages or personal blogs that link to major websites before any external sites have linked back to them.
3. OUT Section Features: Contains web pages that can be reached via links from the SCC, but do not contain links leading back into the SCC (effectively dead-ends / inlinks only with respect to the core).
4. OUT Section Example: Static corporate PDF documents, reference archives, or government policy pages that receive links from the web but do not link back out.

Mentor tip

Contrast the link direction relative to the SCC for both IN and OUT, and give a concrete webpage example for each.

(c)

4 marks

Explain the recursive traversal algorithm used by web crawlers:
1. Fetch & Store: The crawler visits the current URL, downloads a copy of the web page for indexing, and marks the URL as visited.
2. Parse for Outlinks: The algorithm scans through the page's HTML source code sequentially looking for unvisited hyperlink URLs (`<a href="...">`).
3. Recursive Call (Depth-First Jump): When it encounters an unvisited URL, it invokes the crawl procedure recursively on that new target URL.
4. Base Case & Backtracking: When a page contains no further unvisited links (a dead end), the current call terminates and returns control to the previous caller URL to resume scanning remaining links until all reachable pages are visited.

Mentor tip

Clearly describe all four aspects: fetching the current page, scanning for links, jumping recursively to a found link, and backtracking to the parent page upon hitting a dead end.

Question 13

2 marks

Ontological systems for integrating heterogeneous publisher book data

Web science

2 marks

Describe how an ontology resolves heterogeneous field labels across book publishers:
1. An ontology defines a formal, shared vocabulary and semantic mapping of concepts and relationships within a domain.
2. When different publishers use conflicting tag names for the same attribute (e.g., `<Publisher>`, `<Publishing_House>`, or `<Imprint>`), the ontology maps all synonymous tags onto a single standardized schema concept so the bookshop's database can automatically merge records without manual reformatting.

Mentor tip

Define an ontology as a standardized semantic vocabulary/mapping and give a concrete example of resolving two different publisher tag names into one unified field.

Question 14

6 marks

Ambient intelligence technologies and workplace surveillance ethics

Web science

6 marks

Describe the ambient intelligence hardware and network infrastructure (3 marks):
1. Wearable/Embedded Identifiers: Employees carry ID badges or devices embedded with active RFID tags, Bluetooth Low Energy (BLE) beacons, or biometric sensors.
2. Distributed Sensor Network: Unobtrusive wireless readers/receivers installed throughout doorways, corridors, and offices detect the badges automatically without manual check-ins.
3. Central Real-Time Processing: The sensors transmit location coordinates wirelessly to a central server that logs and models employee positions in real time.
Discuss the ethical implications for employers and employees (3 marks):
4. Employer Perspective (Productivity & Safety): Management may argue that since staff are compensated during working hours, tracking location ensures accountability, optimizes building security, and locates personnel rapidly during emergency evacuations.
5. Employee Perspective (Privacy & Autonomy): Continuous spatial tracking infringes on personal privacy, creates a high-stress culture of micromanagement and distrust (e.g., monitoring washroom breaks), and exposes private workplace social interactions or union associations.

Mentor tip

Option D: Object-oriented programming · Option

Questions 15–18 · 65 marks

Question 15

19 marks

College library OOP advantages, developer ethics, Student and Loan encapsulation, and static variables

Object-oriented programming

(a)

4 marks

Outline two advantages of using an OOP approach for the development team:
1. Parallel Modular Development: Because the team is divided into small sub-groups, each sub-group can independently develop and test distinct classes (such as `Student`, `Loan`, `shortTerm`) simultaneously, speeding up delivery.
2. Code Reusability & Easier Maintenance: Objects and methods developed by one sub-group can be reused across the project without rewriting code, and isolated class modules are faster to debug and unit-test.

Mentor tip

Link OOP features (modularity, reuse, independent testing) directly to the scenario of a team split into small sub-groups.

(b)

4 marks

Explain two ethical responsibilities of the software development team:
1. Rigorous Testing for Fairness and Accuracy: Developers have an ethical duty to thoroughly test fine calculations and suspension rules so students are not unfairly overcharged or wrongly barred from borrowing academic resources.
2. Data Privacy & Intellectual Property: Developers must protect confidential student borrowing records from unauthorized disclosure and properly acknowledge/license any external code libraries used to avoid plagiarism or copyright infringement.

