📚 CAIE IGCSE Computer Science Exam Techniques and Marking Criteria | CAIE IGCSE 计算机科学答题技巧与评分标准
Mastering the CAIE IGCSE Computer Science exam requires more than just knowing the content; you must understand how to answer questions precisely to meet the mark scheme. This guide breaks down essential techniques and criteria for both Paper 1 (Theory) and Paper 2 (Problem-solving and Programming), helping you maximise your marks in Year 11.
掌握 CAIE IGCSE 计算机科学考试不仅需要掌握知识，还必须理解如何精确答题以满足评分标准。本指南详细解析试卷一（理论）和试卷二（问题解决与编程）的关键答题技巧和评分要点，帮助 Year 11 的同学拿到最高分。
1. Understanding Command Words | 理解指令词
Command words like 'State', 'Describe', 'Explain' indicate the depth of response required. 'State' expects a brief answer, often one word or a short phrase. For example, 'State the name of the network topology shown' could be answered with 'Star'.
指令词如“State”（陈述）、“Describe”（描述）、“Explain”（解释）指示所需回答的深度。“State”要求简短回答，通常一个词或一个短语。例如“陈述图中所示网络拓扑的名称”，可回答“星型”。
'Describe' requires a more detailed account, such as 'Describe how data is transmitted in a star network'. You should mention that each device is connected to a central hub/switch, data is sent to the hub which forwards it, and only the intended recipient reads it.
“Describe”要求更详细的说明，如“描述数据在星型网络中是如何传输的”。你应提到每台设备都连接到中央集线器/交换机，数据发送到集线器后再转发，只有目标接收方会读取数据。
'Explain' goes further, asking for reasons or effects. If asked 'Explain why a star network is more reliable than a bus network', you need to state that if one cable fails in a star, only that node is affected, whereas in a bus, the whole network can fail, and then explain the implication on network uptime.
“Explain”更进一步，要求给出理由或影响。如果被问“解释为什么星型网络比总线型网络更可靠”，你需要指出在星型网络中一根电缆故障只影响该节点，而在总线型中整个网络可能瘫痪，并解释这对网络正常运行时间的影响。
Other crucial command words include 'Identify' (name or pick out from a list), 'Compare' (point out similarities and differences), and 'Calculate' (work out using given formulas). Always check the command word before writing your answer so that you provide the right level of detail.
其他重要的指令词还有“Identify”（识别，从列表中选出）、“Compare”（比较，指出相同与不同之处）、“Calculate”（计算，用给定的公式算出结果）。作答前务必先看指令词，确保你给出了正确深度的答案。
2. Answering Theory Questions Effectively | 高效回答理论题
Paper 1 theory questions often award one mark per valid point. You must use precise subject terminology. For instance, when asked about advantages of SSDs over HDDs, write 'no moving parts, faster data access, lower power consumption' rather than 'they are better and faster'.
试卷一中的理论题通常每给出一个有效要点就得一分。你必须使用准确的学科术语。例如，被问到固态硬盘(SSD)相比机械硬盘(HDD)的优点时，要写“没有移动部件、数据访问更快、功耗更低”，而不是“它们更好更快”。
If a question expects two advantages, list exactly two distinct points. Adding a third one will not earn extra marks and wastes time. Use bullet points in your answer if the question format allows, as this helps examiners see your points clearly.
如果题目要求写两个优点，就精确列出两个不同的要点。多写第三个并不会加分，反而浪费时间。如果题目格式允许，可以在答案中使用项目符号，这样便于阅卷人看清你的每个得分点。
For extended response questions, plan a short structure: definition, explanation, example. Stick to the mark allocation: a 4-mark question may need four distinct statements or two well-explained ones, depending on the command word.
对于扩展作答的题目，可以简单构思一个结构：定义、解释、举例。紧扣分值分配：一道4分的题可能需要四个独立的陈述，或两个到位解释的要点，这取决于指令词。
3. Writing Pseudocode for Algorithms | 编写算法伪代码
CAIE does not mandate a strict pseudocode syntax, but your solution must be clear, logical, and use consistent conventions. Examiners award marks for correct use of INPUT/OUTPUT, variable declaration, iteration (FOR, WHILE), selection (IF…THEN…ELSE…ENDIF), and tracing logic.
