📚 Year 12 CIE Computer Science: Exam Techniques and Marking Criteria | Year 12 CIE 计算机：答题技巧与评分标准
Mastering the CIE AS Computer Science (9618) examinations requires more than just subject knowledge. Understanding how examiners award marks, what command words demand, and how to structure your answers can make a significant difference between a good grade and an excellent one. This guide breaks down the essential exam techniques and marking criteria, offering clear strategies for every question type you will encounter in Papers 1 and 2.
掌握 CIE AS 计算机科学（9618）考试，仅仅掌握学科知识还不够。理解考官如何评分、指令词要求什么以及如何组织答案，会对最终成绩产生显著影响。本指南详细解析关键的答题技巧与评分标准，为你应对 Paper 1 和 Paper 2 中的各类题型提供清晰策略。
1. Understanding Command Words | 理解指令词
Every question on a CIE Computer Science paper uses specific command words that tell you exactly what to do. ‘State’ means a brief, concise answer without explanation, often just a single word or short phrase. ‘Describe’ requires you to state the main characteristics of something, often in context. ‘Explain’ asks for reasons, causes, or mechanisms, so your answer must include ‘because’ or a logical chain of reasoning. ‘Compare’ needs you to identify similarities and differences, usually in a structured way.
CIE 计算机科学试卷中的每一道题都使用特定的指令词，明确告诉你该做什么。’State’ 意味着给出简短、精确的答案，无需解释，通常只是一个词或短语。’Describe’ 要求你陈述事物的主要特征，往往要结合上下文。’Explain’ 要求给出原因、起因或机制，所以你的回答必须包含 ‘因为’ 或逻辑推理链。’Compare’ 要求你辨识相同点和不同点，通常要以结构化的方式呈现。
Significant marks are lost every year when students write a long description for a ‘state’ question, or give a simple list for an ‘explain’ question. For a question that says ‘Give three reasons why…’, do not write a paragraph; list each reason clearly, ideally on separate numbered lines. The table below summarises the most common command words and their expected responses.
每年都有大量的分数，因为学生在 ‘state’ 问题上写了冗长的描述，或在 ‘explain’ 问题上只列出要点而丢失。对于 ‘Give three reasons why…’ 这样的问题，不要写成段落；清楚地列出每个理由，最好用编号分行。下表总结了最常见的指令词及其期望的回答形式。
| Command Word | Meaning | 指令词 | 含义 |
|---|---|---|---|
| State | Give a short, factual answer | 陈述 | 给出简短的事实性答案 |
| Describe | Write the main features or characteristics | 描述 | 写出主要特征或特点 |
| Explain | Give reasons or mechanisms; show cause and effect | 解释 | 给出原因或机制；展示因果关系 |
| Compare | Identify similarities and differences | 比较 | 识别相同点和不同点 |
| Define | Give the precise meaning of a term or concept | 定义 | 给出术语或概念的精确含义 |
| Identify | Name or select the correct answer from given information | 识别 | 从给定信息中命名或选择正确答案 |
2. Answering Explanation Questions | 回答解释类问题
In Paper 1, ‘explain’ questions often carry the highest marks per item, typically 3–6 marks each. A common mistake is to give a one-sentence answer and expect full credit. The mark scheme usually expects a step-by-step reasoning process. For example, when explaining how an operating system manages interrupts, you should describe the sequence from interrupt signal generation to saving the processor state, calling the ISR, and restoring the state.
在 Paper 1 中，’explain’ 类问题通常每题分值最高，一般是 3 到 6 分。常见的错误是给出一个句子就想拿到全分。评分方案通常期望循序渐进的推理过程。例如，在解释操作系统如何管理中断时，你应当描述从中断信号产生，到保存处理器状态、调用中断服务程序（ISR）以及恢复状态这一完整序列。
Use linking words like ‘first’, ‘then’, ‘as a result’, and ‘this causes’ to make your logical flow explicit. If the question asks for an explanation of a concept like virtual memory, do not just state what it is; explain why it is needed, how the paging mechanism works, and what happens when a page fault occurs. Always connect your sentences to the ‘why’ and ‘how’.
使用像 ‘first’、’then’、’as a result’ 和 ‘this causes’ 这样的连接词，让你的逻辑流程变得清晰明确。如果题目要求解释虚拟内存这样的概念，不要只说它是什么；要说明为什么需要它、分页机制如何工作，以及发生缺页时会发生什么。始终把你的句子与 ‘为什么’ 和 ‘如何’ 连接起来。
3. Describing Processes and Algorithms | 描述过程与算法
Description questions require you to present a clear and ordered account of how something works. In computer science, this often involves algorithms, protocols, or hardware interactions. When describing a search algorithm like binary search, you must state the prerequisite (sorted list), explain the midpoint calculation, and describe the iterative or recursive decision process. Avoid vague language like ‘it checks the middle and splits’; instead, be precise: ‘The algorithm calculates the middle index as floor((low + high)/2) and compares the target value with the element at that index.’
