Command terms

Every part of an IB Computer Science exam question starts with a verb such as state, outline, explain or evaluate. That verb tells you how much to write and what kind of thinking the examiner will reward. Two questions about the same idea, such as cache memory, can ask for one word or for a balanced argument, and the answer that scores full marks on one scores almost nothing on the other.

Each command term has a fixed meaning, an assessment objective and a recognisable shape of full-mark answer in computer science. Knowing them also shapes how you revise and how you spend your time on the day.

How command terms work

The IB publishes a fixed list of command terms for the course, each with an official meaning and an assessment objective. The meanings are the same in every question and every year, so learning them once pays off in every answer you write.

Each command term is linked to one of the course's assessment objectives. The written papers assess three of them; a fourth, solving a real-world problem with computational thinking, is assessed only through the internal assessment.

The AO staircase groups all 21 command terms by objective. The higher the step, the more reasoning each mark needs: an AO1 mark is earned by a correct fact, while an AO3 mark is earned by a reason, a judgement or something you build.

Both Paper 1 and Paper 2 split their marks roughly 50% AO1 and AO2 together, and 50% AO3. Half of every paper is therefore questions that ask you to explain, compare, construct, discuss, evaluate or justify. Recalling and applying facts, however well, can earn only the other half.

What a markscheme is

Examiners mark against a markscheme written for each paper. Knowing how markschemes are laid out explains most exam advice: why one point earns one mark, why the second half of an explanation matters, and why extra answers don't help.

The command terms one by one

The tables group the terms by AO. Each row gives the term's meaning in plain words, what it demands in a computer science answer, and a short example question with a typical mark value. The marks in the examples are realistic, not fixed: the same term can carry different marks in different questions.

AO1: knowledge and understanding

AO1 terms test whether you know a fact precisely. The answers are short, and the marks are lost through vagueness more than through missing knowledge.

The difference between a 0 and a 1 on a Define question is usually one qualifying phrase. "A cache hit is when data is in the cache" misses what makes the event matter: the data is found in the cache, so it is supplied without reading the slower main memory.

AO2: applying and using knowledge

AO2 terms ask you to use what you know: give an account, work something out, pick the right answer, or follow an algorithm. Identify and Trace are AO2 in this course, even though some other subjects treat them differently.

The Calculate and Estimate rows use the same image, which shows how the two terms differ. Calculate wants an exact figure with every stage shown; Estimate accepts a rounded figure as long as the method is sensible.

Trace questions appear mainly in Paper 2, but the technique is the same wherever they appear: one column per variable, one row each time a value changes, and the output recorded when it happens. This short program adds up the even numbers from 1 to 4.

total = 0
for i in range(1, 5):
    if i % 2 == 0:
        total = total + i
print(total)

int total = 0;
for (int i = 1; i < 5; i++) {
    if (i % 2 == 0) {
        total = total + i;
    }
}
System.out.println(total);

AO3: constructing, analysing and evaluating

AO3 terms need reasons, judgement or something built from scratch. Construct and Sketch belong here: both ask you to produce a working artefact (a truth table, a query, a diagram) from the information in the question, not to reproduce one from memory.

Compare is the term students misread most. The official wording asks for similarities, referring to both items throughout. In practice, good answers give similarities and differences, each sentence naming both items, because a comparison with no differences rarely says much about two technologies. Some syllabus statements say "compare and contrast" (subtopic A2.2, on networking models, is one), which asks for both explicitly.

Quirks of the course's list

Four details of this course's list catch students who learned the terms in another subject or from older Computer Science material.

• Identify is AO2. It still needs only the right item, but it often asks you to pick from code or a diagram, which is an application skill.

• Trace is AO2. It is a method question: marks go to the record of each step.

• Construct and Sketch are AO3. Building a correct query or diagram is treated as a higher-order skill, and these parts often carry 3–6 marks.

• Contrast is not on the list. The guide says other terms may sometimes be used, and "contrast" appears in some syllabus statements. If it appears in a question, read it in its ordinary meaning: differences only, like Distinguish.

One idea at three depths

Answering the same question at three depths shows how Outline, Describe and Explain differ. The topic here is why a computer with more cache can run a program faster.

Each sentence of the Explain answer contains a cause and an effect, and the final mark goes to linking the mechanism back to the question's "faster". A student who writes six true facts about cache without the links can still score only two.

How marks scale

Command terms and mark values work together. The term says what kind of point earns a mark; the number in brackets says how many such points the examiner expects. The patterns in the table, like the rules after it, are inferred from earlier IB Computer Science papers, so treat them as a guide, not a guarantee.

Some rules apply across every term:

1. Answer the number of items asked for. If a question says "two", examiners mark only your first two. A third answer earns nothing, and a wrong one listed first can cost a mark.

2. Don't repeat a point in other words. "It is faster" and "it has less delay" are usually the same point and score once.

