Loading…

Worked solutions

IB Computer Science HL · Paper 2

Computer Science, Higher Level

Session: May 2023
Marks: 260
Duration: 1 hour 20 min
Calculator: Not allowed
Questions: 16

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

Keep your copy of the paper open alongside. We don't reproduce the questions: each part below is labelled to match the May 2023 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

21 marks

Quick Rent-a-Car relational database keys, SQL queries, validation, views, DDL, and data consistency

Databases

(a)(i)

1 mark

State the relationship cardinality between CUSTOMER and RENTAL:
One-to-many (1:M) — one customer can enter into multiple rental agreements over time.

Mentor tip

Look at the crow's foot symbol on the RENTAL side of the ERD link.

(a)(ii)

1 mark

State the name of any one primary key from the schema:
CustID (in CUSTOMER), RentalNum (in RENTAL), or CarID (in CAR).

Mentor tip

Any one of the three underlined/key identifiers (CustID, RentalNum, or CarID) is sufficient.

(a)(iii)

1 mark

State the table that contains foreign keys:
RENTAL (which holds CustID referencing CUSTOMER and CarID referencing CAR).

Mentor tip

The junction/transaction table at the 'many' end of both 1:M relationships holds the foreign keys.

(b)

1 mark

Trace the SQL query against the CAR table (WHERE Man = "Lexus" AND Year = 2016):
Row DHE7623 matches both Man = "Lexus" and Year = 2016; its Make attribute is:
EF475

Mentor tip

Output only the projected Make value (EF475) rather than the entire record.

(c)

4 marks

Check 1 — Format / Input Mask Check:
1) Verifies that CarID strictly follows the pattern of 3 uppercase letters followed by 4 digits (e.g. LLL0000 such as WRE2345).
Check 2 — Length Check (or Presence/Uniqueness Check):
2) Verifies that the entered CarID string contains exactly 7 characters (not shorter or longer).

Mentor tip

Name two distinct validation checks (such as Format check LLL9999 and Length check = 7) and explain how each applies to CarID.

(d)

4 marks

Identify the query steps to retrieve the customer's surname:
1) Select the target attribute: SELECT SName
2) Specify both required tables: FROM CUSTOMER INNER JOIN RENTAL
3) Join the tables on their shared key: ON CUSTOMER.CustID = RENTAL.CustID
4) Filter by the rental dates: WHERE DateOut = '22/01/2020' AND DateRet = '26/01/2020';

Mentor tip

Ensure you include the projection (SName), both table names, the join condition on CustID, and both date conditions.

(e)

3 marks

Explain why queries are used to create database views:
1) Virtual Table Abstraction: A view encapsulates a complex multi-table join query into a reusable virtual table that staff can query simply without rewriting SQL.
2) Dynamic Updating: Because a view does not store separate physical data, it automatically reflects real-time changes made to the underlying CUSTOMER, RENTAL, and CAR tables.
3) Role-Based Security: Views can expose only specific non-sensitive columns or rows to front-desk employees while hiding restricted attributes in the base tables.

Mentor tip

Highlight that views act as virtual tables that simplify complex queries, stay dynamically updated, and restrict column/row visibility for security.

(f)

3 marks

Explain how a Data Definition Language (DDL) implements the data model:
1) DDL commands (such as CREATE TABLE) define the database schema by creating the CUSTOMER, RENTAL, and CAR tables along with their field names and data types.
2) DDL specifies integrity constraints, including PRIMARY KEY declarations (CustID, RentalNum, CarID) and FOREIGN KEY relationships.
3) DDL statements (such as ALTER TABLE or DROP TABLE) allow the DBA to modify or maintain the table structure over time.

Mentor tip

Connect DDL to defining table structures, field data types, and primary/foreign key constraints.

(g)

3 marks

Explain the importance of maintaining data consistency in the car rental database:
1) Referential and Transactional Accuracy: Consistency ensures that every record written to the database satisfies all defined constraints, cascades, and validation rules (e.g. a rental cannot reference a non-existent CarID or CustID).
2) Reliable Availability Tracking: Accurate, consistent state across tables prevents double-booking vehicles or charging customers for the wrong rental period.
3) Trustworthy Analytics and Compliance: Managers and external financial audits can rely on consistent data formats for reporting and decision-making.

Mentor tip

Link data consistency to enforcing integrity constraints, preventing booking conflicts, and ensuring reliable operational decisions.

Question 2

9 marks

Database recovery strategies, recovery methods, and restricting staff access to sensitive data

Databases

(a)

1 mark

Identify one factor a DBA considers when selecting a recovery strategy:
Recovery Time Objective (maximum tolerable downtime before service is restored), Recovery Point Objective (maximum tolerable data loss), financial budget for failover hardware, or legal/regulatory requirements.

Mentor tip

State one specific factor such as tolerable downtime (RTO), tolerable data loss (RPO), or implementation cost.

(b)

4 marks

Method 1 — Periodic Backup and Restore with Transaction Log Replay:
1) The DBA restores the most recent scheduled full or incremental backup image of the database, then replays committed transactions from the write-ahead transaction log (roll-forward) or undoes incomplete transactions (rollback) to reach a consistent point in time.
Method 2 — Database Mirroring / Instant Failover Recovery:
2) A synchronized standby replica server continuously mirrors live transactions; if the primary database fails, user traffic is immediately redirected to the mirror server with zero downtime while the primary server is repaired.

Mentor tip

Describe two distinct recovery techniques in detail, such as backup restoration with log replay and live failover mirroring.

(c)

4 marks

Method 1 — Role-Based Access Control (RBAC) and Granular Permissions:
1) Assign staff accounts to specific security roles (e.g. receptionist, manager, HR) with least-privilege permissions and restricted database views so employees can only query tables or columns required for their job duties.
Method 2 — Data Anonymization / Pseudonymization (Splitting PII):
2) Separate Personally Identifiable Information (such as names and phone numbers) into a restricted identity table and replace them in operational tables with anonymous surrogate identifiers (e.g. PER0001) so general staff cannot link sensitive records to specific individuals.

Mentor tip

Focus on access control levels/views and data anonymization (separating PII from operational tables) rather than general login passwords or encryption.

Question 3

15 marks

Artisan College student course relation 1NF analysis, 3NF normalization, and data modelling

Databases

(a)

2 marks

Outline why the STUDENT table violates First Normal Form (1NF):
1) 1NF requires every attribute cell to contain only a single atomic value.
2) In the given table, CourseNum and CourseName contain repeating groups / multiple values within a single record (for example, StudentID 2907 has three CourseNum values and three CourseName values stacked in one row).

Mentor tip

Point out both the rule (atomic values / no repeating groups) and the specific violating columns (CourseNum and CourseName).

(b)

8 marks

Construct the 3rd Normal Form (3NF) schema across four tables, clearly identifying each Primary Key [PK] and Foreign Key [FK]:
1) Isolate the transitive dependency between MajorID and Major into a MAJOR table:
MAJOR (MajorID [PK], Major)
2) Store student details in the STUDENT table referencing MajorID:
STUDENT (StudentID [PK], FName, SName, MajorID [FK])
3) Isolate course details (eliminating partial dependency on CourseNum) into a COURSE table:
COURSE (CourseNum [PK], CourseName)
4) Resolve the many-to-many enrolment relationship between students and courses using an ENROLLMENT junction table with a composite primary key:
ENROLLMENT (StudentID [PK, FK], CourseNum [PK, FK])

Mentor tip

Ensure you have 4 distinct tables (STUDENT, MAJOR, COURSE, ENROLLMENT), with correct attributes and underlined/labelled Primary Keys for each.

(c)

5 marks

Explain why data modelling is essential when designing a database:
1) Aligning with Business Purpose: Data modelling clarifies system requirements upfront so the database captures all necessary entities without omitting critical operational data.
2) Defining Entities and Attributes: It establishes the exact tables, atomic fields, and appropriate data types needed to avoid storage inefficiencies.
3) Identifying Primary and Foreign Keys: It specifies unique keys and relational links so records can be retrieved unambiguously.
4) Mapping Relationships for Complex Queries: Defining 1:M and M:N cardinalities ensures tables can be joined accurately for reporting.
5) Eliminating Redundancy via Normalization: Proper logical modelling removes duplicate data and prevents insertion, update, and deletion anomalies.

Mentor tip

Cover identifying entities, attributes, keys, relationships for multi-table queries, and avoiding redundancy/anomalies.

Question 4

20 marks

Lacto Dairy Products data warehouse, network data model, ETL data scrubbing, segmentation, and analytics

Databases

(a)

2 marks

Define a data warehouse:
1) A large, centralized repository of integrated, subject-oriented, time-variant, and non-volatile historical data consolidated from multiple disparate operational sources.
2) Designed specifically to support Online Analytical Processing (OLAP), data mining, and strategic management decision-making.

Mentor tip

Include its core characteristics (integrated, historical/time-variant, non-volatile repository) and its purpose (supporting OLAP/management decision-making).

(b)

4 marks

Characteristic 1 — Multiple Parent-Child Ownership (Graph Structure):
1) Unlike a strict hierarchical tree where each child node has only one parent, a network data model organizes records as a web-like graph where a child record (member) can have multiple parent records (owners), naturally modelling many-to-many relationships.
Characteristic 2 — Direct Pointer / Link Navigation Paths:
2) Logical associations between record types are maintained via explicit physical pointers (sets/links), providing multiple fast traversal paths directly between connected nodes.

Mentor tip

Highlight that child records can have multiple parent owners (supporting M:N relationships) and nodes are connected in a web/graph structure.

(c)

4 marks

Way 1 — Standardizing Formats and Decoding Values (Data Scrubbing):
1) During the Transform stage, ETL converts inconsistent representations across source systems into a uniform standard (such as converting 'Male'/'Female' to 'M'/'F', unifying date formats, or mapping cryptic numeric codes into meaningful labels like 'Wholesaler' and 'Retailer').
Way 2 — Deduplication and Handling Missing/Outlier Values:
2) ETL identifies and removes duplicate records merged from different sources, flags out-of-bounds numerical errors, and populates missing/null attributes with sensible default or mean values.

Mentor tip

Describe two concrete data-cleaning actions: standardizing inconsistent formats/codes and removing duplicates or fixing missing/erroneous values.

(d)

5 marks

Advantages of Data Segmentation for Lacto Dairy Products:
1) Targeted Pricing and Promotions: Partitioning data into distinct customer segments (such as wholesalers vs retailers, or by geographic region) allows Lacto to tailor bulk pricing tiers, product packaging, and marketing campaigns to each group's purchasing behaviour.
2) Efficient Resource Allocation and Retention: Helps management identify high-value segments, anticipate specific demand patterns, and improve customer retention.
Disadvantages of Data Segmentation:
3) Over-Generalization Within Segments: Assumes all buyers within a segment (e.g. all retailers in a region) share identical needs, potentially overlooking niche opportunities or innovative product lines.
4) Increased Management Cost and Complexity: Collecting, maintaining, and analysing multiple segmented datasets increases administrative overhead and data management costs (and over-segmenting can fragment strategic focus).
Conclusion:
5) Data segmentation is highly effective for differentiating broad pricing structures between wholesalers and retailers, provided the company periodically reviews sub-segments so it does not miss emerging market trends.

Mentor tip

Provide two clear advantages (with reference to wholesalers/retailers), two disadvantages (such as over-generalization and higher cost/complexity), and a balanced conclusion.

(e)

5 marks

Explain how data analytics assists Lacto Dairy Products managers in future decision-making:
1) Predictive Forecasting: Analytics and machine learning models examine historical sales time series to forecast future seasonal demand for specific dairy products.
2) Scenario and Pricing Analysis: Managers can evaluate how discount structures or price changes impact purchasing volume across wholesalers and retailers.
3) Product Portfolio Optimization: Association rule mining and profitability analysis reveal which product lines should be expanded, bundled, or discontinued.
4) Processing Massive Data Volumes: Automated data mining extracts actionable patterns and correlations from vast warehouse datasets that would be impossible to spot manually.
5) Evidence-Based Strategy: Replaces guesswork with quantitative business intelligence for supply chain and inventory planning.

Mentor tip

Connect data mining/forecasting to concrete dairy management decisions (predicting seasonal demand, testing discount impacts, pruning unprofitable product lines).

Option B: Modelling and simulation · Option

Questions 5–8 · 65 marks

Question 5

21 marks

Restaurant monthly electricity bill computer model, variables, and min/max usage pseudocode

Modelling and simulation

(a)

1 mark

Define a computer model:
A mathematical and algorithmic representation of a real-world system or process implemented as a computer program to calculate outcomes or test scenarios.

Mentor tip

Define it as a programmatic/mathematical representation of a real-life situation.

(b)

3 marks

State three variables and their data types for the electricity model:
1) UNITS_USED (or DAYS_IN_MONTH): Integer
2) MONTH_NAME: String
3) TOTAL_BILL (or DAILY_CHARGE / UNIT_COST / TAX_RATE): Real / Float / Double

Mentor tip

Always pair each variable name with its explicit data type (Integer, String, Real/Float).

(c)

5 marks

Construct pseudocode to input monthly units and days, then calculate and output the total bill (including 10% tax) and daily average usage:
```
UNIT_COST = 0.20
DAILY_CHARGE = 0.75
TAX_RATE = 0.10
output "Enter electricity units used in the month: "
input UNITS_MONTH
output "Enter number of days in the month: "
input DAYS_MONTH
SUBTOTAL = (UNITS_MONTH * UNIT_COST) + (DAYS_MONTH * DAILY_CHARGE)
TOTAL_BILL = SUBTOTAL + (SUBTOTAL * TAX_RATE)
UNITS_PER_DAY = UNITS_MONTH / DAYS_MONTH
output "Total monthly electricity bill: $", TOTAL_BILL
output "Average electricity units used per day: ", UNITS_PER_DAY
```

Mentor tip

Include prompts/messages with your inputs and outputs, remember to add the 10% tax to the bill, and divide units by days for daily usage.