CAIE 并未强制规定伪代码的严格语法，但你的方案必须清晰、合乎逻辑，并且使用一致的书写约定。阅卷人会给分在正确使用 INPUT/OUTPUT、变量声明、循环(FOR, WHILE)、选择(IF…THEN…ELSE…ENDIF)以及追踪逻辑上。
Always initialise variables before using them, e.g., Total ← 0. Include comments using // to explain non-obvious steps. For a loop that sums an array of 5 numbers, a clean pseudocode would be:
务必在使用变量前对其进行初始化，例如 Total ← 0。用 // 加注注释以解释不明显的步骤。对于求包含5个数字的数组之和的循环，一段简洁的伪代码可以是：
Total ← 0
FOR Index ← 1 TO 5
INPUT Num
Total ← Total + Num
NEXT Index
OUTPUT Total
Marks are often given for correct loop bounds (1 TO 5), correct accumulation, and final output. Indentation improves readability and is sometimes credited under clarity marks in Paper 1.
得分的点常在于正确的循环边界（1 TO 5）、正确的累加操作以及最终输出。缩进可以提高可读性，在试卷一的评分里，清晰度有时也可以得分。
4. Programming Questions and Mark Allocation | 编程题与分值分配
In Paper 2, you will write actual code (typically in Python). The mark scheme allocates points for: valid input statements, correct use of data types, proper control structures, correct logic/calculations, and formatted output. Even if the final output is wrong, you can earn marks for the approach.
在试卷二中你需要编写实际代码（通常使用 Python）。评分标准按以下分配分数：有效的输入语句、正确使用数据类型、恰当的控制结构、正确的逻辑/计算、格式正确的输出。即使最终输出有误，你仍可因解题思路得分。
For example, a task to input a temperature and output 'Hot' if above 30, 'Cold' if below 10, otherwise 'Mild' might be worth 5 marks: 1 for input, 1 for correct IF structure, 1 for each condition correct, and 1 for appropriate output message.
例如，一道要求输入温度，若高于30输出'Hot'，低于10输出'Cold'，否则输出'Mild'的题目可能值5分：1分给输入，1分给正确的 IF 结构，每种条件正确各得1分，1分给恰当的输出信息。
Always handle edge cases: if a range is 0 to 100 inclusive, check for values outside this range and reject them, earning validation marks. Use meaningful variable names; while they may not be directly awarded, they reduce logical errors.
永远要处理边界情况：如果范围是0到100（含），检查超出该范围的值并拒绝它们，这样能获得验证分数。使用有意义的变量名，虽然不直接得分，但能减少逻辑错误。
5. Handling Trace Table Questions | 处理跟踪表题目
Trace tables test your ability to step through an algorithm line by line. The examiner expects a column for each variable and condition, with rows showing values after each significant step. Leave no cell blank; write the value or a dash.
跟踪表考查你逐行执行算法的能力。阅卷人期望每个变量和条件各占一列，每一行显示执行某重要步骤后的数值。不要留空单元格，填入数值或短划线。
Given the code x = 2; for i = 1 to 3: x = x + i, the trace table would look like:
给定代码 x = 2; for i = 1 to 3: x = x + i，其跟踪表如下：
| Step | i | x |
|---|---|---|
| Initial | - | 2 |
| 1 | 1 | 3 |
| 2 | 2 | 5 |
| 3 | 3 | 8 |
Common mistakes include skipping iterations, updating values out of sequence, or ignoring that a variable may change after a loop body. Each correctly filled row typically scores one mark.
常见错误包括跳过迭代、不按顺序更新值，或忽略变量可能在循环体结束后改变。通常每个正确填充的行可以得1分。
6. Data Representation and Calculation Tips | 数据表示与计算技巧
Questions on binary, hexadecimal, and binary addition require you to show all working steps. For a binary-to-hexadecimal conversion, group bits from right to left in fours, then convert each nibble. A clear layout with interim steps earns method marks even if the final answer is wrong.
