📚 Cambridge IGCSE Computer Science: Exam Techniques and Marking Criteria | 剑桥IGCSE计算机科学：答题技巧与评分标准
Scoring top marks in Cambridge IGCSE Computer Science is not just about knowing the theory; it is about understanding how to present your knowledge in a way that ticks every box on the examiner's mark scheme. This guide breaks down essential exam techniques, command word interpretations, and subject-specific strategies that will help you avoid common pitfalls and convert your understanding into marks. Whether you are tackling Paper 1 theory or Paper 2 problem-solving and programming, the following advice is designed to sharpen your exam performance.
在剑桥 IGCSE 计算机科学中拿到高分，不仅在于掌握理论知识，更在于懂得如何以符合评分标准的方式呈现你的答案。本指南拆解了关键的答题技巧、指令词解读以及各主题的策略，帮助你避开常见错误，将知识转化为分数。无论你面对的是试卷一的理论题，还是试卷二的问题解决与编程题，以下建议都旨在提升你的应试表现。
1. Understanding the Mark Scheme | 理解评分标准
Every question is matched to a specific set of marking points, and examiners award marks solely based on the presence of these points. Reading past mark schemes reveals patterns: for instance, a two-mark 'describe' question often expects one mark for identification and one for a linked expansion. Familiarise yourself with how marks are allocated for common question types, and always check if the answer space (number of lines) hints at the amount of detail required.
每道题目都有一套明确的得分点，考官仅根据是否出现这些得分点来给分。阅读历年评分方案可以发现规律：例如，一道两分的“描述”题通常会要求一个识别点和一个相关的扩展点。要熟悉常见题型中分数是如何分配的，并始终留意答题空间（行数）是否暗示了所需的详细程度。
Many candidates lose marks by writing vague statements. The mark scheme often lists acceptable alternatives separated by a slash, e.g., 'storage/media' or 'volatile/temporary'. Train yourself to use the same precise technical vocabulary found in the syllabus, because synonyms are not always rewarded. When reviewing your answers, underline the key terms from the mark scheme to see if you have actually included them.
许多考生因表述模糊而丢分。评分方案常以斜线分隔可接受的替代答案，如“存储/介质”或“易失性/临时性”。要训练自己使用大纲中出现的精确技术词汇，因为同义词并不总能得分。检查答案时，划出评分方案中的关键词，确认你是否确实写入了这些词。
2. Command Words in IGCSE CS | IGCSE 计算机科学指令词
Understanding command words is the foundation of targeted answering. 'State' requires a short, factual piece of information with no explanation. 'Describe' asks for characteristics or a sequence of events – you should paint a picture in words. 'Explain' demands a reason or mechanism, often using 'because' or 'so that', linking a cause to an effect. Using these distinctions will prevent you from overwriting where it is not needed or under‑explaining where marks are available.
理解指令词是精准答题的基础。“State”（陈述）要求给出简短的事实信息，无需解释。“Describe”（描述）需要陈述特征或事件顺序——你要用文字描绘出来。“Explain”（解释）则要求给出原因或机制，常使用“因为”或“为了”，将原因与结果联系起来。掌握这些区别可以防止你在不需要时过度写作，或者在可以得分的地方解释不足。
| Command Word | What It Means | 指令词 | 含义 |
|---|---|---|---|
| State | Give a fact, name, or value. | 陈述 | 给出事实、名称或数值。 |
| Describe | Say what something is like or what happens. | 描述 | 说明某物是什么样的或发生了什么。 |
| Explain | Give reasons or purposes; cause and effect. | 解释 | 给出原因或目的；因果关系。 |
| Identify | Name or pick out from given information. | 识别 | 从给定信息中找出或命名。 |
| Justify | Support a choice with evidence/reasons. | 论证 | 用证据/理由支持一个选择。 |
When a question asks you to 'Give two reasons why…', write two distinct, fully expressed reasons, not one reason phrased in two ways. Even if your first reason is excellent, you will only score one mark if the second does not add a genuinely new point. Practise extracting the command word from past papers and writing a few bullet points that mirror the expected answer structure.