(d)

2 marks

Outline one limitation of this model for helping the owners reduce costs:
1) The model only tracks aggregate monthly/daily electricity consumption across the entire building.
2) It provides no breakdown by individual appliance or subsystem (cooking vs refrigeration vs air conditioning), so the owners cannot identify which specific equipment is wasting energy.

Mentor tip

Point out that aggregate monthly totals do not isolate which specific appliances or times of day consume the most electricity.

(e)

2 marks

State two ways the model could be implemented:
1) Spreadsheet software (using built-in formulas and functions like MIN/MAX/INDEX/MATCH).
2) Custom program code written in a high-level programming language (using arrays and search/sort algorithms).

Mentor tip

State 'Spreadsheet' and 'Program code / High-level programming language'.

(f)

8 marks

Construct pseudocode using parallel arrays to identify and output the month names with the highest and lowest electricity usage:
```
// Parallel arrays storing the 12 month names and their corresponding unit usage
MONTHS = ["January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December"]
USAGE = [5800, 5000, 5200, 4800, 4750, 4900, 5500, 6500, 5600, 5960, 5850, 6280]
MAX_USAGE = USAGE[0]
MIN_USAGE = USAGE[0]
HIGH_MONTH = MONTHS[0]
LOW_MONTH = MONTHS[0]
loop I from 1 to 11
if USAGE[I] > MAX_USAGE then
MAX_USAGE = USAGE[I]
HIGH_MONTH = MONTHS[I]
end if
if USAGE[I] < MIN_USAGE then
MIN_USAGE = USAGE[I]
LOW_MONTH = MONTHS[I]
end if
end loop
output "Month with highest electricity usage: ", HIGH_MONTH
output "Month with lowest electricity usage: ", LOW_MONTH
```

Mentor tip

Use two 1D parallel arrays (MONTHS and USAGE), initialize min/max trackers, loop across all 12 months updating both the extreme value and the corresponding month name, and output the month names.

Question 6

15 marks

Government budget resource allocation simulation, data collection accuracy, and citizen privacy ethics

Modelling and simulation

(a)

2 marks

Describe the difference between a model and a simulation in this government budgeting scenario:
1) The model is the static set of mathematical equations and rules defining the relationships between population size, tax revenue, and healthcare/education spending.
2) The simulation is the dynamic execution of that model over time using varied demographic and tax rate inputs to observe how future national budgets behave.

Mentor tip

Contrast the underlying mathematical rules/relationships (model) with running those rules over time with changing tax/population inputs (simulation).

(b)(i)

2 marks

Outline one advantage of using a simulation for future government budget allocation:
1) Enables policymakers to test "what-if" tax and demographic scenarios safely in advance, identifying potential funding shortfalls in healthcare or education years before they occur.
2) Allows proactive policy adjustments to prevent budget deficits without risking real public services.

Mentor tip

Connect testing future scenarios in advance to preventing healthcare/education budget shortages.

(b)(ii)

2 marks

Outline one disadvantage of relying on a simulation for government resource allocation:
1) If the underlying mathematical assumptions are flawed or the input data (collected only every 3 years) is outdated or incomplete, the simulation will produce inaccurate projections.
2) Relying on faulty predictions can lead to misallocated tax funds and underfunded public hospitals or schools.

Mentor tip

Explain how flawed assumptions or outdated inputs lead directly to misallocated public budgets.

(c)

4 marks

Improvement 1 — Increase Data Collection Frequency:
1) Collect demographic and economic data annually or continuously in real time rather than only once every three years, allowing the simulation to capture rapid economic shifts or sudden population changes accurately.
Improvement 2 — Expand the Scope and Granularity of Variables:
2) Gather additional relevant variables (such as age-group distribution, inflation rates, employment levels, and regional disease trends) to reduce the impact of unmodelled confounding factors.

Mentor tip

Propose two clear improvements: collecting data more frequently than every 3 years, and capturing a broader set of demographic/economic variables.

(d)

5 marks

Reasons Why Citizen Ethical Concerns ARE Justified:
1) Invasion of Privacy and Function Creep: Detailed personal data on income, health, and family size might be repurposed by government agencies for surveillance or shared with unauthorized third parties beyond budget modelling.
2) Data Breach Vulnerability: Centralizing sensitive citizen records creates a high-value target for cyberattacks; a breach could expose confidential medical and tax histories.
Reasons Why Government Data Collection IS Justified:
3) Essential for Accurate Public Planning: High-resolution data on earnings, demographics, and healthcare usage is vital to build an accurate simulation that ensures hospitals and schools receive fair, adequate funding.
4) Anonymization Mitigates Harm: If data is properly aggregated and stripped of personally identifiable markers before entering the simulation, individual privacy is preserved.
Balanced Appraisal:
5) Citizen concerns are valid regarding raw personal records, so the government must enforce strict anonymization, encryption, and legal data-protection boundaries to balance privacy rights against effective public resource planning.

Mentor tip

Discuss both sides (privacy/surveillance/breach risks vs the public benefit of accurate healthcare and education funding) and conclude with how anonymization balances the trade-off.

Question 7

9 marks

2D vs 3D visualization examples, keyframe animation, and hardware resource implications

Modelling and simulation

(a)(i)

1 mark

State one example of a 2D visualization:
A flat 2D map, line graph, bar chart, blueprint diagram, or 2D sprite image.

Mentor tip

Give any clear flat two-dimensional graphical representation (such as a 2D map or chart).

(a)(ii)

1 mark

State one example of a 3D visualization:
A 3D wireframe/polygon model, 3D CGI architectural walkthrough, or 3D virtual reality environment.

Mentor tip

Give any three-dimensional graphical representation with depth/perspective.

(b)

2 marks

Describe the significance of a key frame in 3D visualization/animation:
1) A key frame defines the exact position, pose, and attributes of 3D objects at a critical starting or ending point of a movement sequence.
2) Software interpolation (tweening) automatically calculates and generates all the intermediate frames between two key frames to create a smooth transition.

Mentor tip

State that keyframes define the start and end states of a movement so the software can interpolate the smooth transition frames between them.

(c)

5 marks

Explain the implications of 3D animation on computer hardware resources during game creation:
1) Intensive GPU/CPU Processing: Rendering 3D polygon meshes, calculating lighting/shadows (ray tracing or rasterization), and interpolating character movement across scenes demand massive parallel floating-point computation from the GPU and CPU.
2) Compositing Multiple Layers: Combining background environments, moving character models, particle effects, and textures into each final frame increases rendering latency.
3) High RAM / VRAM Capacity Requirements: Storing high-resolution texture maps, 3D geometry buffers, and multiple animation frames simultaneously requires substantial primary memory and dedicated video memory (VRAM).
4) Secondary Storage Bandwidth: Loading large 3D scene assets and saving rendered animation sequences requires fast SSD storage and high memory bus bandwidth to avoid bottlenecks.
5) Extended Rendering Times: High-detail scenes require long render farms or compilation times if hardware resources are insufficient.

Mentor tip

Address CPU/GPU processing load (rendering polygons, lighting, and layer transitions), RAM/VRAM capacity for textures/frames, and storage speed.

Question 8

20 marks

Autonomous vehicle genetic algorithms, supervised vs unsupervised voice commands, NLP, and human vs machine language learning

Modelling and simulation

(a)

2 marks

Outline how a genetic algorithm is iterative:
1) Starting from an initial population of potential solutions, the algorithm repeatedly loops through successive generations.
2) In each iteration (generation), individuals are evaluated, selected, crossed over, and mutated to progressively refine the population toward an optimal driving behaviour.

Mentor tip

Explain that the algorithm repeats a cycle of evaluation, selection, crossover, and mutation across multiple generations.

(b)

2 marks

Outline the role of the fitness function in a genetic algorithm:
1) A quantitative objective function that evaluates each potential solution in a generation and assigns a numerical score measuring how well it achieves the target criteria (e.g. safe, smooth navigation).
2) This score determines which high-performing individuals are selected to reproduce into the next generation.

Mentor tip

State that it quantitatively scores how close each potential solution is to the optimum to guide selection.

(c)(i)

2 marks

Outline one benefit of supervised learning for learning verbal driving commands:
1) Uses a labelled dataset mapping known voice commands to exact, verified vehicle actions.
2) Ensures the autonomous vehicle's response to critical safety commands (like "stop" or "pull over") is highly predictable, reliable, and verifiable.

Mentor tip

Link labelled input-output training pairs to predictable, safety-verified vehicle responses.

(c)(ii)

2 marks

Outline one benefit of unsupervised learning for learning verbal driving commands:
1) Can discover patterns in unlabelled natural speech without requiring every possible phrasing to be pre-labelled in advance.
2) Allows the vehicle to adapt to novel accents, regional dialects, informal phrasing, or emotional vocal tone (such as an urgent shout warning of a hazard).

Mentor tip

Link unlabelled pattern discovery to handling unforeseen phrasing, new dialects, or urgent vocal tones.

(d)

5 marks

Explain how advances in Natural Language Processing (NLP) enable appropriate responses to verbal commands:
1) Acoustic and Phonemic Recognition: Deep neural speech-to-text models accurately transcribe spoken words even over background road and engine noise.
2) Semantic Parsing and Intent Classification: Modern NLP extracts the underlying meaning and intent of a sentence rather than relying on rigid keyword matching (e.g. understanding "it's too chilly in here" as a command to raise the cabin temperature).
3) Contextual Disambiguation: Transformer/attention models use conversational and environmental context to resolve ambiguous words or pronouns.
4) Sentiment and Urgency Detection: Prosody and tone analysis detect panic or urgency in the passenger's voice (e.g. shouting "Watch out!"), triggering immediate emergency braking.
5) Low-Latency Edge Processing: Optimized NLP models execute in real time on the vehicle's onboard processors so safety-critical actions occur without network delay.

Mentor tip

Cover semantic intent extraction, noise/accent robustness, urgency/tone detection, and real-time execution for safety.

(e)

5 marks

Explain why human language acquisition differs from machine language learning:
1) Human Learning — Meaning, Social Context, and Grammar Rules:
- Humans first acquire language through immersive social interaction, grounding words in real-world sensory experiences and physical context.
- Humans internalize grammatical structures and generative syntax rules, enabling them to understand metaphors, sarcasm, and novel expressions with very few examples.
2) Machine Learning — Statistical Pattern Matching on Massive Corpora:
- Machine language models do not possess conscious comprehension or lived sensory experience; instead, they compute statistical probability distributions over vast datasets of text and audio tokens.
- Machines match input sequences to stored mathematical patterns and weights, which can struggle when human speech violates formal syntax or uses novel cultural idioms.

Mentor tip

Contrast human learning (experiential meaning, social context, and grammatical rule internalization) with machine learning (statistical probability matching across massive datasets).

(f)

2 marks

Describe the use of a chatbot:
1) A conversational software agent designed to simulate natural human dialogue via text or speech interfaces.
2) Commonly deployed in customer service, virtual assistants, or entertainment to automatically answer routine queries, guide users, or execute tasks 24/7.

Mentor tip

State both what a chatbot does (simulates human conversation) and a practical application (answering customer queries or virtual assistance).

Option C: Web science · Option

Questions 9–12 · 65 marks

Question 9

10 marks

URLs, DNS resolution, HTTPS characteristics, protocols vs standards, and connectionless IP

Web science

(a)

1 mark

Define a Uniform Resource Locator (URL):
The unique human-readable address used to specify the location of a resource (such as a web page or file) on the World Wide Web / Internet and the protocol used to retrieve it.

Mentor tip

State clearly that a URL is the unique address of a resource/webpage on the internet.

(b)

3 marks

Describe how a Domain Name Server (DNS) functions:
1) When a user enters a URL in a browser, the DNS receives the human-readable domain hostname and looks it up in its database to resolve it to the corresponding numeric IP address.
2) If the local/primary DNS server does not have the IP cached, it forwards the query hierarchically to root, Top-Level Domain (TLD), and authoritative DNS servers until a match is found.
3) Once resolved, the DNS returns the IP address to the client browser so TCP/IP packets can be routed to the target web server (or returns an error if the domain does not exist).

Mentor tip

Cover translating the domain name to an IP address, querying higher-level DNS servers if not cached locally, and returning the IP to the browser.

(c)

2 marks

Identify two characteristics of HTTPS:
1) Encrypts all HTTP communication traffic between client and server using TLS/SSL (symmetric session encryption negotiated via asymmetric keys).
2) Authenticates the identity of the web server using a verifiable digital SSL/TLS certificate issued by a trusted Certificate Authority.

Mentor tip

Mention TLS/SSL encryption of data in transit and server authentication via digital certificates.

(d)

2 marks

Distinguish between a network protocol and a technical standard:
1) A protocol is a formal set of procedural rules governing how data packets are formatted, transmitted, received, and acknowledged during communication across a network (e.g. HTTP, TCP, IP).
2) A standard is a broader, agreed-upon technical specification for hardware, software, data formats, or languages (e.g. HTML5, Unicode, IEEE 802.11) that ensures universal compatibility and interoperability across platforms.