涉及二进制、十六进制和二进制加法的题目要求你展示所有解题步骤。做二进制到十六进制转换时，从右向左每四位一组，然后转换每个半字节。即使最终答案错误，清晰的布局和中间步骤仍可获得方法分。
When adding two binary numbers, write them aligned, carry over bits explicitly, and note any overflow. If the question is about two's complement, explicitly invert bits and add 1 for conversion. Provide your answer in the requested format, e.g., 8-bit representation.
进行两个二进制数相加时，要对齐书写，明确标出进位，并记录任何溢出。如果题目涉及二进制补码，要明确写出取反加1的转换过程。将答案写成要求的格式，例如8位表示。
For calculations like file size estimation, show the formula: File size = sample rate × bit depth × duration and substitute numbers carefully. Marks are awarded for correct substitution and unit conversion (e.g., bits to bytes).
对于估算文件大小等计算题，列出公式：文件大小 = 采样率 × 位深度 × 时长，仔细代入数值。正确代入数据和单位换算（如位转字节）都能得分。
7. Validation, Verification and Error Handling | 验证、核实与错误处理
to check if age ≥ 0 AND age ≤ 120 before processing is a typical 1-mark validation point.
在试卷二的编程任务中，如果题目规范要求特定的数据类型或范围，务必加入输入验证。在处理前用 IF 语句检查年龄是否在0至120之间，这就构成了一个典型的1分验证点。
Verification aims to ensure data entered is correct, e.g., asking the user to re-enter a password and comparing the two entries. You might write a loop that repeats until both inputs match, which shows understanding of verification logic and earns additional marks.
核实旨在确保输入的数据正确，例如让用户再次输入密码并对比两次输入。你可能会写一个循环，直到两次输入匹配才继续，这体现出对核实逻辑的理解，并能获得额外分数。
Even if the specification does not explicitly ask for error handling, adding a simple check for negative values or empty input can indicate 'robust code' in a high-band answer. Do not over-engineer, though, and leave time for the core logic.
即使题目规范里没有明确要求错误处理，增加一个简单的负数或空输入检查可以在高分段回答中体现“代码健壮性”。但不要过度设计，要为核心逻辑留下时间。
8. Using Examples to Support Answers | 用例子支持答案
Many 'Describe' and 'Explain' questions explicitly require an example. If a question asks about cyber security threats, stating 'one threat is phishing' and then giving a brief example, like 'an email pretending to be from a bank asking for login details', secures full marks.
很多“描述”和“解释”类题目明确要求举例。如果一道题询问网络安全威胁，指出“一个威胁是网络钓鱼”，然后给出简单例子，如“一封假装来自银行索取登录信息的电子邮件”，便能拿全分数。
When explaining protocols like HTTP and HTTPS, you could write: 'HTTPS encrypts data using SSL/TLS; for example, when you enter credit card details on a shopping site, the data is scrambled so eavesdroppers cannot read it.' Your example must clearly illustrate the concept.
在解释 HTTP 和 HTTPS 等协议时，你可以写：“HTTPS 使用 SSL/TLS 加密数据；例如，当你在购物网站输入信用卡信息时，数据被加密，窃听者无法读取。”你的例子必须清楚说明概念。
Non-exampled answers often lose marks because the mark scheme explicitly allocates a mark for 'suitable example given'. Train yourself to read 'Give an example' in the question and respond with a real-world or programming instance.
没有例子的答案经常丢分，因为评分标准特地为“给出合适的例子”分配了分值。要训练自己识别题目中“举一个例子”的提示，并回答一个现实世界或编程中的实例。
9. Common Mistakes to Avoid | 常见错误避免
One of the most frequent mistakes is not reading the full question. Many students provide a brilliant answer to a related but unasked question. Highlight key terms and the command word to stay focused.
最常见的错误之一是没有读完整道题目。很多学生给出了一个与题目相关但并非所问的优秀答案。要高亮标注关键词和指令词，保持答题焦点。
In pseudocode, forgetting to initialise a counter or sum variable leads to runtime logic errors and lost marks. Always set variables to 0 or '' before a loop. Also, ensure final OUTPUT is inside the code snippet; marks are often reserved for the output statement.