3. Write each point as a separate idea. A list or one sentence per point makes it easy for the examiner to award each mark.

4. Use the scenario. When the question gives a context, markschemes often say the answer must be applied to it, and generic answers can be capped.

5. Show working. In calculations, traces and constructions, method marks can be earned even when the final answer is wrong, and a wrong value carried correctly into a later part can still score there.

Answers to AO3 terms that build or judge

Evaluate, Discuss and To what extent need the most practice, because they carry the most marks per question and are marked as a whole, often against level descriptors. The shorter AO3 terms, Justify and Suggest, follow the same logic in fewer marks: a claim, then a reason that depends on the scenario.

Construct and Deduce are AO3 too, but they are marked by the correctness of what you produce. For SQL, that means the right columns, the right table, the right condition and the right order.

SELECT name
FROM Student
WHERE year_group = 12 AND house = 'Blue'
ORDER BY name;

Common pitfalls with command terms

Most lost marks on command terms come from a small number of habits. Each is easy to fix once you can see it in your own answers.

Another common slip is writing a definition that uses the word being defined ("Encryption is when data is encrypted"). Define questions need the meaning in other words, such as "converting plaintext into ciphertext using an algorithm and a key so that only someone with the right key can read it".

Students also over-answer AO1 terms. A State question needs a value or a name, and any explanation added is ignored. The time is better spent on the AO3 parts at the end of the question.

Revising with command terms

Good revision turns every syllabus statement into the questions it can produce. The statement's command term tells you the main depth to prepare, so you can predict most question types for a topic before you have seen a paper.

Plan by teaching hours and paper

The recommended teaching hours give a rough guide to where the marks fall. Topics with more hours have more content and appear in more questions, so they deserve more revision time.

Source: IB Computer Science guide, first assessment 2027, recommended teaching hours.

Programming alone is 40 of the SL course's 105 hours of syllabus content (the case study included), which is why Paper 2 is dominated by code. At HL, theme A's four topics have 72 hours and the case study 30, against 93 for theme B, so Paper 1 needs at least as much revision time as Paper 2.

Active recall, not rereading

Rereading notes feels productive but builds little memory. Retrieval practice, where you answer from memory and then check, works far better. Three methods suit this course:

• Definition cards. Put each glossary term on a card or in a flashcard app, with the definition on the back. Test yourself until you can write each one with its qualifying words.

• Command-term drills. Take one syllabus statement, write a 2-mark Outline, a 3-mark Describe and a 4-mark Explain on it, then check them against your notes.

• Past-style questions under time. No past papers for this course exist yet, so use the specimen papers from your teacher and exam-style questions written for each subtopic. Mark your answers strictly against the markscheme, one point at a time.

Write code by hand

Paper 2 is written without an IDE, a compiler or autocomplete. Practise writing functions and methods on paper, then type them in and run them to find your mistakes. Pay attention to indentation in Python and braces and semicolons in Java; markschemes usually tolerate small syntax slips but not ambiguous logic, so clear layout protects your marks. Some Paper 2 questions also ban built-in functions such as sort, max or len, so revise the hand-written loop for each one as well as the one-line call. Layout, marking and the hand-written replacements are covered in full on the page "Paper 2: reading and writing code".

On the day

Exam technique is mostly about reading carefully and spending time where the marks are. The following routine works for every question on both papers.

1. Read the whole question first, including every part, so you know where it is going and don't use up a good point too early.

2. Mark the command term and the marks for each part. Decide how many points the part needs before you write.

3. Mark the context. Underline the scenario's names and details, and use them in your answer.

4. Answer the number of items asked for, each as a separate sentence or bullet.

5. Show working for calculations, traces and constructions.

6. Keep time for the long parts. Discuss, Evaluate and To what extent parts often come last in a question and carry the most marks.

7. Leave a few minutes at the end to check that every part has an answer. A blank part scores nothing, and a short attempt often scores a mark.

Paper 1's two sections are sat together without a break, so you decide how to split your time between them; the page "Paper 1: concepts questions" gives the split for each level.

Checklist

• To do:

  I can give the meaning of all 21 command terms and the AO each belongs to.

• To do:

  I know that Identify and Trace are AO2, and Construct and Sketch are AO3.

• To do:

  I can answer one question at Outline, Describe and Explain depth and say what each extra mark is for.

• To do:

  I write Compare answers that name both items in every point.

• To do:

  I can tell how many points a part needs from its command term and marks.

• To do:

  I answer only the number of items asked for.

• To do:

  I show working in every Calculate, Trace and Construct answer.

• To do:

  I give strengths, limitations and a conclusion in every Evaluate and Discuss answer.

• To do:

  I revise with active recall, using definition cards and timed questions.

• To do:

  I can write code by hand, including the hand-written version of common built-ins.

• To do:

  I can pace a paper at about 1.5 minutes per mark.