Mentor tip

Contrast rules for transmitting/exchanging data over a network (protocol) with broader specifications for formats, languages, or hardware interoperability (standard).

(e)

2 marks

Outline what connectionless means in the context of IP packet transmission:
1) Data packets (datagrams) are transmitted onto the network without any prior handshake or dedicated end-to-end circuit being established between the client and server.
2) Each packet carries its own complete destination IP address in its header and is routed independently across network hops (potentially arriving out of order).

Mentor tip

Highlight that there is no prior handshake/setup circuit and each packet is addressed and routed independently.

Question 10

19 marks

Server-side PHP/SQL database insertion, CGI script execution, SEO, link ranking, and crawling the deep web

Web science

(a)(i)

3 marks

Describe the process executed by the PHP/SQL code fragment:
1) Constructs an SQL INSERT statement to add a new record ('Sonia', 'Smith', 'sonia@smith.com', '+61231456789') into the FirstName, Surname, Email, and Cellphone columns of GuestsTable.
2) Executes the query on the database connection ( $co nn - > q u er y ($ sql)) and checks whether the execution succeeded (=== TRUE).
3) Outputs "New record created successfully" to the web page if the insert succeeded, or outputs an error message containing the SQL string and database connection error details if it failed.

Mentor tip

Mention inserting the four values into GuestsTable, executing the query via $conn, and echoing either the success message or the error details.

(a)(ii)

3 marks

Describe how the Common Gateway Interface (CGI) enables server-side script execution:
1) When a web browser sends an HTTP request for a dynamic resource, the web server uses the CGI specification to pass the client's request data (such as form inputs or query strings) to an external script/executable on the server.
2) The server-side CGI script executes at runtime (e.g. interacting with a database).
3) The script outputs dynamically generated HTML back through the web server, which sends the rendered response to the client's browser.

Mentor tip

Trace the three steps: client request passed by web server to CGI script, script execution on the server, and dynamic HTML returned to the client.

(b)(i)

2 marks

Identify two Search Engine Optimization (SEO) techniques:
1) Including relevant, descriptive keywords in HTML <title>, heading tags, image alt attributes, and <meta> description/keyword tags.
2) Building high-quality incoming and outgoing hyperlinks (backlinks from authoritative websites), maintaining clean mobile-responsive HTML, and updating site content regularly.

Mentor tip

Name two specific technical practices (such as relevant meta-tags/alt attributes and authoritative backlinks/regular content updates) rather than generic labels like 'white hat'.

(b)(ii)

4 marks

Explain why incoming and outgoing links are significant factors in search engine ranking:
1) Core Ranking Input: Search algorithms such as PageRank and HITS rely heavily on the hyperlink graph structure to evaluate page importance.
2) Incoming Links as Votes of Authority: In PageRank, an incoming link (backlink) from another website acts as a vote of confidence; links coming from high-authority pages pass greater weight to the target page.
3) Outgoing Links and Hub Value: In HITS, a page that links out to many reputable, relevant resources achieves a high Hub score, which in turn reinforces the Authority scores of the pages it cites.
4) Vulnerability to Manipulation: Because link counts strongly boost rankings, some site operators attempt black-hat manipulation such as link farms or comment spam, forcing search engines to evaluate link quality alongside quantity.

Mentor tip

Reference both PageRank (incoming links as weighted votes of authority) and HITS (hubs and authorities).

(c)

3 marks

Describe how a web crawler indexes a webpage:
1) Starting from a list of known seed URLs, the crawler visits the webpage (checking robots.txt permissions and sitemaps first).
2) It parses the HTML document, extracts body text, headings, and <meta> tags, and stores these keywords alongside the URL in the search engine's inverted index database.
3) It extracts all outgoing hyperlinks (<a href="...">) on the page and adds those newly discovered URLs to its crawl queue to visit next (using breadth-first or depth-first traversal).

Mentor tip

Cover starting from seed URLs/checking robots.txt, parsing and storing keywords/metadata in the index, and queuing outgoing hyperlinks for further crawling.

(d)

4 marks

Suggest why the inability to access the deep web is an increasing problem for search engines:
1) Shift from Static to Dynamic Database-Driven Sites: A rapidly growing proportion of modern websites generate content dynamically on the fly from backend databases only in response to specific user queries or form submissions, leaving no static URLs for crawlers to follow.
2) Proliferation of Paywalls and Authenticated Portals: More high-value information (academic research, social media feeds, subscription news, cloud repositories) is locked behind login screens or paywalls.
3) Exponential Growth of Unindexed Data: As deep web databases expand much faster than the static surface web, traditional crawlers miss the vast majority of online information.
4) Incomplete and Less Relevant Search Results: Users relying on surface-web search engines receive an increasingly narrow, superficial subset of available human knowledge, reducing search comprehensiveness and reliability.

Mentor tip

Focus consistently on either the technical barriers (proliferation of dynamic database-driven pages, logins, paywalls, and exponential growth) or the consequential impact on search completeness.

Question 11

16 marks

Peer-to-peer (P2P) networks, lossy compression for audio vs eBooks, cloud computing, and IP protection

Web science

(a)

2 marks

Identify two characteristics of a Peer-to-Peer (P2P) network:
1) Decentralized Architecture: There is no single central server storing the files; nodes communicate directly with one another, eliminating a single point of failure.
2) Dual Client-Server Role: Every connected peer simultaneously acts as both a client (downloading data chunks from other peers) and a server (uploading data chunks and sharing bandwidth).

Mentor tip

State that P2P is decentralized without a central server and that each peer acts as both client and server (supplying and consuming resources).

(b)

6 marks

General Effect of Lossy Compression:
1) Permanently discards data deemed less perceptible to achieve a much smaller file size, resulting in faster download times over mobile networks and lower storage requirements.
Appropriateness for the Music Album (Audio):
2) Highly Appropriate: Psychoacoustic lossy formats (like MP3 or AAC) remove frequencies outside the range of human hearing or masked by louder sounds.
3) At reasonable bitrates, the file size is drastically reduced (~90%) with negligible perceptible loss in audio quality for mobile listeners (though extreme compression causes audible distortion).
Appropriateness for the eBook (Text and Embedded Images):
4) Inappropriate for Text: Applying lossy compression to character text is completely unsuitable because discarding bytes corrupts characters and renders the book unreadable (text requires lossless compression).
5) Mixed for Embedded Cover/Photo Images: Lossy compression (like JPEG) can shrink photographic illustrations inside the eBook, though excessive compression introduces blocky pixelation, colour banding, and blurry text overlays.
Overall Judgment:
6) Lossy compression is well-suited for streaming/downloading the audio album on mobile devices, but the eBook should use lossless compression for its text and moderate compression only for photographic images.

Mentor tip

Evaluate both media types separately: explain why lossy works well for audio (removing inaudible frequencies) but corrupts eBook text or pixelates heavily compressed images.

(c)

2 marks

Identify two characteristics of cloud computing:
1) On-Demand Remote Hosting: Storage, processing, and software applications (SaaS/PaaS/IaaS) are hosted on remote internet servers and accessible from any location or device.
2) Elastic Scalability and Pay-As-You-Go: Computing capacity and storage scale dynamically to match user demand without the user managing physical hardware.

Mentor tip

Mention internet-hosted location-independent services (SaaS/IaaS) and elastic scalability.

(d)

6 marks

Evaluation of Method 1 — Protecting Images Behind a Registered User Login:
1) Strengths: Enforced on the server side, preventing anonymous public visitors and automated web scrapers from accessing the portfolio; allows the artist to authenticate users and track who views the images.
2) Weaknesses: Creates friction that deters casual fans or potential buyers unwilling to register; blocks search engine crawlers from indexing the artwork (reducing publicity); and logged-in users can still screenshot or redistribute downloaded files.
Evaluation of Method 2 — Client-Side Script Disabling the Right-Click 'Save Image' Menu:
3) Strengths: Allows the public and search engines to browse and discover the gallery freely without login friction, while deterring casual, non-technical visitors from saving copies with a simple right-click.
4) Weaknesses: Easily bypassed because the script runs on the client side—users can disable JavaScript in their browser, extract the image directly from the browser cache/developer tools, or take a screen capture.
Overall Conclusion:
5) Neither method guarantees absolute protection once an image is displayed on a client screen; combining public low-resolution watermarked previews with server-side authenticated login for high-resolution files provides a more balanced solution.

Mentor tip

Give a balanced 3+3 evaluation covering both server-side login (secure access vs reduced visibility/signup friction) and client-side script (easy public browsing vs easily bypassed via cache/screenshots/disabling JS).

Question 12

20 marks

Bowtie web graph model, SCC characteristics, directed sub-graphs, IN/OUT evolution, and ontologies vs folksonomies

Web science

(a)(i)

1 mark

Define a node in the bowtie web graph:
A single web page (or website) on the World Wide Web.

Mentor tip

State clearly that a node represents a webpage or website.

(a)(ii)

1 mark

Define an edge in the bowtie web graph:
A directed hyperlink pointing from one web page (node) to another.

Mentor tip

State clearly that an edge represents a directed hyperlink between two pages.

(b)

2 marks

Identify two characteristics of a node in the Strongly Connected Core (SCC):
1) Mutual Reachability: Any node in the SCC can reach any other node in the SCC by following a directed path of hyperlinks (and therefore has both incoming and outgoing links).
2) Component Relationships: An SCC node can be reached from IN nodes and can reach OUT nodes via directed paths, but cannot navigate back to IN nodes or return from OUT nodes.

Mentor tip

Mention that SCC nodes have both incoming and outgoing links such that every SCC node can reach every other SCC node (and can reach OUT nodes but not IN nodes).

(c)

4 marks

Construct the directed sub-graph with circles for nodes {A..I} and directed arrows for edges:
1) IN Component (Nodes A, B, C): Draw circles A, B, and C with directed arrows pointing into the SCC (e.g. A -> D, B -> E, C -> D) and no arrows returning from the SCC to A, B, or C.
2) SCC Component (Nodes D, E, F, G): Draw circles D, E, F, and G connected in a strongly connected directed cycle (e.g. D -> E -> F -> G -> D) so every node in {D, E, F, G} can reach every other node in the core.
3) OUT Component (Nodes H, I): Draw circles H and I with directed arrows pointing from the SCC into them (e.g. F -> H, G -> I) and no arrows pointing back into the SCC.
4) Tube Connection (C to I): Draw a direct arrow from IN node C to OUT node I (C -> I) bypassing the SCC.

Mentor tip

Ensure D, E, F, G form a closed directed loop (SCC), A, B, C point into {D,E,F,G}, {D,E,F,G} point out to H and I, and include the direct tube arrow C -> I.

(d)

4 marks

Suggest why the IN and OUT proportions changed from 21% each in 2000 to 32% IN and 6% OUT in 2012:
1) Shrinkage of OUT (21% to 6%) due to Ubiquitous Outbound Linking: Modern websites rarely act as dead-end terminal pages; social media buttons, advertising networks, partner links, and SEO best practices cause previously isolated OUT sites to link back to major platforms in the SCC, pulling them into the SCC.
2) Growth of IN (21% to 32%) due to Easy Self-Publishing: Blogs, personal sites, and new small-business pages can be created effortlessly and immediately link out to authoritative SCC sites (like Wikipedia, YouTube, or news portals), but receive few or no incoming links from established SCC sites.
3) Expansion and Density of the SCC (27% to 51%): Social networking, collaborative wikis, and content-sharing platforms have made the core of the web much more densely interconnected.
4) Crawl Scale and Methodology Differences: The 2012 crawl covered 3.5 billion nodes compared to ~200 million in 2000, capturing a much larger long-tail of newly published leaf pages (IN).

Mentor tip

Explain both shifts: why OUT shrank (few dead-end sites remain because ads/social widgets link back to the SCC) and why IN grew (millions of new blogs/small sites link to popular SCC hubs without receiving backlinks).

(e)

2 marks

Identify two characteristics of collective intelligence:
1) Shared Group Knowledge: Emerges from the decentralized collaboration, crowdsourcing, and combined contributions of large communities of users over the web.
2) Continuous Co-Creation: Participants actively generate, edit, review, and refine shared resources (such as Wikipedia articles or open-source repositories on GitHub).

Mentor tip

State that it is shared knowledge derived from mass collaboration/crowdsourcing and give an example or characteristic of co-creation.

(f)

6 marks

Significance and Strengths of Ontologies in the Semantic Web:
1) Formal Structure: An ontology formally defines a domain's classes, properties, and relationships using standardized machine-readable frameworks (such as RDF, OWL, and Dublin Core).
2) Automated Reasoning and Interoperability: Because ontological vocabulary is unambiguous, consistent, and hierarchical, software agents can automatically query, integrate, and infer new knowledge across heterogeneous databases.
3) Limitation of Ontologies: Designing and maintaining strict ontologies requires specialized expertise, high cost, and rigid agreement, making them slow to adapt to new topics.
Significance and Strengths of Folksonomies:
4) User-Driven Collaborative Tagging: A folksonomy arises when everyday users freely annotate web content (photos, videos, posts) with informal, uncontrolled keywords or hashtags.
5) Low Barrier and Adaptability: Folksonomies are inexpensive, democratic, and rapidly reflect emerging trends and user vocabulary without requiring central coordination.
6) Limitation of Folksonomies for the Semantic Web: Because tags are uncontrolled, they suffer from spelling variations, synonyms, homonyms, and ambiguity, making automated machine reasoning and consistent semantic indexing difficult.

Mentor tip

Discuss both ontologies (formal RDF/OWL structure enabling machine reasoning, but rigid/complex) and folksonomies (flexible user tagging, but ambiguous/inconsistent for machines).