描述类问题要求你对某事物如何工作给出清晰、有序的陈述。在计算机科学中，这往往涉及算法、协议或硬件交互。当描述二分查找这样的搜索算法时，你必须声明前提条件（有序列表），解释如何计算中点，并描述迭代或递归的决策过程。避免使用模糊的语言，比如 ‘它检查中间并分割’；而要精确地表述：’算法计算中间索引为 floor((low + high)/2)，并将目标值与位于该索引的元素进行比较。’
For processes like the fetch–decode–execute cycle, stepwise bullet points or numbered lists can be very effective. However, ensure that each step is a sentence, not just a keyword. The mark scheme allocates marks for each logical step correctly identified and explained in sequence. Practise writing these descriptions under timed conditions, as they are predictable and high-yield when you master the wording.
对于像取指–译码–执行周期这样的过程，逐步的要点或编号列表非常有效。但请确保每一步都是一句话，而不仅仅是一个关键词。评分方案会为按顺序正确识别和解释的每个逻辑步骤分配分数。在限时条件下练习书写这些描述，因为一旦你掌握了措辞，这些内容既可预测又容易得分。
4. Comparing and Contrasting | 比较与对比
Compare questions frequently appear in topics like networking (e.g., LAN vs WAN), scheduling algorithms, and data structures. A side-by-side comparison is often the best approach, either in a single table if the question allows, or by clearly addressing one item for similarity, then the other, then a difference. Never write about one item completely and then the other; the mark scheme rewards direct, paired comparisons.
比较类问题经常出现在网络（例如 LAN 与 WAN）、调度算法和数据结构等主题中。并排比较通常是最好的方式，如果题目允许，可采用表格形式；或者先说明一个项目的相似点，再说明另一个，然后指出不同点。绝不要先完整描述一个再描述另一个；评分方案奖励直接的、成对的比较。
Use comparative phrases: ‘Both A and B use…’, ‘Unlike A, B relies on…’, ‘A is faster whereas B is more reliable’. When comparing programming concepts like iteration and recursion, highlight the stack usage, performance, and readability. Aim for two columns of similarities and differences in your mind, even if you write them as continuous text.
使用比较性短语：’Both A and B use…’、’Unlike A, B relies on…’、’A is faster whereas B is more reliable’。在比较迭代和递归等编程概念时，要突出栈的使用、性能和可读性。即使你写成连贯的文字，也要在脑海中呈现两列相似点和不同点。
5. Pseudocode and Programming Tasks | 伪代码与编程任务
Paper 2 contains significant marks for writing pseudocode, tracing algorithms, and completing program code. CIE has a published pseudocode guide; your syntax must be consistent and match the expected format. Use keywords like INPUT, OUTPUT, FOR … TO … NEXT, WHILE … DO … ENDWHILE, and IF … THEN … ELSE … ENDIF. Indentation is crucial for readability – use it consistently to show the structure of loops and selection.
Paper 2 中包含大量关于书写伪代码、跟踪算法和完成程序代码的分数。CIE 发布了官方的伪代码指南；你的语法必须一致并符合预期格式。使用 INPUT、OUTPUT、FOR … TO … NEXT、WHILE … DO … ENDWHILE 以及 IF … THEN … ELSE … ENDIF 等关键词。缩进对可读性至关重要——始终如一地使用缩进来展示循环和选择的结构。
When tracing an algorithm, fill in a trace table neatly. Each variable should have a column, and each row should represent one iteration or step. Do not skip steps, even if the algorithm loops many times; examiners check for correct intermediate values. If the question asks ‘what is the purpose of this algorithm?’, your answer must summarise the overall function in one clear sentence, not repeat the code line by line.
跟踪算法时，工整地填写跟踪表。每个变量应有一列，每一行代表一次迭代或一步。即使算法循环多次，也不要跳过步骤；考官会检查中间值是否正确。如果题目问 ‘what is the purpose of this algorithm?’，你的回答必须用一句清晰的话概括整体功能，而不是逐行复述代码。
6. Diagrams and Representations | 图表与表示法
Exam questions often ask you to draw a logic circuit, a flowchart, a tree data structure, or a network diagram. These must be drawn with a ruler and pencil, clearly labelled. For logic gates, use the standard symbols (AND, OR, NOT, NAND, NOR, XOR) from the syllabus and ensure connections are neat. In flowcharts, use the correct shapes: ovals for start/stop, parallelograms for input/output, rectangles for processes, and diamonds for decisions.