Mentor tip

State each ethical issue (e.g., thorough testing of fine/suspension logic, protecting student privacy, or acknowledging code sources) and explain why it matters.

(c)

3 marks

Outline how encapsulation enhances security in the `Student` class:
1. In `Student`, attributes such as `private int studentID` and `private String studentName` are declared with the `private` access modifier.
2. This hides the internal state so external classes cannot directly read or accidentally corrupt student data.
3. External code can only interact with these fields through controlled `public` accessor/mutator methods (`getStudentID()`, `addLoan()`).

Mentor tip

Explicitly reference the `private` keyword on attributes like `studentID` or `booksBorrowed` in the `Student` class.

(d)

2 marks

Construct the Java statement to instantiate the `Student` object:
1. Declare a `Student` reference variable and invoke the `new` operator.
2. Pass the `int` ID `93003` and `String` name `"Smith"` to the `Student` constructor:

Student newStudent = new Student(93003, "Smith");

Mentor tip

Match the exact parameter order of the constructor `public Student(int studentID, String studentName)`.

(e)

3 marks

Explain the purpose of `static` in `static int numBooksLoaned = 0;`:
1. A `static` variable belongs to the `Loan` class itself rather than to any individual `Loan` instance, so only a single shared copy exists in memory.
2. It is initialized once when the class loads, not reset to zero each time a new `Loan` object is instantiated.
3. Because the constructor increments `numBooksLoaned` every time a `Loan` is created, it tracks the cumulative total of all books borrowed across the entire library.

Mentor tip

Contrast a class-level variable (one shared copy across all instances) with an instance variable, and explain its role as a global loan counter.

(f)

3 marks

Construct the Java code to add the specified `Loan` to `Student` object `ST`:
1. Instantiate a new `Loan` object with `new Loan(212000, "The Stars")`.
2. Pass that newly created `Loan` instance into `ST.addLoan(...)`:

ST.addLoan(new Loan(212000, "The Stars"));

Mentor tip

You can either instantiate the `Loan` inline inside `ST.addLoan(...)` or assign it to a temporary `Loan` variable first and pass that variable to `ST.addLoan(...)`.

Question 16

15 marks

Tracing Student/Loan arrays, constructing showDetails(), and addressing missing requirements

Object-oriented programming

(a)(i)

1 mark

Trace `borrowers[93001].getStudentName()`:
`borrowers[93001]` references the `Student` initialized with ID `93001` and name `"Jones"`. Output: `Jones`.

Mentor tip

Look up index 93001 in the `borrowers` array.

(a)(ii)

1 mark

Trace `borrowers[93001].getLoan(1).getBookTitle()`:
Student `93001` had `"The Sky"` added at loan index `0`, and `"The Spooks"` added second at loan index `1`. Output: `The Spooks`.

Mentor tip

Remember Java arrays are 0-indexed: `getLoan(0)` is the first loan (`The Sky`), so `getLoan(1)` is the second loan (`The Spooks`).

(a)(iii)

1 mark

Trace `borrowers[3012].getLoan(0).getBookTitle()`:
Student `3012` (`Zang`) had `"The Animals"` added as their first loan (index `0`). Output: `The Animals`.

Mentor tip

Index `0` retrieves the first `Loan` added to `borrowers[3012]`.

(b)