Option D: Object-oriented programming · Option

Questions 13–16 · 65 marks

Question 13

14 marks

Real estate UML class diagram, primitive types, relationships, static counters, and object arrays

Object-oriented programming

(a)

1 mark

Define a primitive data type:
A fundamental, predefined basic data type built directly into the programming language (such as int, boolean, double, char) that stores a single raw value directly in memory rather than an object reference.

Mentor tip

State that it is a basic, built-in language type that holds a raw value directly in memory (and serves as a building block for composite object types).

(b)(i)

1 mark

State a suitable Boolean attribute for Property:
isSold (or hasGarden, hasParking, hasPool, isAvailable).

Mentor tip

Name any realistic true/false property attribute such as isSold or hasGarage.

(b)(ii)

1 mark

State a suitable integer attribute for Property:
numBedrooms (or numRooms, numBathrooms, floorArea, yearBuilt).

Mentor tip

Name any realistic whole-number property attribute such as numBedrooms or area.

(c)(i)

1 mark

State the relationship between Owner and Property:
Aggregation (or Association / "has-a" relationship).

Mentor tip

Look at the open diamond on the UML link between Owner and Property, which denotes aggregation.

(c)(ii)

1 mark

State the relationship between House and Property:
Inheritance ("is-a" relationship — House is a subclass of the superclass Property).

Mentor tip

Look at the hollow triangle arrow pointing from House up to Property, which denotes inheritance.

(d)

2 marks

Distinguish between a class and an instantiation in this real estate scenario:
1) A class is the abstract blueprint or template that defines the attributes and methods common to a category (for example, the House or Property class definition).
2) An instantiation is a concrete object created in memory at runtime from that class blueprint containing actual attribute values (for example, a specific House instance at "12 Oak Street" priced at $350,000).

Mentor tip

Provide both the theoretical distinction (blueprint vs concrete object in memory) and a scenario-specific example (the House class vs a specific house object with an address/price).

(e)

2 marks

Outline how the static modifier affects a variable:
1) The variable belongs to the class itself rather than to individual object instances, so a single shared copy exists across all instances of that class.
2) It can be accessed or updated directly via the class name without needing to instantiate an object, and any change to its value is seen by all instances.

Mentor tip

Highlight that a static variable is a single class-level variable shared by all instances.

(f)

2 marks

Describe how static variables can track the total number of House and Apartment objects created:
1) Declare a static integer counter variable (e.g. private static int totalCount = 0;) in the Property superclass (or separate static counters in both House and Apartment).
2) Increment this static counter by 1 inside the constructor every time a new House or Apartment object is instantiated, and provide a public static accessor method to retrieve the combined count.

Mentor tip

Explain declaring a static integer counter and incrementing it inside the constructor(s) whenever a House or Apartment is instantiated.

(g)

3 marks

Construct the Java statement to declare and instantiate the wishList array:
```java
House[] wishList = new House[10];
```
1) Declares wishList with type House[].
2) Uses the new keyword.
3) Allocates an array of size 10: House[10].

Mentor tip

Include the exact array identifier wishList, the new operator, and the size 10 of type House.

Question 14

10 marks

Encapsulation, inheritance for student housing, and programming language internationalization

Object-oriented programming

(a)

1 mark

Define encapsulation:
Bundling data (attributes) and the methods that operate on that data into a single unit (class), while restricting direct external access to internal state using private visibility modifiers.

Mentor tip

Mention both bundling data with methods inside a class and hiding internal data via private attributes.

(b)

3 marks

Explain one benefit provided by encapsulation (Data Hiding / Protection):
1) By declaring class attributes as private, external classes cannot directly read or accidentally corrupt an object's internal state.
2) External code must interact with the attributes strictly through public accessor (getter) and mutator (setter) methods.
3) Mutator methods can enforce validation rules (e.g. ensuring a property price is never set to a negative number), improving program reliability and modularity.

Mentor tip

Develop a single cohesive chain of reasoning: private variables -> controlled access via getters/setters -> preventing accidental data corruption (or easier isolated maintenance).

(c)

3 marks

Explain one benefit of using inheritance to create the new Room class:
1) Code Reuse: By extending House (or Property), the Room subclass automatically inherits existing attributes (ID, address, city, price) and their accessor/mutator methods.
2) Faster, Less Error-Prone Development: Programmers do not need to rewrite or retest duplicated code, saving development time and only needing to add attributes unique to a student room.
3) Easier Maintenance: Any future bug fix or update made to the superclass is automatically propagated to Room and all other subclasses.

Mentor tip

Focus on inheriting existing attributes/methods from Property/House, which eliminates code duplication and saves development/testing time.

(d)

3 marks

Explain one feature of modern programming languages that supports internationalization:
1) Native Unicode Support: Modern languages (like Java) represent characters using Unicode (such as 16-bit UTF-16 or UTF-8) rather than legacy 7/8-bit ASCII.
2) Expanded Character Set: Unicode can encode tens of thousands of characters and symbols (2^16 = 65,536 in 16-bit, or over 1 million code points).
3) Global Alphabet Compatibility: This allows student names, addresses, and user interfaces to accurately store and display non-Latin scripts (such as Chinese, Arabic, Cyrillic, and accented characters) for international users.

Mentor tip

Identify Unicode (16+ bits vs 8-bit ASCII), state its large character capacity, and explain that it represents diverse international alphabets and symbols.

Question 15

21 marks

Method signatures, String immutability, null pointer handling, sorting House objects, and top-3 budget selection

Object-oriented programming

(a)

2 marks

Define a method signature:
1) The combination of the method's identifier (name) and its parameter list (the number, data types, and order of its parameters).
2) Together with the return type / access modifier in the method header, it uniquely identifies the method for invocation and overloading.

Mentor tip

Include both the method name/identifier and the parameter list (types and order).

(b)

2 marks

Describe how an original String variable passed to a method can be assigned a new value:
1) In Java, arguments are passed by value and String objects are immutable, so reassigning the parameter inside a void method will not modify the caller's original variable.
2) Instead, the method must declare a return type of String, return the newly constructed String value, and the caller must assign the method's return value back to the original String variable (e.g. myStr = modifyString(myStr);).

Mentor tip

Explain that Java uses pass-by-value / Strings are immutable, so the method must return a String that is assigned back to the original variable.

(c)

1 mark

State the intended purpose of unknown(String x):
To output/print the addresses of all houses in the allHouses array that are located in the specified city x.

Mentor tip

Make sure to state 'all houses' matching city x, not just a single house.

(d)(i)

2 marks

Outline the runtime error likely to occur when unknown(String x) is called:
1) Because allHouses is only partially filled, the loop iterating up to allHouses.length will eventually encounter unpopulated array elements that are null.
2) Calling allHouses[i].getCity() on a null element throws a NullPointerException at runtime.

Mentor tip

Link the partially filled array (null elements) to throwing a NullPointerException when invoking .getCity().

(d)(ii)

2 marks

Outline how this runtime error can be corrected:
1) Add a null check before invoking .getCity(): if (allHouses[i] != null && allHouses[i].getCity().equals(x)).
2) Alternatively, maintain an integer counter count of the actual number of House objects stored and loop only while i < count (or use a dynamic ArrayList<House>).

Mentor tip

Propose either guarding the condition with if (allHouses[i] != null) or stopping the loop at a count variable tracking populated elements.

(e)

5 marks

Construct the houseSort() method to sort allHouses in ascending order of price (using Bubble Sort or Selection Sort):
```java
public void houseSort() {
House temp;
for (int i = 0; i < allHouses.length - 1; i++) {
for (int j = 0; j < allHouses.length - 1 - i; j++) {
if (allHouses[j] != null && allHouses[j + 1] != null) {
if (allHouses[j].getPrice() > allHouses[j + 1].getPrice()) {
temp = allHouses[j];
allHouses[j] = allHouses[j + 1];
allHouses[j + 1] = temp;
}
}
}
}
}
```
1) Declares a temporary swap variable of type House.
2) Uses nested outer and inner loops to traverse the array.
3) Compares adjacent prices using allHouses[j].getPrice() > allHouses[j + 1].getPrice().
4) Swaps the complete House object references (not just the integer prices).

Mentor tip

Be sure to declare your swap variable as type House and swap the House objects in allHouses[j] and allHouses[j+1] based on .getPrice().

(f)

7 marks

Construct the selectThree(int budget) method returning the 3 most expensive houses <= budget in ascending order:
```java
public House[] selectThree(int budget) {
House[] result = new House[3];
houseSort();
int i = 0;
while (i < allHouses.length && allHouses[i] != null && allHouses[i].getPrice() <= budget) {
i++;
}
// i-1 is now the index of the most expensive house within budget
result[0] = allHouses[i - 3];
result[1] = allHouses[i - 2];
result[2] = allHouses[i - 1];
return result;
}
```
1) Declares the method with return type House[] and parameter int budget.
2) Instantiates House[] result = new House[3];.
3) Calls houseSort() first so allHouses is sorted in ascending order of price.
4) Advances index i while i < allHouses.length, allHouses[i] != null, and allHouses[i].getPrice() <= budget.
5) Copies the three highest-priced qualifying houses (at indices i - 3, i - 2, and i - 1) into result[0..2] in ascending order and returns result.

Mentor tip

Sort allHouses first with houseSort(), find the last index where getPrice() <= budget (checking for null!), and copy that element and the two preceding elements into your size-3 result array.

Question 16

20 marks

Abstract data types, singly linked list ClientList operations, list splitting, and recursive binary search

Object-oriented programming

(a)

2 marks

Identify two features of an Abstract Data Type (ADT):
1) Defined by Operations and Behaviour: An ADT is a logical model defined by the data it holds and the standardized set of operations performed on it (such as push, pop, enqueue, addToEnd).
2) Implementation Hiding (Information Hiding): The underlying memory representation and internal code implementation are hidden from the user, who only interacts with the public method interface.

Mentor tip

Highlight that an ADT is defined by its standardized operations/behaviour while hiding its internal implementation details from the user.

(b)

3 marks

Describe the meaning of openDay.enList(new ClientNode("Sophie Bella", "072456721"));:
1) Instantiates a new ClientNode object in memory.
2) Initializes that node with name "Sophie Bella" and phone "072456721" (with attended = false and next = null).
3) Calls enList on the openDay list to append this new ClientNode to the tail/end of the openDay linked list.

Mentor tip

Break the statement into three parts: instantiating a ClientNode, initializing it with Sophie Bella's details, and appending it to the end of the openDay list.

(c)

5 marks

Construct the enList(ClientNode visitor) method in ClientList:
```java
void enList(ClientNode visitor) {
if (root == null) {
root = visitor;
} else {
ClientNode current = root;
while (current.getNext() != null) {
current = current.getNext();
}
current.setNext(visitor);
}
}
```
1) Checks if the list is currently empty (root == null or isEmpty()).
2) If empty, assigns root = visitor.
3) Otherwise, initializes a traversal pointer ClientNode current = root.
4) Loops while current.getNext() != null, advancing current = current.getNext() until reaching the last node.
5) Links the new node at the tail by calling current.setNext(visitor).

Mentor tip

Handle the empty list case (root == null) first, then traverse using while (current.getNext() != null) so current stops on the actual tail node before calling current.setNext(visitor).

(d)

4 marks

Outline the steps to split openDay into hasAttended and notAttended without writing code:
1) Instantiate two new empty ClientList objects named hasAttended and notAttended.
2) Start a traversal pointer at the root node of openDay and iterate sequentially through every ClientNode until the end of the list (null) is reached.
3) At each node, inspect the boolean value of its attended attribute.
4) If attended is true, append a node with that client's details to hasAttended using enList; otherwise (if false), append it to notAttended.

Mentor tip

State clearly: instantiate both new ClientList objects, traverse every node in openDay, check the attended boolean flag, and enlist the node into hasAttended or notAttended accordingly.

(e)

6 marks

Construct the recursive binary search method binSearch(String name, int low, int high):
```java
public Owner binSearch(String name, int low, int high) {
if (high < low) {
return null;
}
int middle = (low + high) / 2;
Owner current = contacts.get(middle);
if (current.getName().equals(name)) {
return current;
} else if (name.compareTo(current.getName()) < 0) {
return binSearch(name, low, middle - 1);
} else {
return binSearch(name, middle + 1, high);
}
}
```
1) Base Case 1 (Not Found): Checks if high < low (search range exhausted) and returns null.
2) Midpoint Calculation & Access: Computes int middle = (low + high) / 2 and retrieves Owner current = contacts.get(middle).
3) Base Case 2 (Found): Checks if current.getName().equals(name) and returns the matching Owner object.
4) Recursive Step (Left Sublist): If name is alphabetically smaller (name.compareTo(current.getName()) < 0), recursively returns binSearch(name, low, middle - 1).
5) Recursive Step (Right Sublist): Otherwise, recursively returns binSearch(name, middle + 1, high).

Mentor tip

Include both base cases (high < low returning null, and equals(name) returning the Owner) and ensure your recursive calls pass middle - 1 and middle + 1 with return statements.

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
Object-oriented programming65 marks · 25%
Q13, Q14, Q15, Q16

Mentor insights

Traps to watch for

01
Databases: Quick Rent-a-Car relational database keys, SQL queries, validatio · Q1
Look at the crow's foot symbol on the RENTAL side of the ERD link.
02
Modelling and simulation: Restaurant monthly electricity bill computer model, variables, an · Q5
Define it as a programmatic/mathematical representation of a real-life situation.
03
Web science: Bowtie web graph model, SCC characteristics, directed sub-graphs, · Q12
State clearly that a node represents a webpage or website.
04
Object-oriented programming: Method signatures, String immutability, null pointer handling, so · Q15
Include both the method name/identifier and the parameter list (types and order).

