Theme A – Concepts of Computer Science: Revision Notes

Condensed for quick recall of Theme A – Concepts of computer science, which covers how computing systems actually work (SL: 38 hours; HL: 72 hours). The case study is listed separately, with its own hours at each level (see below). For the full two-theme syllabus map, use the IB DP Computer Science syllabus guide. Theme A’s Paper 1 questions also draw on the pre-released case study, so revising these four topics in isolation from the case study leaves a real gap close to the exam.

What Theme A covers

Theme A has four official sub-topics, all studied by both SL and HL students (unlike Theme B, where B.4 is HL only).

A.1 Computer fundamentals

• The hardware layer: CPU components (control unit, ALU, registers), the fetch-execute cycle, and primary vs secondary storage.
• How data of every type – number, text, image, sound – is ultimately represented as binary, and the basic units (bit, byte) used to measure it.
• The role of the operating system in managing hardware resources and mediating between applications and the underlying machine.

A.2 Networks

• Network fundamentals: clients, servers, and the distinction between a LAN and a WAN.
• The function of network devices – gateways, hardware firewalls, modems, network interface cards, routers, switches and wireless access points – and how the devices map to the layers of the TCP/IP model (A2.1.3, SL and HL).
• The network protocols used for transport and application: TCP, UDP, HTTP, HTTPS and DHCP (A2.1.4).
• HL only (A2.1.5): the function of the TCP/IP model itself – the role of each of its application, transport, internet and network interface layers, and how they interact to ensure reliable data transmission over a network.
• Data transmission concepts – packet switching, bandwidth, and common causes of transmission error – and why the internet’s architecture is described as decentralised.

A.3 Databases

• Why structured data is organised into databases rather than flat files: reducing redundancy, preserving integrity, and controlling access.
• The relational model – tables, records, fields, primary and foreign keys – and how relationships between tables are represented.
• Database programming in SQL, required at both SL and HL: the difference between SQL’s data definition language (defining the structure of the database) and data manipulation language (working with the data it holds); constructing queries between two tables, including joins, relational operators, filtering, pattern matching and ordering (SELECT, DISTINCT, FROM, WHERE, BETWEEN, ORDER BY, GROUP BY, HAVING, ASC, DESC, JOIN, LIKE with the % wildcard, AND, OR, NOT – exact syntax can vary between database systems); and how SQL is used to update data in a database.
• HL only (A3.3.4-A3.3.6): calculations with SQL’s aggregate functions on grouped data (AVERAGE, COUNT, MAX, MIN, SUM); virtual and materialized (snapshot) database views and what they offer (hiding data complexity, data consistency, independence, performance, query simplification, read-only or updatable data, security); and how transactions maintain data integrity – ACID (atomicity, consistency, isolation, durability) and the commands BEGIN TRANSACTION, COMMIT and ROLLBACK.
• HL only (A3.4): alternative databases and data warehouses.

A.4 Machine learning

• The distinction between traditional rule-based programming and machine learning, where a system derives its own rules from data.
• Core concepts: training data, features, and the difference between supervised and unsupervised learning approaches.
• Real-world applications of machine learning, and the ethical issues they raise – bias in training data, transparency of automated decisions, and the environmental cost of training large models – which the course treats as integral content, not an optional add-on.

Theme A vs Theme B, and why the split matters for revision

Theme A asks “how does a computing system work?”; Theme B (Computational thinking and problem-solving) asks “how do we use one to solve a problem?” Concretely, that means Theme A’s four topics are largely descriptive and conceptual – explaining architecture, protocols, data models and machine-learning principles – while Theme B is where programming, algorithm design and object-oriented programming live. Students sometimes over-invest revision time in Theme B because it feels more “practical”, but Paper 1 is worth an equal share of the external assessment and draws entirely from Theme A plus the case study, so under-revising A.1-A.4 leaves real marks on the table independently of programming ability.

How it’s tested

Paper 1 is set on Theme A together with the pre-released case study, so A.1-A.4 content routinely appears embedded in case-study scenarios rather than as standalone recall questions. Programming-heavy content lives in Theme B, not Theme A, so Theme A questions are more conceptual: explaining how something works, evaluating a design choice, or applying a concept (e.g. normalisation, in A.3) to an unfamiliar scenario. The exception is A.3’s SQL content, where you must be able to construct queries between two tables.

Common exam pitfalls

• Describing the fetch-execute cycle from memory as a fixed sequence without being able to explain what each stage actually does to data in the registers – examiners reward explanation of mechanism, not just naming the stages.
• In A.3, confusing a primary key (uniquely identifies a record within its own table) with a foreign key (a field that references a primary key in a different table) – these are tested as a pair, and mixing them up is one of the most common Theme A errors.
• Treating A.4 as purely conceptual and skipping the ethical dimension – the course syllabus brief explicitly frames “raising ethical issues” as one of the course’s defining characteristics, and Theme A extended-response questions can ask you to evaluate a machine-learning application on ethical as well as technical grounds.
• Assuming LAN/WAN terminology from everyday use is precise enough for exam answers – define terms using the syllabus’s own vocabulary (client, server, protocol) rather than informal substitutes.

Self-test

1. What is the difference between primary storage and secondary storage?
2. Why is a relational database preferred over storing all data in a single flat file?
3. Name two ethical issues machine learning applications can raise.
4. What is the fetch-execute cycle, in outline?
5. Tables Borrowers(BorrowerID, Surname) and Loans(LoanID, BorrowerID, BookTitle) are linked by BorrowerID. Write an SQL query listing each borrower’s surname and the titles of the books they have on loan, for surnames beginning with “Mc”, in alphabetical order of surname.

Answers: 1. Primary storage (e.g. RAM) is directly accessible to the CPU and is typically volatile and fast but limited in capacity; secondary storage (e.g. an SSD) is non-volatile, larger-capacity, but slower, and used to hold data and programs when not actively in use. 2. A relational database reduces data redundancy (data is stored once and referenced via keys rather than duplicated across records), preserves data integrity when records are updated, and allows structured, controlled access – a flat file duplicates and fragments related data with no way to enforce consistency between copies. 3. Any two of: bias in training data leading to unfair or discriminatory outputs, lack of transparency in how automated decisions are reached, and the environmental/energy cost of training large models. 4. The CPU fetches an instruction from memory, decodes what operation it specifies, and executes that operation (which may read or write data via the registers), before repeating the cycle for the next instruction. 5. SELECT Borrowers.Surname, Loans.BookTitle FROM Borrowers JOIN Loans ON Borrowers.BorrowerID = Loans.BorrowerID WHERE Borrowers.Surname LIKE 'Mc%' ORDER BY Borrowers.Surname ASC; – the JOIN links the tables through the foreign key, LIKE with the % wildcard filters by pattern, and ORDER BY sorts the result.

Official syllabus

International Baccalaureate Organization, Diploma Programme Subject Brief – Sciences: Computer science, first assessment 2027, © 2024.