7 marks

Construct the `showDetails(int studentID)` method:
1. Define a `void` method taking `int studentID` as a parameter.
2. Retrieve the `Student` object from `borrowers[studentID]` and output `getStudentName()`.
3. Loop through the `booksBorrowed` array from index `0` to `9` (`i < 10`).
4. Retrieve each `Loan` via `s.getLoan(i)`, check that it is not `null`, and print `getBookTitle()` and `getDate()`.
```java public void showDetails(int studentID) { Student s = borrowers[studentID]; System.out.println(s.getStudentName()); for (int i = 0; i < 10; i++) { Loan currentLoan = s.getLoan(i); if (currentLoan != null) { System.out.println(currentLoan.getBookTitle()); System.out.println(currentLoan.getDate()); } } } ```

Mentor tip

Check `if (currentLoan != null)` before calling `.getBookTitle()` and `.getDate()` because a student may have borrowed fewer than 10 books.

(c)

3 marks

State three user requirements missing from the current class design:
1. No distinction between short-term (5-day) and long-term (30-day) book loans.
2. No attributes or methods to calculate overdue fines ( $1.50/ d a y f or s h or t - t er m v s$ 0.10/day for long-term).
3. No flag or validation logic to suspend a student's borrowing rights when a short-term loan is overdue by more than 4 days (nor a boundary check preventing `addLoan` from exceeding 10 books).

Mentor tip

Compare the bulleted user requirements at the start of Option D against the attributes/methods actually present in `Student` and `Loan`.

(d)

2 marks

Describe a modification to the `Student` class to address one missing requirement:
1. Add a `private boolean canBorrow = true;` (or `borrowingSuspended`) attribute to `Student` along with getter/setter methods.
2. Alternatively, modify `addLoan(Loan book)` to check `if (numBooks < 10 && canBorrow)` before inserting a new loan into `booksBorrowed`.

Mentor tip

Specify both the new attribute/check added to `Student` and how its value changes or governs borrowing.

Question 17

11 marks

Inheritance hierarchy, shortTerm/longTerm subclasses, UML class diagram, and static counters

Object-oriented programming

(a)

2 marks

Describe one benefit of inheritance in OOP:
1. Inheritance promotes code reuse and eliminates duplication
2. by allowing subclasses (`shortTerm` and `longTerm`) to automatically inherit common attributes (`bookID`, `bookTitle`, `d`) and methods from the `Loan` superclass.

Mentor tip

Connect reduced code duplication to subclasses inheriting shared fields and methods from their superclass.

(b)

2 marks

State two additional attributes required in the `shortTerm` subclass:
1. `private double dailyFineRate;` (or `fine`, e.g., `$1.50`).
2. `private int maxLoanDays;` (borrowing limit of 5 days) or `private boolean causesSuspension;` / `maxOverdueDays` (4 days).

Mentor tip

Pick any two attributes specific to short-term loans: daily fine rate, loan duration limit, or overdue suspension threshold.

(c)

4 marks

Construct the UML class diagram showing relationships among `Student`, `Loan`, `shortTerm`, and `longTerm`:
1. Draw four class boxes labelled `Student`, `Loan`, `shortTerm`, and `longTerm`.
2. Draw a `has-a` association/aggregation line between `Student` and `Loan` (a `Student` holds an array of `Loan` objects).
3. Draw `is-a` inheritance arrows (hollow triangle arrowheads pointing to the superclass) from both `shortTerm` and `longTerm` up to `Loan`.
``` [Student] <>---- (has-a) ----> [Loan] ^ | (is-a / inherits) +---------+---------+ | | [shortTerm] [longTerm] ```

Mentor tip

Clearly distinguish the 'has-a' relationship (`Student` to `Loan`) from the 'is-a' inheritance relationships (`shortTerm` and `longTerm` pointing to `Loan`).

(d)

3 marks

Outline how to display the total number of books on loan without traversing `borrowers`:
1. Declare a `static` counter variable (or use `Loan.numBooksLoaned`, or `static` counters in `shortTerm` and `longTerm`) that increments whenever a loan is instantiated and decrements when a book is returned.
2. Provide a `public static` accessor method in `Loan` (or sum the subclass static counters).
3. Call this `static` method directly on the class to output the total count in $O (1)$ time without looping through `borrowers`.

Mentor tip

Use `static` class variables and a `static` method so the count is maintained at the class level rather than by summing array elements.

Question 18

20 marks

Searching borrowed books, binary search trees for fast lookup, and replacing arrays with linked lists

Object-oriented programming

(a)

4 marks

Explain why locating who borrowed a specific book is slow in the current design:
1. Encapsulated Unindexed Arrays: Borrowed `Loan` objects are scattered across individual `booksBorrowed` arrays inside separate `Student` objects within the `borrowers` array.
2. Sparse Array Traversal: The `borrowers` array has 100,000 slots (most of which are `null`), and each student's `booksBorrowed` array has 10 slots (many of which are also `null`).
3. Exhaustive Linear Search: Because there is no index keyed by `bookID`, the system must perform a nested linear search across every non-null `Student` and every non-null `Loan` until a matching `bookID` is found.
4. High Time Complexity: This exhaustive scan results in poor $O (N \times M)$ performance.