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 2023
Marks: 260
Duration: 1 hour 20 min
Calculator: Not allowed
Questions: 16

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

Keep your copy of the paper open alongside. We don't reproduce the questions: each part below is labelled to match the May 2023 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

21 marks

Quick Rent-a-Car relational database keys, SQL queries, validation, views, DDL, and data consistency

Databases

(a)(i)

1 mark

State the relationship cardinality between CUSTOMER and RENTAL:
One-to-many (1:M) — one customer can enter into multiple rental agreements over time.

Mentor tip

Look at the crow's foot symbol on the RENTAL side of the ERD link.

(a)(ii)

1 mark

State the name of any one primary key from the schema:
CustID (in CUSTOMER), RentalNum (in RENTAL), or CarID (in CAR).

Mentor tip

Any one of the three underlined/key identifiers (CustID, RentalNum, or CarID) is sufficient.

(a)(iii)

1 mark

State the table that contains foreign keys:
RENTAL (which holds CustID referencing CUSTOMER and CarID referencing CAR).

Mentor tip

The junction/transaction table at the 'many' end of both 1:M relationships holds the foreign keys.

(b)

1 mark

Trace the SQL query against the CAR table (WHERE Man = "Lexus" AND Year = 2016):
Row DHE7623 matches both Man = "Lexus" and Year = 2016; its Make attribute is:
EF475

Mentor tip

Output only the projected Make value (EF475) rather than the entire record.

(c)

4 marks

Check 1 — Format / Input Mask Check:
1) Verifies that CarID strictly follows the pattern of 3 uppercase letters followed by 4 digits (e.g. LLL0000 such as WRE2345).
Check 2 — Length Check (or Presence/Uniqueness Check):
2) Verifies that the entered CarID string contains exactly 7 characters (not shorter or longer).

Mentor tip

Name two distinct validation checks (such as Format check LLL9999 and Length check = 7) and explain how each applies to CarID.

(d)

4 marks

Identify the query steps to retrieve the customer's surname:
1) Select the target attribute: SELECT SName
2) Specify both required tables: FROM CUSTOMER INNER JOIN RENTAL
3) Join the tables on their shared key: ON CUSTOMER.CustID = RENTAL.CustID
4) Filter by the rental dates: WHERE DateOut = '22/01/2020' AND DateRet = '26/01/2020';

Mentor tip

Ensure you include the projection (SName), both table names, the join condition on CustID, and both date conditions.

(e)

3 marks

Explain why queries are used to create database views:
1) Virtual Table Abstraction: A view encapsulates a complex multi-table join query into a reusable virtual table that staff can query simply without rewriting SQL.
2) Dynamic Updating: Because a view does not store separate physical data, it automatically reflects real-time changes made to the underlying CUSTOMER, RENTAL, and CAR tables.
3) Role-Based Security: Views can expose only specific non-sensitive columns or rows to front-desk employees while hiding restricted attributes in the base tables.

Mentor tip

Highlight that views act as virtual tables that simplify complex queries, stay dynamically updated, and restrict column/row visibility for security.

(f)

3 marks

Explain how a Data Definition Language (DDL) implements the data model:
1) DDL commands (such as CREATE TABLE) define the database schema by creating the CUSTOMER, RENTAL, and CAR tables along with their field names and data types.
2) DDL specifies integrity constraints, including PRIMARY KEY declarations (CustID, RentalNum, CarID) and FOREIGN KEY relationships.
3) DDL statements (such as ALTER TABLE or DROP TABLE) allow the DBA to modify or maintain the table structure over time.

Mentor tip

Connect DDL to defining table structures, field data types, and primary/foreign key constraints.

(g)

3 marks

Explain the importance of maintaining data consistency in the car rental database:
1) Referential and Transactional Accuracy: Consistency ensures that every record written to the database satisfies all defined constraints, cascades, and validation rules (e.g. a rental cannot reference a non-existent CarID or CustID).
2) Reliable Availability Tracking: Accurate, consistent state across tables prevents double-booking vehicles or charging customers for the wrong rental period.
3) Trustworthy Analytics and Compliance: Managers and external financial audits can rely on consistent data formats for reporting and decision-making.

Mentor tip

Link data consistency to enforcing integrity constraints, preventing booking conflicts, and ensuring reliable operational decisions.

Question 2

9 marks

Database recovery strategies, recovery methods, and restricting staff access to sensitive data

Databases

(a)

1 mark

Identify one factor a DBA considers when selecting a recovery strategy:
Recovery Time Objective (maximum tolerable downtime before service is restored), Recovery Point Objective (maximum tolerable data loss), financial budget for failover hardware, or legal/regulatory requirements.

Mentor tip

State one specific factor such as tolerable downtime (RTO), tolerable data loss (RPO), or implementation cost.

(b)

4 marks

Method 1 — Periodic Backup and Restore with Transaction Log Replay:
1) The DBA restores the most recent scheduled full or incremental backup image of the database, then replays committed transactions from the write-ahead transaction log (roll-forward) or undoes incomplete transactions (rollback) to reach a consistent point in time.
Method 2 — Database Mirroring / Instant Failover Recovery:
2) A synchronized standby replica server continuously mirrors live transactions; if the primary database fails, user traffic is immediately redirected to the mirror server with zero downtime while the primary server is repaired.

Mentor tip

Describe two distinct recovery techniques in detail, such as backup restoration with log replay and live failover mirroring.

(c)

4 marks

Method 1 — Role-Based Access Control (RBAC) and Granular Permissions:
1) Assign staff accounts to specific security roles (e.g. receptionist, manager, HR) with least-privilege permissions and restricted database views so employees can only query tables or columns required for their job duties.
Method 2 — Data Anonymization / Pseudonymization (Splitting PII):
2) Separate Personally Identifiable Information (such as names and phone numbers) into a restricted identity table and replace them in operational tables with anonymous surrogate identifiers (e.g. PER0001) so general staff cannot link sensitive records to specific individuals.

Mentor tip

Focus on access control levels/views and data anonymization (separating PII from operational tables) rather than general login passwords or encryption.

Question 3

15 marks

Artisan College student course relation 1NF analysis, 3NF normalization, and data modelling

Databases

(a)

2 marks

Outline why the STUDENT table violates First Normal Form (1NF):
1) 1NF requires every attribute cell to contain only a single atomic value.
2) In the given table, CourseNum and CourseName contain repeating groups / multiple values within a single record (for example, StudentID 2907 has three CourseNum values and three CourseName values stacked in one row).

Mentor tip

Point out both the rule (atomic values / no repeating groups) and the specific violating columns (CourseNum and CourseName).

(b)

8 marks

Construct the 3rd Normal Form (3NF) schema across four tables, clearly identifying each Primary Key [PK] and Foreign Key [FK]:
1) Isolate the transitive dependency between MajorID and Major into a MAJOR table:
MAJOR (MajorID [PK], Major)
2) Store student details in the STUDENT table referencing MajorID:
STUDENT (StudentID [PK], FName, SName, MajorID [FK])
3) Isolate course details (eliminating partial dependency on CourseNum) into a COURSE table:
COURSE (CourseNum [PK], CourseName)
4) Resolve the many-to-many enrolment relationship between students and courses using an ENROLLMENT junction table with a composite primary key:
ENROLLMENT (StudentID [PK, FK], CourseNum [PK, FK])

Mentor tip

Ensure you have 4 distinct tables (STUDENT, MAJOR, COURSE, ENROLLMENT), with correct attributes and underlined/labelled Primary Keys for each.

(c)

5 marks

Explain why data modelling is essential when designing a database:
1) Aligning with Business Purpose: Data modelling clarifies system requirements upfront so the database captures all necessary entities without omitting critical operational data.
2) Defining Entities and Attributes: It establishes the exact tables, atomic fields, and appropriate data types needed to avoid storage inefficiencies.
3) Identifying Primary and Foreign Keys: It specifies unique keys and relational links so records can be retrieved unambiguously.
4) Mapping Relationships for Complex Queries: Defining 1:M and M:N cardinalities ensures tables can be joined accurately for reporting.
5) Eliminating Redundancy via Normalization: Proper logical modelling removes duplicate data and prevents insertion, update, and deletion anomalies.

Mentor tip

Cover identifying entities, attributes, keys, relationships for multi-table queries, and avoiding redundancy/anomalies.

Question 4

20 marks

Lacto Dairy Products data warehouse, network data model, ETL data scrubbing, segmentation, and analytics

Databases

(a)

2 marks

Define a data warehouse:
1) A large, centralized repository of integrated, subject-oriented, time-variant, and non-volatile historical data consolidated from multiple disparate operational sources.
2) Designed specifically to support Online Analytical Processing (OLAP), data mining, and strategic management decision-making.

Mentor tip

Include its core characteristics (integrated, historical/time-variant, non-volatile repository) and its purpose (supporting OLAP/management decision-making).

(b)

4 marks

Characteristic 1 — Multiple Parent-Child Ownership (Graph Structure):
1) Unlike a strict hierarchical tree where each child node has only one parent, a network data model organizes records as a web-like graph where a child record (member) can have multiple parent records (owners), naturally modelling many-to-many relationships.
Characteristic 2 — Direct Pointer / Link Navigation Paths:
2) Logical associations between record types are maintained via explicit physical pointers (sets/links), providing multiple fast traversal paths directly between connected nodes.

Mentor tip

Highlight that child records can have multiple parent owners (supporting M:N relationships) and nodes are connected in a web/graph structure.

(c)

4 marks

Way 1 — Standardizing Formats and Decoding Values (Data Scrubbing):
1) During the Transform stage, ETL converts inconsistent representations across source systems into a uniform standard (such as converting 'Male'/'Female' to 'M'/'F', unifying date formats, or mapping cryptic numeric codes into meaningful labels like 'Wholesaler' and 'Retailer').
Way 2 — Deduplication and Handling Missing/Outlier Values:
2) ETL identifies and removes duplicate records merged from different sources, flags out-of-bounds numerical errors, and populates missing/null attributes with sensible default or mean values.

Mentor tip

Describe two concrete data-cleaning actions: standardizing inconsistent formats/codes and removing duplicates or fixing missing/erroneous values.

(d)

5 marks

Advantages of Data Segmentation for Lacto Dairy Products:
1) Targeted Pricing and Promotions: Partitioning data into distinct customer segments (such as wholesalers vs retailers, or by geographic region) allows Lacto to tailor bulk pricing tiers, product packaging, and marketing campaigns to each group's purchasing behaviour.
2) Efficient Resource Allocation and Retention: Helps management identify high-value segments, anticipate specific demand patterns, and improve customer retention.
Disadvantages of Data Segmentation:
3) Over-Generalization Within Segments: Assumes all buyers within a segment (e.g. all retailers in a region) share identical needs, potentially overlooking niche opportunities or innovative product lines.
4) Increased Management Cost and Complexity: Collecting, maintaining, and analysing multiple segmented datasets increases administrative overhead and data management costs (and over-segmenting can fragment strategic focus).
Conclusion:
5) Data segmentation is highly effective for differentiating broad pricing structures between wholesalers and retailers, provided the company periodically reviews sub-segments so it does not miss emerging market trends.

Mentor tip

Provide two clear advantages (with reference to wholesalers/retailers), two disadvantages (such as over-generalization and higher cost/complexity), and a balanced conclusion.

(e)

5 marks

Explain how data analytics assists Lacto Dairy Products managers in future decision-making:
1) Predictive Forecasting: Analytics and machine learning models examine historical sales time series to forecast future seasonal demand for specific dairy products.
2) Scenario and Pricing Analysis: Managers can evaluate how discount structures or price changes impact purchasing volume across wholesalers and retailers.
3) Product Portfolio Optimization: Association rule mining and profitability analysis reveal which product lines should be expanded, bundled, or discontinued.
4) Processing Massive Data Volumes: Automated data mining extracts actionable patterns and correlations from vast warehouse datasets that would be impossible to spot manually.
5) Evidence-Based Strategy: Replaces guesswork with quantitative business intelligence for supply chain and inventory planning.

Mentor tip

Connect data mining/forecasting to concrete dairy management decisions (predicting seasonal demand, testing discount impacts, pruning unprofitable product lines).

Option B: Modelling and simulation · Option

Questions 5–8 · 65 marks

Question 5

21 marks

Restaurant monthly electricity bill computer model, variables, and min/max usage pseudocode

Modelling and simulation

(a)

1 mark

Define a computer model:
A mathematical and algorithmic representation of a real-world system or process implemented as a computer program to calculate outcomes or test scenarios.

Mentor tip

Define it as a programmatic/mathematical representation of a real-life situation.

(b)

3 marks

State three variables and their data types for the electricity model:
1) UNITS_USED (or DAYS_IN_MONTH): Integer
2) MONTH_NAME: String
3) TOTAL_BILL (or DAILY_CHARGE / UNIT_COST / TAX_RATE): Real / Float / Double

Mentor tip

Always pair each variable name with its explicit data type (Integer, String, Real/Float).