在伪代码中，忘记初始化计数器或求和变量会导致运行时逻辑错误并失分。始终在循环前将变量设为0或''。同时确保最终的 OUTPUT 语句在代码片段内；输出语句常常有独立赋分。
For trace tables, jumping directly to the final value without showing intermediate steps often results in half marks at best. Take the time to list every iteration. In theory answers, using vague terms like 'it’s better' instead of 'it has a higher data transfer rate' loses marks for precision.
在跟踪表中，直接写最终值而不展示中间步骤往往最多得一半分数。要花时间列出每次迭代。在理论题答案中，使用模糊说法如“它更好”而不是“它具有更高的数据传输率”会因不够精确而失分。
10. Exam Paper Structure and Time Management | 试卷结构与时间管理
Paper 1 (Theory) lasts 1 hour 45 minutes and is worth 75 marks. Paper 2 (Problem-solving and Programming) also lasts 1 hour 45 minutes for 50 marks. You should allocate roughly 1.4 minutes per mark in Paper 1 and 2.1 minutes per mark in Paper 2, since Paper 2 questions often involve coding.
试卷一（理论）时长1小时45分钟，满分75分。试卷二（问题解决与编程）同样时长1小时45分钟，满分50分。你应当大致按试卷一每分1.4分钟、试卷二每分2.1分钟来分配时间，因为试卷二的题目常涉及编码。
Start with questions you find easiest to secure quick marks. Leave the more challenging 6-mark algorithm or extended response questions until you have attempted all compulsory ones. Keep an eye on the clock and write down your progress checkpoints, e.g., '30 minutes left, should be at question 5'.
从你觉得最简单的题目入手，快速拿到分数。把较难的6分算法题或扩展作答题目留到完成所有必答题之后。留意时间，可以写下进度检查点，例如“剩余30分钟，应该做到第5题”。
In Paper 2, avoid spending 20 minutes debugging one syntax error. Use the comment feature or write a note to the examiner if you cannot fix a minor bug, because method marks can still be awarded. Never leave a question completely blank; even a partial attempt may earn 1 or 2 marks.
在试卷二中，避免花20分钟去调试一个小小的语法错误。如果无法修正一个小 bug，可以用注释或给阅卷人写一张便条说明，因为方法分仍然可以得到。千万不要把题目完全空着，哪怕是部分尝试也可能拿到1到2分。
11. Leveraging the Mark Scheme for Revision | 利用评分标准进行复习
The official CAIE mark schemes are public and reveal exactly what examiners expect. After practising a past paper, compare your answer to the mark scheme line by line. Notice recurring patterns: 'any two from' means you need only a couple of points; 'max three marks' means do not overload.
CAIE 官方评分标准是公开的，精确揭示了阅卷人的期望。做完一套真题后，逐行对比你的答案与评分标准。留意常出现的模式：“任选其二”意味着你只需写两个点；“最多三分”意味着不要写太多。
Many topics such as data transmission, logic gates, and high/low-level languages follow predictable mark allocations. Once you internalise those, you can tailor your answers to match them. This is one of the most effective revision strategies for Year 11 students.
许多主题，如数据传输、逻辑门和高级/低级语言的评分点分配是可以预测的。一旦你内化了这些，便可以量身定制自己的答案。对 Year 11 学生来说，这是最有效的复习策略之一。
12. Final Preparations and Exam-Day Tips | 考前准备与考试日贴士
In the last days before the exam, do timed past papers under realistic conditions. Practise writing pseudocode on paper without an IDE to get comfortable with hand-writing code. Review common algorithms like linear search, bubble sort, and reading/writing to files.
考前最后几天，模拟真实环境限时做真题。练习在纸上手写伪代码，不依赖开发环境，让自己习惯手写代码。回顾线性搜索、冒泡排序和读写文件等常见算法。
Get a good night's sleep before the exam. During reading time, scan through the whole paper and note the command words. Bring a black pen and a clear ruler for any tables. Stay calm, trust your preparation, and remember that every mark counts.
考试前一晚保证良好睡眠。阅读时间里，快速浏览整份试卷，标出指令词。带上黑色笔和清晰尺子以备画表格。保持冷静，相信自己的准备，记住每一分都重要。
Published by TutorHao | Computer Science Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图