试题常要求你绘制逻辑电路、流程图、树形数据结构或网络图。这些必须用直尺和铅笔绘制，并清晰标注。对于逻辑门，使用教学大纲中的标准符号（AND、OR、NOT、NAND、NOR、XOR），并确保连线整洁。在流程图中，使用正确的形状：椭圆形表示开始/结束，平行四边形表示输入/输出，矩形表示处理过程，菱形表示判断。
In tree diagrams, show the nodes and pointers accurately; when inserting or deleting nodes, redraw the final tree. Marks are awarded for correct structure, correct positions, and correct labels. Never cross arrows if you can avoid it, and if a diagram becomes messy, start again – a clean diagram can secure 2–3 extra marks that a messy one might lose.
在树形图中，准确显示节点和指针；在插入或删除节点后重绘最终树形。结构正确、位置正确和标签正确都能得分。尽量避免箭头交叉，如果图表变得混乱，就重新绘制——一张清晰的图表能多获得 2–3 分，而混乱的图表可能导致失分。
7. Time Management | 时间管理
Paper 1 has 75 marks in 90 minutes – roughly 1.2 marks per minute. Paper 2 also has 75 marks but in 120 minutes. Use the mark count as a guide for how many minutes to spend on a question. A 4-mark question should take about 5–6 minutes, not 15. If you get stuck, move on and return later. Always reserve the last 10 minutes for reviewing your answers, especially checking for missed command words or calculation errors.
Paper 1 有 75 分，用时 90 分钟——大约每分钟争取 1.2 分。Paper 2 同样有 75 分，但用时 120 分钟。使用分值为指导，决定一道题花费多少分钟。一道 4 分的题应花大约 5–6 分钟，而非 15 分钟。如果卡住了，就跳到下一题，最后再返回。务必留出最后 10 分钟检查答案，尤其要注意遗漏的指令词或计算错误。
In Paper 2, the programming problem at the end often requires extensive code writing. Allocate at least 25 minutes to it. Read the entire question before starting to code – understand the inputs, outputs, and edge cases. Quickly jot down the algorithm in pseudocode before translating it into program code. This planning step saves time and reduces logic errors.
在 Paper 2 中，最后的编程题通常需要编写大量代码。至少分配 25 分钟给它。在开始编码之前通读全题——理解输入、输出和边缘情况。在转化为程序代码前，先用伪代码快速草拟算法。这个规划步骤能节约时间并减少逻辑错误。
8. Common Mistakes to Avoid | 常见错误避免
One of the most frequent errors is answering the question you wish you were asked, rather than the one actually set. Circle the command word and underline keywords in every question. Another mistake is providing answers that are too vague: ‘It is faster’ or ‘It uses less memory’ without quantifying or explaining why. In Paper 1, always include technical terms; for example, say ‘data is transmitted in packets’ rather than ‘data is sent in small pieces’.
最常见的错误之一，是回答了你希望被问到的问题，而不是实际设定的问题。圈出每道题中的指令词，并在关键词下划线。另一个错误是答案过于模糊：只说 ‘它更快’ 或 ‘它占用更少内存’，而不量化或解释原因。在 Paper 1 中，始终使用技术术语；例如，说 ‘数据以数据包形式传输’，而不是 ‘数据以小块形式发送’。
In Paper 2, forgetting to initialise variables, using incorrect loop boundaries, or not handling empty input cases are typical slip-ups. Trace your code with a simple test case mentally. Also, many students write the full program but fail to output the required result, or output extra debug text that does not match the specification. Follow the output format exactly.
在 Paper 2 中，忘记初始化变量、使用错误的循环边界或不处理空输入情况是典型的疏忽。在心里用一个简单的测试用例跟踪你的代码。此外，许多学生写出了完整的程序，却未能输出所需结果，或者输出了与规范不符的额外调试文本。要严格遵循输出格式。
9. Understanding Mark Schemes | 理解评分方案
The CIE mark scheme is your best friend during revision. It shows exactly what the examiner expects for each point. For explanation questions, marks are often awarded for ‘each mark point’ – a single correct fact or logical step. For example, in a 4-mark explanation of how SSL/TLS works, you might need to mention the handshake, certificate exchange, symmetric key generation, and encrypted data transfer. Study past mark schemes to understand the granularity of marking.
CIE 评分方案是你在复习时最好的帮手。它精确地展示了考官期望的每一个要点。对于解释类问题，分数通常按 ‘评分点’ 给予——每个正确的事实或逻辑步骤得 1 分。例如，在解释 SSL/TLS 工作原理的 4 分题中，你可能需要提到握手、证书交换、对称密钥生成以及加密数据传输。研究往年评分方案，以理解评分的颗粒度。
For programming tasks, the mark scheme often has separate rows for input, processing, output, and correct syntax. Therefore, even if your code doesn’t fully work, you can still gain marks for correct input statements and output formatting. In descriptive questions, look for the essential concepts mentioned in the mark scheme; use the same keywords as the syllabus.