(c)

5 marks

Construct pseudocode to input monthly units and days, then calculate and output the total bill (including 10% tax) and daily average usage:
```
UNIT_COST = 0.20
DAILY_CHARGE = 0.75
TAX_RATE = 0.10
output "Enter electricity units used in the month: "
input UNITS_MONTH
output "Enter number of days in the month: "
input DAYS_MONTH
SUBTOTAL = (UNITS_MONTH * UNIT_COST) + (DAYS_MONTH * DAILY_CHARGE)
TOTAL_BILL = SUBTOTAL + (SUBTOTAL * TAX_RATE)
UNITS_PER_DAY = UNITS_MONTH / DAYS_MONTH
output "Total monthly electricity bill: $", TOTAL_BILL
output "Average electricity units used per day: ", UNITS_PER_DAY
```

Mentor tip

Include prompts/messages with your inputs and outputs, remember to add the 10% tax to the bill, and divide units by days for daily usage.

(d)

2 marks

Outline one limitation of this model for helping the owners reduce costs:
1) The model only tracks aggregate monthly/daily electricity consumption across the entire building.
2) It provides no breakdown by individual appliance or subsystem (cooking vs refrigeration vs air conditioning), so the owners cannot identify which specific equipment is wasting energy.

Mentor tip

Point out that aggregate monthly totals do not isolate which specific appliances or times of day consume the most electricity.

(e)

2 marks

State two ways the model could be implemented:
1) Spreadsheet software (using built-in formulas and functions like MIN/MAX/INDEX/MATCH).
2) Custom program code written in a high-level programming language (using arrays and search/sort algorithms).

Mentor tip

State 'Spreadsheet' and 'Program code / High-level programming language'.

(f)

8 marks

Construct pseudocode using parallel arrays to identify and output the month names with the highest and lowest electricity usage:
```
// Parallel arrays storing the 12 month names and their corresponding unit usage
MONTHS = ["January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December"]
USAGE = [5800, 5000, 5200, 4800, 4750, 4900, 5500, 6500, 5600, 5960, 5850, 6280]
MAX_USAGE = USAGE[0]
MIN_USAGE = USAGE[0]
HIGH_MONTH = MONTHS[0]
LOW_MONTH = MONTHS[0]
loop I from 1 to 11
if USAGE[I] > MAX_USAGE then
MAX_USAGE = USAGE[I]
HIGH_MONTH = MONTHS[I]
end if
if USAGE[I] < MIN_USAGE then
MIN_USAGE = USAGE[I]
LOW_MONTH = MONTHS[I]
end if
end loop
output "Month with highest electricity usage: ", HIGH_MONTH
output "Month with lowest electricity usage: ", LOW_MONTH
```

Mentor tip

Use two 1D parallel arrays (MONTHS and USAGE), initialize min/max trackers, loop across all 12 months updating both the extreme value and the corresponding month name, and output the month names.

Question 6

15 marks

Government budget resource allocation simulation, data collection accuracy, and citizen privacy ethics

Modelling and simulation

(a)

2 marks

Describe the difference between a model and a simulation in this government budgeting scenario:
1) The model is the static set of mathematical equations and rules defining the relationships between population size, tax revenue, and healthcare/education spending.
2) The simulation is the dynamic execution of that model over time using varied demographic and tax rate inputs to observe how future national budgets behave.

Mentor tip

Contrast the underlying mathematical rules/relationships (model) with running those rules over time with changing tax/population inputs (simulation).

(b)(i)

2 marks

Outline one advantage of using a simulation for future government budget allocation:
1) Enables policymakers to test "what-if" tax and demographic scenarios safely in advance, identifying potential funding shortfalls in healthcare or education years before they occur.
2) Allows proactive policy adjustments to prevent budget deficits without risking real public services.

Mentor tip

Connect testing future scenarios in advance to preventing healthcare/education budget shortages.

(b)(ii)

2 marks

Outline one disadvantage of relying on a simulation for government resource allocation:
1) If the underlying mathematical assumptions are flawed or the input data (collected only every 3 years) is outdated or incomplete, the simulation will produce inaccurate projections.
2) Relying on faulty predictions can lead to misallocated tax funds and underfunded public hospitals or schools.

Mentor tip

Explain how flawed assumptions or outdated inputs lead directly to misallocated public budgets.

(c)

4 marks

Improvement 1 — Increase Data Collection Frequency:
1) Collect demographic and economic data annually or continuously in real time rather than only once every three years, allowing the simulation to capture rapid economic shifts or sudden population changes accurately.
Improvement 2 — Expand the Scope and Granularity of Variables:
2) Gather additional relevant variables (such as age-group distribution, inflation rates, employment levels, and regional disease trends) to reduce the impact of unmodelled confounding factors.

Mentor tip

Propose two clear improvements: collecting data more frequently than every 3 years, and capturing a broader set of demographic/economic variables.

(d)

5 marks

Reasons Why Citizen Ethical Concerns ARE Justified:
1) Invasion of Privacy and Function Creep: Detailed personal data on income, health, and family size might be repurposed by government agencies for surveillance or shared with unauthorized third parties beyond budget modelling.
2) Data Breach Vulnerability: Centralizing sensitive citizen records creates a high-value target for cyberattacks; a breach could expose confidential medical and tax histories.
Reasons Why Government Data Collection IS Justified:
3) Essential for Accurate Public Planning: High-resolution data on earnings, demographics, and healthcare usage is vital to build an accurate simulation that ensures hospitals and schools receive fair, adequate funding.
4) Anonymization Mitigates Harm: If data is properly aggregated and stripped of personally identifiable markers before entering the simulation, individual privacy is preserved.
Balanced Appraisal:
5) Citizen concerns are valid regarding raw personal records, so the government must enforce strict anonymization, encryption, and legal data-protection boundaries to balance privacy rights against effective public resource planning.

Mentor tip

Discuss both sides (privacy/surveillance/breach risks vs the public benefit of accurate healthcare and education funding) and conclude with how anonymization balances the trade-off.

Question 7

9 marks

2D vs 3D visualization examples, keyframe animation, and hardware resource implications

Modelling and simulation

(a)(i)

1 mark

State one example of a 2D visualization:
A flat 2D map, line graph, bar chart, blueprint diagram, or 2D sprite image.

Mentor tip

Give any clear flat two-dimensional graphical representation (such as a 2D map or chart).

(a)(ii)

1 mark

State one example of a 3D visualization:
A 3D wireframe/polygon model, 3D CGI architectural walkthrough, or 3D virtual reality environment.

Mentor tip

Give any three-dimensional graphical representation with depth/perspective.

(b)

2 marks

Describe the significance of a key frame in 3D visualization/animation:
1) A key frame defines the exact position, pose, and attributes of 3D objects at a critical starting or ending point of a movement sequence.
2) Software interpolation (tweening) automatically calculates and generates all the intermediate frames between two key frames to create a smooth transition.

Mentor tip

State that keyframes define the start and end states of a movement so the software can interpolate the smooth transition frames between them.

(c)

5 marks

Explain the implications of 3D animation on computer hardware resources during game creation:
1) Intensive GPU/CPU Processing: Rendering 3D polygon meshes, calculating lighting/shadows (ray tracing or rasterization), and interpolating character movement across scenes demand massive parallel floating-point computation from the GPU and CPU.
2) Compositing Multiple Layers: Combining background environments, moving character models, particle effects, and textures into each final frame increases rendering latency.
3) High RAM / VRAM Capacity Requirements: Storing high-resolution texture maps, 3D geometry buffers, and multiple animation frames simultaneously requires substantial primary memory and dedicated video memory (VRAM).
4) Secondary Storage Bandwidth: Loading large 3D scene assets and saving rendered animation sequences requires fast SSD storage and high memory bus bandwidth to avoid bottlenecks.
5) Extended Rendering Times: High-detail scenes require long render farms or compilation times if hardware resources are insufficient.

Mentor tip

Address CPU/GPU processing load (rendering polygons, lighting, and layer transitions), RAM/VRAM capacity for textures/frames, and storage speed.

Question 8

20 marks

Autonomous vehicle genetic algorithms, supervised vs unsupervised voice commands, NLP, and human vs machine language learning

Modelling and simulation

(a)

2 marks

Outline how a genetic algorithm is iterative:
1) Starting from an initial population of potential solutions, the algorithm repeatedly loops through successive generations.
2) In each iteration (generation), individuals are evaluated, selected, crossed over, and mutated to progressively refine the population toward an optimal driving behaviour.

Mentor tip

Explain that the algorithm repeats a cycle of evaluation, selection, crossover, and mutation across multiple generations.

(b)

2 marks

Outline the role of the fitness function in a genetic algorithm:
1) A quantitative objective function that evaluates each potential solution in a generation and assigns a numerical score measuring how well it achieves the target criteria (e.g. safe, smooth navigation).
2) This score determines which high-performing individuals are selected to reproduce into the next generation.

Mentor tip

State that it quantitatively scores how close each potential solution is to the optimum to guide selection.

(c)(i)

2 marks

Outline one benefit of supervised learning for learning verbal driving commands:
1) Uses a labelled dataset mapping known voice commands to exact, verified vehicle actions.
2) Ensures the autonomous vehicle's response to critical safety commands (like "stop" or "pull over") is highly predictable, reliable, and verifiable.

Mentor tip

Link labelled input-output training pairs to predictable, safety-verified vehicle responses.

(c)(ii)

2 marks

Outline one benefit of unsupervised learning for learning verbal driving commands:
1) Can discover patterns in unlabelled natural speech without requiring every possible phrasing to be pre-labelled in advance.
2) Allows the vehicle to adapt to novel accents, regional dialects, informal phrasing, or emotional vocal tone (such as an urgent shout warning of a hazard).

Mentor tip

Link unlabelled pattern discovery to handling unforeseen phrasing, new dialects, or urgent vocal tones.

(d)

5 marks

Explain how advances in Natural Language Processing (NLP) enable appropriate responses to verbal commands:
1) Acoustic and Phonemic Recognition: Deep neural speech-to-text models accurately transcribe spoken words even over background road and engine noise.
2) Semantic Parsing and Intent Classification: Modern NLP extracts the underlying meaning and intent of a sentence rather than relying on rigid keyword matching (e.g. understanding "it's too chilly in here" as a command to raise the cabin temperature).
3) Contextual Disambiguation: Transformer/attention models use conversational and environmental context to resolve ambiguous words or pronouns.
4) Sentiment and Urgency Detection: Prosody and tone analysis detect panic or urgency in the passenger's voice (e.g. shouting "Watch out!"), triggering immediate emergency braking.
5) Low-Latency Edge Processing: Optimized NLP models execute in real time on the vehicle's onboard processors so safety-critical actions occur without network delay.

Mentor tip

Cover semantic intent extraction, noise/accent robustness, urgency/tone detection, and real-time execution for safety.

(e)

5 marks

Explain why human language acquisition differs from machine language learning:
1) Human Learning — Meaning, Social Context, and Grammar Rules:
- Humans first acquire language through immersive social interaction, grounding words in real-world sensory experiences and physical context.
- Humans internalize grammatical structures and generative syntax rules, enabling them to understand metaphors, sarcasm, and novel expressions with very few examples.
2) Machine Learning — Statistical Pattern Matching on Massive Corpora:
- Machine language models do not possess conscious comprehension or lived sensory experience; instead, they compute statistical probability distributions over vast datasets of text and audio tokens.
- Machines match input sequences to stored mathematical patterns and weights, which can struggle when human speech violates formal syntax or uses novel cultural idioms.

Mentor tip

Contrast human learning (experiential meaning, social context, and grammatical rule internalization) with machine learning (statistical probability matching across massive datasets).

(f)

2 marks

Describe the use of a chatbot:
1) A conversational software agent designed to simulate natural human dialogue via text or speech interfaces.
2) Commonly deployed in customer service, virtual assistants, or entertainment to automatically answer routine queries, guide users, or execute tasks 24/7.

Mentor tip

State both what a chatbot does (simulates human conversation) and a practical application (answering customer queries or virtual assistance).

Option C: Web science · Option

Questions 9–12 · 65 marks

Question 9

10 marks

URLs, DNS resolution, HTTPS characteristics, protocols vs standards, and connectionless IP

Web science

(a)

1 mark

Define a Uniform Resource Locator (URL):
The unique human-readable address used to specify the location of a resource (such as a web page or file) on the World Wide Web / Internet and the protocol used to retrieve it.

Mentor tip

State clearly that a URL is the unique address of a resource/webpage on the internet.

(b)

3 marks

Describe how a Domain Name Server (DNS) functions:
1) When a user enters a URL in a browser, the DNS receives the human-readable domain hostname and looks it up in its database to resolve it to the corresponding numeric IP address.
2) If the local/primary DNS server does not have the IP cached, it forwards the query hierarchically to root, Top-Level Domain (TLD), and authoritative DNS servers until a match is found.
3) Once resolved, the DNS returns the IP address to the client browser so TCP/IP packets can be routed to the target web server (or returns an error if the domain does not exist).

Mentor tip

Cover translating the domain name to an IP address, querying higher-level DNS servers if not cached locally, and returning the IP to the browser.

(c)

2 marks

Identify two characteristics of HTTPS:
1) Encrypts all HTTP communication traffic between client and server using TLS/SSL (symmetric session encryption negotiated via asymmetric keys).
2) Authenticates the identity of the web server using a verifiable digital SSL/TLS certificate issued by a trusted Certificate Authority.

Mentor tip

Mention TLS/SSL encryption of data in transit and server authentication via digital certificates.

(d)

2 marks

Distinguish between a network protocol and a technical standard:
1) A protocol is a formal set of procedural rules governing how data packets are formatted, transmitted, received, and acknowledged during communication across a network (e.g. HTTP, TCP, IP).
2) A standard is a broader, agreed-upon technical specification for hardware, software, data formats, or languages (e.g. HTML5, Unicode, IEEE 802.11) that ensures universal compatibility and interoperability across platforms.