Mentor tip

Explain that loans are distributed inside individual `Student` objects with no `bookID` index, forcing a slow linear search across sparse arrays.

(b)(i)

2 marks

Describe a suitable course data structure for fast book lookup:
1. A Binary Search Tree (BST) of nodes representing active book loans,
2. ordered/indexed by the unique `bookID` key, where each node stores both the `bookID` and the borrowing student's `studentID`.

Mentor tip

Name a Binary Search Tree (BST) and specify that `bookID` is used as the ordering key.

(b)(ii)

4 marks

Outline the steps to populate and search the Binary Search Tree:
1. Insertion on Loan: Whenever a book is borrowed, insert a new node containing `(bookID, studentID)` into the BST ordered by `bookID`.
2. Binary Search Traversal: To find who borrowed a target `bookID`, start at the root node and compare the target `bookID` with the current node's `bookID`, branching left if smaller or right if larger ( $O (lo g n)$ average comparisons).
3. Retrieve Student ID: Once the matching `bookID` node is reached, extract the stored `studentID`.
4. Direct Array Access: Use `borrowers[studentID]` to immediately access the corresponding `Student` object and output their details.

Mentor tip

Cover both how nodes `(bookID, studentID)` are inserted when loaned and how the tree is searched by `bookID` to retrieve `studentID`.

(c)

4 marks

Describe one advantage of replacing the `borrowers` array with a linked list (2 marks):
1. Dynamic Memory Efficiency: A linked list allocates memory nodes dynamically only for currently enrolled students.
2. Eliminates Wasted Space: Avoids reserving a fixed 100,000-element `borrowers` array where the vast majority of indices remain `null`.
Describe one disadvantage of using a linked list (2 marks):
3. Loss of $O (1)$ Direct Indexing: You can no longer access a student directly via `borrowers[studentID]`.
4. Slower Sequential Lookup: Finding a specific student requires traversing the linked list sequentially from the `head` pointer via a linear search ( $O (n)$ time).

Mentor tip

Give one clear advantage with elaboration (dynamic sizing avoids wasted memory on empty array slots) and one clear disadvantage with elaboration (sequential traversal replaces instant index lookup).

(d)

6 marks

Construct the `addToEnd()` algorithm to append a `shortTerm` loan node to the end of a linked list:
1. Instantiate or receive the `shortTerm` node to be inserted and set its `next` pointer to `null`.
2. Obtain the `head` reference of the linked list.
3. Handle the empty-list edge case: if `head == null`, set `head` to the new `shortTerm` node.
4. Otherwise, initialize a temporary traversal pointer `temp = head`.
5. Loop `while (temp.getNext() != null)` to advance `temp = temp.getNext()` until the last node is reached.
6. Link the last node's `next` pointer to the new `shortTerm` object (`temp.setNext(newBook)`).
```java public void addToEnd(shortTerm newBook) { newBook.setNext(null); if (head == null) { head = newBook; } else { Node temp = head; while (temp.getNext() != null) { temp = temp.getNext(); } temp.setNext(newBook); } } ```

Mentor tip

Our analysis

Where the marks are

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

Databases65 marks · 25%
Q1, Q2, Q3, Q4
Modelling and simulation65 marks · 25%
Q5, Q6, Q7, Q8
Web science65 marks · 25%
Q9, Q10, Q11, Q12, Q13, Q14
Object-oriented programming65 marks · 25%
Q15, Q16, Q17, Q18

Mentor insights

Traps to watch for

01
Databases: Shelving unit orders normalization from unnormalized form to 3NF · Q3
Check the three classic 1NF requirements: atomic values in every cell, no repeating groups/blank multi-row cells, and a defined primary key.
02
Modelling and simulation: Neural network hidden units and weights, OCR vs speech recognitio · Q8
Trace the signal flow: input nodes -> multiplied by connection weights -> summed/processed at hidden layer nodes -> sent to output layer.
03
Web science: HTML vs HTTP, XML data structures, DNS resolution, and web protoc · Q9
Contrast a markup language for page structure (HTML) with a network communication protocol for data transfer (HTTP).
04
Object-oriented programming: Searching borrowed books, binary search trees for fast lookup, an · Q18
Explain that loans are distributed inside individual `Student` objects with no `bookID` index, forcing a slow linear search across sparse arrays.

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