对于编程任务，评分方案通常为输入、处理、输出和正确语法分别设行。因此，即使你的代码不能完全运行，正确的输入语句和输出格式仍能得分。在描述类问题中，寻找评分方案中提到的核心概念；使用与教学大纲相同的关键词。
10. Using Past Papers Effectively | 有效使用历年真题
Past papers are not just a test of knowledge; they are a training tool. Start by attempting a paper under timed conditions, then use the mark scheme to mark it strictly, noting every lost mark. Categorise your errors: was it a knowledge gap, a misinterpretation, or a careless mistake? Then re-attempt the same paper a few days later, aiming for full marks. This repeated exposure builds exam fluency.
历年真题不仅是知识的检验，更是一种训练工具。先在限时条件下完成一份试卷，然后对照评分方案严格评分，记录每一个丢失的分数。将你的错误分类：是知识漏洞、理解偏差，还是粗心大意？几天后重新做同一份试卷，力争拿到满分。这种重复训练能增强考试的流畅性。
Create a ‘common mistake log’ where you write down the exact error and the correct approach. For example: ‘Said HTTP is encrypted – actually HTTP is plain text; HTTPS is encrypted’. Revise this log before the actual exam to prevent repeating the same mistakes. Also, practise papers from different sessions, because topics are sometimes repeated with slight variations.
制作一本 ‘常见错误日志’，写下确切的错误及正确的方法。例如：’说 HTTP 是加密的——实际上 HTTP 是明文；HTTPS 是加密的’。在真正考试前复习该日志，以避免重复犯同样的错误。此外，练习不同考季的试卷，因为主题有时会以略有变化的形式重复出现。
11. Tackling Long-answer Questions | 攻克长篇答题
Some questions, especially in Paper 1, are extended response items worth 6–10 marks. These often ask you to discuss the advantages and disadvantages of a technology, or to justify a choice between two solutions. Structure your answer with a brief introduction, then a series of comparative paragraphs, and a conclusion if the question asks for a recommendation. Use technical reasoning rather than personal opinion.
有些问题，特别是 Paper 1 中的长答题，是价值 6–10 分的扩展回答题。这些题目通常要求你讨论某种技术的优点和缺点，或为在两种解决方案之间做出选择提供理由。用简短的引言构建你的回答，然后是一系列比较性段落，如果题目要求给出建议，则加上结论。使用技术性推理而非个人意见。
When discussing a topic like ‘the use of interpreters vs compilers’, pair each advantage with a corresponding disadvantage, e.g., ‘Interpreters provide better error reporting during development, but compilers produce faster-executing standalone code.’ This balanced approach often earns more marks than a one-sided argument. Use examples from real programming environments to strengthen your points.
在讨论诸如 ‘解释器与编译器的使用’ 的话题时，将每个优点与相应的缺点配对，例如，’解释器在开发期间提供更好的错误报告，但编译器生成更快的独立可执行代码。’这种平衡的方法通常比单方面的论证能获得更多分数。使用真实编程环境中的例子来加强你的观点。
12. Final Tips for Exam Day | 考试日最后提示
Read the entire question paper during the reading time, noting which questions you feel most confident about. Start with those to build momentum and reduce anxiety. Bring a clear pencil case with multiple pens, pencils, an eraser, and a ruler (for diagrams). Remember that a blank page yields no marks; even if you are unsure, write down relevant facts or attempt a partial solution. Marks can be awarded for correct sub-steps.
在阅读时间内通读整份试卷，标注出你最有信心回答的题目。从这些题目开始做，以建立答题势头并减少焦虑。携带一个透明的铅笔盒，内装多支笔、铅笔、橡皮和直尺（用于绘图）。记住，空白页得不到任何分数；即使你不确定，也要写下相关事实或尝试部分解答。正确的子步骤也能得分。
Finally, manage your wellbeing: get a full night’s sleep before the exam, eat a balanced meal, and stay hydrated. Cognitive performance drops significantly when you are tired or hungry. During the exam, if you feel panicked, take ten seconds to close your eyes, breathe deeply, and reset. You’ve prepared thoroughly; now trust your training and show the examiner what you know.
最后，管理好你的身心状态：考前一晚保证充足睡眠，吃一顿营养均衡的饭，并保持水分。疲劳或饥饿时，认知能力会显著下降。考试中如果感到慌乱，花十秒钟闭上眼睛，深呼吸，然后重新开始。你已经充分准备；现在要相信自己的训练，向考官展示你所掌握的知识。
Published by TutorHao | Computer Science Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导