Mentor tip

Contrast rules for transmitting/exchanging data over a network (protocol) with broader specifications for formats, languages, or hardware interoperability (standard).

(e)

2 marks

Outline what connectionless means in the context of IP packet transmission:
1) Data packets (datagrams) are transmitted onto the network without any prior handshake or dedicated end-to-end circuit being established between the client and server.
2) Each packet carries its own complete destination IP address in its header and is routed independently across network hops (potentially arriving out of order).

Mentor tip

Highlight that there is no prior handshake/setup circuit and each packet is addressed and routed independently.

Question 10

19 marks

Server-side PHP/SQL database insertion, CGI script execution, SEO, link ranking, and crawling the deep web

Web science

(a)(i)

3 marks

Describe the process executed by the PHP/SQL code fragment:
1) Constructs an SQL INSERT statement to add a new record ('Sonia', 'Smith', 'sonia@smith.com', '+61231456789') into the FirstName, Surname, Email, and Cellphone columns of GuestsTable.
2) Executes the query on the database connection ( $co nn - > q u er y ($ sql)) and checks whether the execution succeeded (=== TRUE).
3) Outputs "New record created successfully" to the web page if the insert succeeded, or outputs an error message containing the SQL string and database connection error details if it failed.

Mentor tip

Mention inserting the four values into GuestsTable, executing the query via $conn, and echoing either the success message or the error details.

(a)(ii)

3 marks

Describe how the Common Gateway Interface (CGI) enables server-side script execution:
1) When a web browser sends an HTTP request for a dynamic resource, the web server uses the CGI specification to pass the client's request data (such as form inputs or query strings) to an external script/executable on the server.
2) The server-side CGI script executes at runtime (e.g. interacting with a database).
3) The script outputs dynamically generated HTML back through the web server, which sends the rendered response to the client's browser.

Mentor tip

Trace the three steps: client request passed by web server to CGI script, script execution on the server, and dynamic HTML returned to the client.

(b)(i)

2 marks

Identify two Search Engine Optimization (SEO) techniques:
1) Including relevant, descriptive keywords in HTML <title>, heading tags, image alt attributes, and <meta> description/keyword tags.
2) Building high-quality incoming and outgoing hyperlinks (backlinks from authoritative websites), maintaining clean mobile-responsive HTML, and updating site content regularly.

Mentor tip

Name two specific technical practices (such as relevant meta-tags/alt attributes and authoritative backlinks/regular content updates) rather than generic labels like 'white hat'.

(b)(ii)

4 marks

Explain why incoming and outgoing links are significant factors in search engine ranking:
1) Core Ranking Input: Search algorithms such as PageRank and HITS rely heavily on the hyperlink graph structure to evaluate page importance.
2) Incoming Links as Votes of Authority: In PageRank, an incoming link (backlink) from another website acts as a vote of confidence; links coming from high-authority pages pass greater weight to the target page.
3) Outgoing Links and Hub Value: In HITS, a page that links out to many reputable, relevant resources achieves a high Hub score, which in turn reinforces the Authority scores of the pages it cites.
4) Vulnerability to Manipulation: Because link counts strongly boost rankings, some site operators attempt black-hat manipulation such as link farms or comment spam, forcing search engines to evaluate link quality alongside quantity.

Mentor tip

Reference both PageRank (incoming links as weighted votes of authority) and HITS (hubs and authorities).

(c)

3 marks

Describe how a web crawler indexes a webpage:
1) Starting from a list of known seed URLs, the crawler visits the webpage (checking robots.txt permissions and sitemaps first).
2) It parses the HTML document, extracts body text, headings, and <meta> tags, and stores these keywords alongside the URL in the search engine's inverted index database.
3) It extracts all outgoing hyperlinks (<a href="...">) on the page and adds those newly discovered URLs to its crawl queue to visit next (using breadth-first or depth-first traversal).

Mentor tip

Cover starting from seed URLs/checking robots.txt, parsing and storing keywords/metadata in the index, and queuing outgoing hyperlinks for further crawling.

(d)

4 marks

Suggest why the inability to access the deep web is an increasing problem for search engines:
1) Shift from Static to Dynamic Database-Driven Sites: A rapidly growing proportion of modern websites generate content dynamically on the fly from backend databases only in response to specific user queries or form submissions, leaving no static URLs for crawlers to follow.
2) Proliferation of Paywalls and Authenticated Portals: More high-value information (academic research, social media feeds, subscription news, cloud repositories) is locked behind login screens or paywalls.
3) Exponential Growth of Unindexed Data: As deep web databases expand much faster than the static surface web, traditional crawlers miss the vast majority of online information.
4) Incomplete and Less Relevant Search Results: Users relying on surface-web search engines receive an increasingly narrow, superficial subset of available human knowledge, reducing search comprehensiveness and reliability.

Mentor tip

Focus consistently on either the technical barriers (proliferation of dynamic database-driven pages, logins, paywalls, and exponential growth) or the consequential impact on search completeness.

Question 11

16 marks

Peer-to-peer (P2P) networks, lossy compression for audio vs eBooks, cloud computing, and IP protection

Web science

(a)

2 marks

Identify two characteristics of a Peer-to-Peer (P2P) network:
1) Decentralized Architecture: There is no single central server storing the files; nodes communicate directly with one another, eliminating a single point of failure.
2) Dual Client-Server Role: Every connected peer simultaneously acts as both a client (downloading data chunks from other peers) and a server (uploading data chunks and sharing bandwidth).

Mentor tip

State that P2P is decentralized without a central server and that each peer acts as both client and server (supplying and consuming resources).

(b)

6 marks

General Effect of Lossy Compression:
1) Permanently discards data deemed less perceptible to achieve a much smaller file size, resulting in faster download times over mobile networks and lower storage requirements.
Appropriateness for the Music Album (Audio):
2) Highly Appropriate: Psychoacoustic lossy formats (like MP3 or AAC) remove frequencies outside the range of human hearing or masked by louder sounds.
3) At reasonable bitrates, the file size is drastically reduced (~90%) with negligible perceptible loss in audio quality for mobile listeners (though extreme compression causes audible distortion).
Appropriateness for the eBook (Text and Embedded Images):
4) Inappropriate for Text: Applying lossy compression to character text is completely unsuitable because discarding bytes corrupts characters and renders the book unreadable (text requires lossless compression).
5) Mixed for Embedded Cover/Photo Images: Lossy compression (like JPEG) can shrink photographic illustrations inside the eBook, though excessive compression introduces blocky pixelation, colour banding, and blurry text overlays.
Overall Judgment:
6) Lossy compression is well-suited for streaming/downloading the audio album on mobile devices, but the eBook should use lossless compression for its text and moderate compression only for photographic images.

Mentor tip

Evaluate both media types separately: explain why lossy works well for audio (removing inaudible frequencies) but corrupts eBook text or pixelates heavily compressed images.

(c)

2 marks

Identify two characteristics of cloud computing:
1) On-Demand Remote Hosting: Storage, processing, and software applications (SaaS/PaaS/IaaS) are hosted on remote internet servers and accessible from any location or device.
2) Elastic Scalability and Pay-As-You-Go: Computing capacity and storage scale dynamically to match user demand without the user managing physical hardware.

Mentor tip

Mention internet-hosted location-independent services (SaaS/IaaS) and elastic scalability.

(d)

6 marks

Evaluation of Method 1 — Protecting Images Behind a Registered User Login:
1) Strengths: Enforced on the server side, preventing anonymous public visitors and automated web scrapers from accessing the portfolio; allows the artist to authenticate users and track who views the images.
2) Weaknesses: Creates friction that deters casual fans or potential buyers unwilling to register; blocks search engine crawlers from indexing the artwork (reducing publicity); and logged-in users can still screenshot or redistribute downloaded files.
Evaluation of Method 2 — Client-Side Script Disabling the Right-Click 'Save Image' Menu:
3) Strengths: Allows the public and search engines to browse and discover the gallery freely without login friction, while deterring casual, non-technical visitors from saving copies with a simple right-click.
4) Weaknesses: Easily bypassed because the script runs on the client side—users can disable JavaScript in their browser, extract the image directly from the browser cache/developer tools, or take a screen capture.
Overall Conclusion:
5) Neither method guarantees absolute protection once an image is displayed on a client screen; combining public low-resolution watermarked previews with server-side authenticated login for high-resolution files provides a more balanced solution.

Mentor tip

Question 12

20 marks

Bowtie web graph model, SCC characteristics, directed sub-graphs, IN/OUT evolution, and ontologies vs folksonomies

Web science

(a)(i)

1 mark

Define a node in the bowtie web graph:
A single web page (or website) on the World Wide Web.

Mentor tip

State clearly that a node represents a webpage or website.

(a)(ii)

1 mark

Define an edge in the bowtie web graph:
A directed hyperlink pointing from one web page (node) to another.

Mentor tip

State clearly that an edge represents a directed hyperlink between two pages.

(b)

2 marks

Identify two characteristics of a node in the Strongly Connected Core (SCC):
1) Mutual Reachability: Any node in the SCC can reach any other node in the SCC by following a directed path of hyperlinks (and therefore has both incoming and outgoing links).
2) Component Relationships: An SCC node can be reached from IN nodes and can reach OUT nodes via directed paths, but cannot navigate back to IN nodes or return from OUT nodes.

Mentor tip

Mention that SCC nodes have both incoming and outgoing links such that every SCC node can reach every other SCC node (and can reach OUT nodes but not IN nodes).

(c)

4 marks

Construct the directed sub-graph with circles for nodes {A..I} and directed arrows for edges:
1) IN Component (Nodes A, B, C): Draw circles A, B, and C with directed arrows pointing into the SCC (e.g. A -> D, B -> E, C -> D) and no arrows returning from the SCC to A, B, or C.
2) SCC Component (Nodes D, E, F, G): Draw circles D, E, F, and G connected in a strongly connected directed cycle (e.g. D -> E -> F -> G -> D) so every node in {D, E, F, G} can reach every other node in the core.
3) OUT Component (Nodes H, I): Draw circles H and I with directed arrows pointing from the SCC into them (e.g. F -> H, G -> I) and no arrows pointing back into the SCC.
4) Tube Connection (C to I): Draw a direct arrow from IN node C to OUT node I (C -> I) bypassing the SCC.

Mentor tip

Ensure D, E, F, G form a closed directed loop (SCC), A, B, C point into {D,E,F,G}, {D,E,F,G} point out to H and I, and include the direct tube arrow C -> I.

(d)

4 marks

Suggest why the IN and OUT proportions changed from 21% each in 2000 to 32% IN and 6% OUT in 2012:
1) Shrinkage of OUT (21% to 6%) due to Ubiquitous Outbound Linking: Modern websites rarely act as dead-end terminal pages; social media buttons, advertising networks, partner links, and SEO best practices cause previously isolated OUT sites to link back to major platforms in the SCC, pulling them into the SCC.
2) Growth of IN (21% to 32%) due to Easy Self-Publishing: Blogs, personal sites, and new small-business pages can be created effortlessly and immediately link out to authoritative SCC sites (like Wikipedia, YouTube, or news portals), but receive few or no incoming links from established SCC sites.
3) Expansion and Density of the SCC (27% to 51%): Social networking, collaborative wikis, and content-sharing platforms have made the core of the web much more densely interconnected.
4) Crawl Scale and Methodology Differences: The 2012 crawl covered 3.5 billion nodes compared to ~200 million in 2000, capturing a much larger long-tail of newly published leaf pages (IN).

Mentor tip

(e)

2 marks

Identify two characteristics of collective intelligence:
1) Shared Group Knowledge: Emerges from the decentralized collaboration, crowdsourcing, and combined contributions of large communities of users over the web.
2) Continuous Co-Creation: Participants actively generate, edit, review, and refine shared resources (such as Wikipedia articles or open-source repositories on GitHub).

Mentor tip

State that it is shared knowledge derived from mass collaboration/crowdsourcing and give an example or characteristic of co-creation.

(f)

6 marks

Significance and Strengths of Ontologies in the Semantic Web:
1) Formal Structure: An ontology formally defines a domain's classes, properties, and relationships using standardized machine-readable frameworks (such as RDF, OWL, and Dublin Core).
2) Automated Reasoning and Interoperability: Because ontological vocabulary is unambiguous, consistent, and hierarchical, software agents can automatically query, integrate, and infer new knowledge across heterogeneous databases.
3) Limitation of Ontologies: Designing and maintaining strict ontologies requires specialized expertise, high cost, and rigid agreement, making them slow to adapt to new topics.
Significance and Strengths of Folksonomies:
4) User-Driven Collaborative Tagging: A folksonomy arises when everyday users freely annotate web content (photos, videos, posts) with informal, uncontrolled keywords or hashtags.
5) Low Barrier and Adaptability: Folksonomies are inexpensive, democratic, and rapidly reflect emerging trends and user vocabulary without requiring central coordination.
6) Limitation of Folksonomies for the Semantic Web: Because tags are uncontrolled, they suffer from spelling variations, synonyms, homonyms, and ambiguity, making automated machine reasoning and consistent semantic indexing difficult.

Mentor tip

Discuss both ontologies (formal RDF/OWL structure enabling machine reasoning, but rigid/complex) and folksonomies (flexible user tagging, but ambiguous/inconsistent for machines).

Option D: Object-oriented programming · Option

Questions 13–16 · 65 marks

Question 13

14 marks

Real estate UML class diagram, primitive types, relationships, static counters, and object arrays

Object-oriented programming

(a)

1 mark

Define a primitive data type:
A fundamental, predefined basic data type built directly into the programming language (such as int, boolean, double, char) that stores a single raw value directly in memory rather than an object reference.

Mentor tip

State that it is a basic, built-in language type that holds a raw value directly in memory (and serves as a building block for composite object types).

(b)(i)

1 mark

State a suitable Boolean attribute for Property:
isSold (or hasGarden, hasParking, hasPool, isAvailable).

Mentor tip

Name any realistic true/false property attribute such as isSold or hasGarage.

(b)(ii)

1 mark

State a suitable integer attribute for Property:
numBedrooms (or numRooms, numBathrooms, floorArea, yearBuilt).

Mentor tip

Name any realistic whole-number property attribute such as numBedrooms or area.

(c)(i)

1 mark

State the relationship between Owner and Property:
Aggregation (or Association / "has-a" relationship).

Mentor tip

Look at the open diamond on the UML link between Owner and Property, which denotes aggregation.

(c)(ii)

1 mark

State the relationship between House and Property:
Inheritance ("is-a" relationship — House is a subclass of the superclass Property).

Mentor tip

Look at the hollow triangle arrow pointing from House up to Property, which denotes inheritance.

(d)

2 marks

Distinguish between a class and an instantiation in this real estate scenario:
1) A class is the abstract blueprint or template that defines the attributes and methods common to a category (for example, the House or Property class definition).
2) An instantiation is a concrete object created in memory at runtime from that class blueprint containing actual attribute values (for example, a specific House instance at "12 Oak Street" priced at $350,000).

Mentor tip

Provide both the theoretical distinction (blueprint vs concrete object in memory) and a scenario-specific example (the House class vs a specific house object with an address/price).

(e)

2 marks

Outline how the static modifier affects a variable:
1) The variable belongs to the class itself rather than to individual object instances, so a single shared copy exists across all instances of that class.
2) It can be accessed or updated directly via the class name without needing to instantiate an object, and any change to its value is seen by all instances.

Mentor tip

Highlight that a static variable is a single class-level variable shared by all instances.

(f)

2 marks

Describe how static variables can track the total number of House and Apartment objects created:
1) Declare a static integer counter variable (e.g. private static int totalCount = 0;) in the Property superclass (or separate static counters in both House and Apartment).
2) Increment this static counter by 1 inside the constructor every time a new House or Apartment object is instantiated, and provide a public static accessor method to retrieve the combined count.

Mentor tip

Explain declaring a static integer counter and incrementing it inside the constructor(s) whenever a House or Apartment is instantiated.

(g)

3 marks

Construct the Java statement to declare and instantiate the wishList array:
```java
House[] wishList = new House[10];
```
1) Declares wishList with type House[].
2) Uses the new keyword.
3) Allocates an array of size 10: House[10].

Mentor tip

Include the exact array identifier wishList, the new operator, and the size 10 of type House.

Question 14

10 marks

Encapsulation, inheritance for student housing, and programming language internationalization

Object-oriented programming

(a)

1 mark

Define encapsulation:
Bundling data (attributes) and the methods that operate on that data into a single unit (class), while restricting direct external access to internal state using private visibility modifiers.

Mentor tip

Mention both bundling data with methods inside a class and hiding internal data via private attributes.

(b)

3 marks

Explain one benefit provided by encapsulation (Data Hiding / Protection):
1) By declaring class attributes as private, external classes cannot directly read or accidentally corrupt an object's internal state.
2) External code must interact with the attributes strictly through public accessor (getter) and mutator (setter) methods.
3) Mutator methods can enforce validation rules (e.g. ensuring a property price is never set to a negative number), improving program reliability and modularity.

Mentor tip

Develop a single cohesive chain of reasoning: private variables -> controlled access via getters/setters -> preventing accidental data corruption (or easier isolated maintenance).

(c)

3 marks

Explain one benefit of using inheritance to create the new Room class:
1) Code Reuse: By extending House (or Property), the Room subclass automatically inherits existing attributes (ID, address, city, price) and their accessor/mutator methods.
2) Faster, Less Error-Prone Development: Programmers do not need to rewrite or retest duplicated code, saving development time and only needing to add attributes unique to a student room.
3) Easier Maintenance: Any future bug fix or update made to the superclass is automatically propagated to Room and all other subclasses.

Mentor tip

Focus on inheriting existing attributes/methods from Property/House, which eliminates code duplication and saves development/testing time.

(d)

3 marks

Explain one feature of modern programming languages that supports internationalization:
1) Native Unicode Support: Modern languages (like Java) represent characters using Unicode (such as 16-bit UTF-16 or UTF-8) rather than legacy 7/8-bit ASCII.
2) Expanded Character Set: Unicode can encode tens of thousands of characters and symbols (2^16 = 65,536 in 16-bit, or over 1 million code points).
3) Global Alphabet Compatibility: This allows student names, addresses, and user interfaces to accurately store and display non-Latin scripts (such as Chinese, Arabic, Cyrillic, and accented characters) for international users.

Mentor tip

Identify Unicode (16+ bits vs 8-bit ASCII), state its large character capacity, and explain that it represents diverse international alphabets and symbols.

Question 15

21 marks

Method signatures, String immutability, null pointer handling, sorting House objects, and top-3 budget selection

Object-oriented programming

(a)

2 marks

Define a method signature:
1) The combination of the method's identifier (name) and its parameter list (the number, data types, and order of its parameters).
2) Together with the return type / access modifier in the method header, it uniquely identifies the method for invocation and overloading.

Mentor tip

Include both the method name/identifier and the parameter list (types and order).

(b)

2 marks

Describe how an original String variable passed to a method can be assigned a new value:
1) In Java, arguments are passed by value and String objects are immutable, so reassigning the parameter inside a void method will not modify the caller's original variable.
2) Instead, the method must declare a return type of String, return the newly constructed String value, and the caller must assign the method's return value back to the original String variable (e.g. myStr = modifyString(myStr);).

Mentor tip

Explain that Java uses pass-by-value / Strings are immutable, so the method must return a String that is assigned back to the original variable.

(c)

1 mark

State the intended purpose of unknown(String x):
To output/print the addresses of all houses in the allHouses array that are located in the specified city x.

Mentor tip

Make sure to state 'all houses' matching city x, not just a single house.

(d)(i)

2 marks

Outline the runtime error likely to occur when unknown(String x) is called:
1) Because allHouses is only partially filled, the loop iterating up to allHouses.length will eventually encounter unpopulated array elements that are null.
2) Calling allHouses[i].getCity() on a null element throws a NullPointerException at runtime.

Mentor tip

Link the partially filled array (null elements) to throwing a NullPointerException when invoking .getCity().

(d)(ii)

2 marks

Outline how this runtime error can be corrected:
1) Add a null check before invoking .getCity(): if (allHouses[i] != null && allHouses[i].getCity().equals(x)).
2) Alternatively, maintain an integer counter count of the actual number of House objects stored and loop only while i < count (or use a dynamic ArrayList<House>).

Mentor tip

Propose either guarding the condition with if (allHouses[i] != null) or stopping the loop at a count variable tracking populated elements.

(e)

5 marks

Construct the houseSort() method to sort allHouses in ascending order of price (using Bubble Sort or Selection Sort):
```java
public void houseSort() {
House temp;
for (int i = 0; i < allHouses.length - 1; i++) {
for (int j = 0; j < allHouses.length - 1 - i; j++) {
if (allHouses[j] != null && allHouses[j + 1] != null) {
if (allHouses[j].getPrice() > allHouses[j + 1].getPrice()) {
temp = allHouses[j];
allHouses[j] = allHouses[j + 1];
allHouses[j + 1] = temp;
}
}
}
}
}
```
1) Declares a temporary swap variable of type House.
2) Uses nested outer and inner loops to traverse the array.
3) Compares adjacent prices using allHouses[j].getPrice() > allHouses[j + 1].getPrice().
4) Swaps the complete House object references (not just the integer prices).

Mentor tip

Be sure to declare your swap variable as type House and swap the House objects in allHouses[j] and allHouses[j+1] based on .getPrice().

(f)

7 marks

Construct the selectThree(int budget) method returning the 3 most expensive houses <= budget in ascending order:
```java
public House[] selectThree(int budget) {
House[] result = new House[3];
houseSort();
int i = 0;
while (i < allHouses.length && allHouses[i] != null && allHouses[i].getPrice() <= budget) {
i++;
}
// i-1 is now the index of the most expensive house within budget
result[0] = allHouses[i - 3];
result[1] = allHouses[i - 2];
result[2] = allHouses[i - 1];
return result;
}
```
1) Declares the method with return type House[] and parameter int budget.
2) Instantiates House[] result = new House[3];.
3) Calls houseSort() first so allHouses is sorted in ascending order of price.
4) Advances index i while i < allHouses.length, allHouses[i] != null, and allHouses[i].getPrice() <= budget.
5) Copies the three highest-priced qualifying houses (at indices i - 3, i - 2, and i - 1) into result[0..2] in ascending order and returns result.

Mentor tip

Sort allHouses first with houseSort(), find the last index where getPrice() <= budget (checking for null!), and copy that element and the two preceding elements into your size-3 result array.

Question 16

20 marks

Abstract data types, singly linked list ClientList operations, list splitting, and recursive binary search

Object-oriented programming

(a)

2 marks

Identify two features of an Abstract Data Type (ADT):
1) Defined by Operations and Behaviour: An ADT is a logical model defined by the data it holds and the standardized set of operations performed on it (such as push, pop, enqueue, addToEnd).
2) Implementation Hiding (Information Hiding): The underlying memory representation and internal code implementation are hidden from the user, who only interacts with the public method interface.

Mentor tip

Highlight that an ADT is defined by its standardized operations/behaviour while hiding its internal implementation details from the user.

(b)

3 marks

Describe the meaning of openDay.enList(new ClientNode("Sophie Bella", "072456721"));:
1) Instantiates a new ClientNode object in memory.
2) Initializes that node with name "Sophie Bella" and phone "072456721" (with attended = false and next = null).
3) Calls enList on the openDay list to append this new ClientNode to the tail/end of the openDay linked list.

Mentor tip

Break the statement into three parts: instantiating a ClientNode, initializing it with Sophie Bella's details, and appending it to the end of the openDay list.

(c)

5 marks

Construct the enList(ClientNode visitor) method in ClientList:
```java
void enList(ClientNode visitor) {
if (root == null) {
root = visitor;
} else {
ClientNode current = root;
while (current.getNext() != null) {
current = current.getNext();
}
current.setNext(visitor);
}
}
```
1) Checks if the list is currently empty (root == null or isEmpty()).
2) If empty, assigns root = visitor.
3) Otherwise, initializes a traversal pointer ClientNode current = root.
4) Loops while current.getNext() != null, advancing current = current.getNext() until reaching the last node.
5) Links the new node at the tail by calling current.setNext(visitor).

Mentor tip

Handle the empty list case (root == null) first, then traverse using while (current.getNext() != null) so current stops on the actual tail node before calling current.setNext(visitor).

(d)

4 marks

Outline the steps to split openDay into hasAttended and notAttended without writing code:
1) Instantiate two new empty ClientList objects named hasAttended and notAttended.
2) Start a traversal pointer at the root node of openDay and iterate sequentially through every ClientNode until the end of the list (null) is reached.
3) At each node, inspect the boolean value of its attended attribute.
4) If attended is true, append a node with that client's details to hasAttended using enList; otherwise (if false), append it to notAttended.

Mentor tip

State clearly: instantiate both new ClientList objects, traverse every node in openDay, check the attended boolean flag, and enlist the node into hasAttended or notAttended accordingly.

(e)

6 marks

Construct the recursive binary search method binSearch(String name, int low, int high):
```java
public Owner binSearch(String name, int low, int high) {
if (high < low) {
return null;
}
int middle = (low + high) / 2;
Owner current = contacts.get(middle);
if (current.getName().equals(name)) {
return current;
} else if (name.compareTo(current.getName()) < 0) {
return binSearch(name, low, middle - 1);
} else {
return binSearch(name, middle + 1, high);
}
}
```
1) Base Case 1 (Not Found): Checks if high < low (search range exhausted) and returns null.
2) Midpoint Calculation & Access: Computes int middle = (low + high) / 2 and retrieves Owner current = contacts.get(middle).
3) Base Case 2 (Found): Checks if current.getName().equals(name) and returns the matching Owner object.
4) Recursive Step (Left Sublist): If name is alphabetically smaller (name.compareTo(current.getName()) < 0), recursively returns binSearch(name, low, middle - 1).
5) Recursive Step (Right Sublist): Otherwise, recursively returns binSearch(name, middle + 1, high).

Mentor tip

Include both base cases (high < low returning null, and equals(name) returning the Owner) and ensure your recursive calls pass middle - 1 and middle + 1 with return statements.

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
Object-oriented programming65 marks · 25%
Q13, Q14, Q15, Q16

Mentor insights

Traps to watch for

01
Databases: Quick Rent-a-Car relational database keys, SQL queries, validatio · Q1
Look at the crow's foot symbol on the RENTAL side of the ERD link.
02
Modelling and simulation: Restaurant monthly electricity bill computer model, variables, an · Q5
Define it as a programmatic/mathematical representation of a real-life situation.
03
Web science: Bowtie web graph model, SCC characteristics, directed sub-graphs, · Q12
State clearly that a node represents a webpage or website.
04
Object-oriented programming: Method signatures, String immutability, null pointer handling, so · Q15
Include both the method name/identifier and the parameter list (types and order).

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