当题目要求“给出……的两个理由”时，写出两个截然不同、表述完整的理由，而不是将同一个理由用两种方式表达。即使第一个理由非常出色，如果第二个理由没有提出真正新的观点，你也只能得到一分。练习从历年试卷中提取指令词，并写出几点与预期答案结构相符的要点。
3. Tackling Algorithm and Pseudocode Questions | 应对算法与伪代码题
Programming questions are marked for logic, correct use of structures, and accuracy. Present your pseudocode clearly: use indentation for loops and conditions, and stick to Cambridge-approved syntax such as INPUT, OUTPUT, WHILE...DO...ENDWHILE, IF...THEN...ELSE...ENDIF. Even if your syntax is not flawless, a correct algorithm design can earn significant method marks.
编程题的评分依据是逻辑、结构使用的正确性以及准确性。清晰地呈现你的伪代码：循环和条件语句使用缩进，并坚持剑桥认可的语法，如 INPUT、OUTPUT、WHILE...DO...ENDWHILE、IF...THEN...ELSE...ENDIF。即使语法并非完美，正确的算法设计也能获得可观的步骤分。
Before writing any code, read the scenario thoroughly and identify the inputs, processes, and outputs expected. Jot down a quick plan or flowchart. In the exam, annotate your pseudocode with brief comments only if it helps clarify your thinking – but don't waste time on excessive commentary. Ensure variable names are meaningful and consistent; a mark may be lost if you suddenly change a variable from Total to Sum without reason.
在编写任何代码之前，仔细阅读题目情境，明确所需的输入、处理和输出。快速写下计划或流程图。考试时，只有在有助于理清思路时才为伪代码添加简短注释——但不要在过多的注释上浪费时间。确保变量名称有意义且前后一致；如果无缘由地将变量从 Total 突然改为 Sum，可能会丢分。
Exam questions often ask you to complete partially written algorithms. Pay careful attention to the variables already declared and the loop conditions provided. When you fill in the blanks, test each change mentally by running a sample data input through the logic. This 'dry run' approach prevents off‑by‑one errors and logic flaws that are heavily penalised.
考题经常要求补全部分给出的算法。仔细留意已声明的变量和已提供的循环条件。填空时，用一组示例数据在脑海中逐行运行逻辑进行测试。这种“演练”方法可以防止被严重扣分的边界错误和逻辑缺陷。
4. Mastering Trace Tables | 掌握追踪表
Trace table questions require you to step through an algorithm line by line, recording variable values at each stage. Start by drawing the table with a column for each variable and condition. Use a new row every time any variable changes. Be systematic: even if the algorithm loops many times, show all iterations unless the question specifies a limit – a completed table proves your understanding and secures full marks.
追踪表题目要求你逐行执行算法，并记录每一步变量的值。首先绘制表格，为每个变量和条件各设一列。每当有任何变量发生变化，就使用新的一行。要有条理：即使算法循环多次，除非题目明确限制，否则要展示所有迭代——一个完整的追踪表能证明你的理解并确保满分。
A common pitfall is forgetting to update the value of a condition flag like flag ← TRUE after a check. Double-check the order of operations: does the algorithm check the condition before or after updating the variable? In WHILE loops, the condition is evaluated at the start of each iteration, so ensure your trace reflects this precisely. Marks are often lost by putting values in the wrong row or skipping a step.
一个常见的陷阱是，在执行检查后忘记更新像 flag ← TRUE 这样的条件标志。仔细核对操作顺序：算法是在更新变量之前还是之后检查条件？在 WHILE 循环中，每次迭代开始时会先判断条件，所以要确保追踪表准确反映这一点。把数值放错行或跳过一个步骤往往会丢分。
After completing a trace, review the final row. Does the output match what you would expect logically? If asked a follow‑up question about the purpose of the algorithm, use the trace results to describe what the algorithm accomplishes, e.g., 'It finds the largest number in the list.' This demonstrates deeper understanding and can gain an extra explanation mark.
完成追踪表后，检查最后一行。输出结果是否符合逻辑预期？如果后续题目询问算法的目的，用追踪结果描述算法实现了什么，例如“它找到了列表中的最大值”。这可以展示更深入的理解，并可能获得额外的解释分。
5. Logic Gates and Boolean Expressions | 逻辑门与布尔表达式
For logic gate questions, you must be able to complete truth tables and draw circuits from Boolean expressions. Use standard symbols for AND, OR, NOT, NAND, NOR, and XOR. When building a truth table with two inputs A and B, write all four combinations (00, 01, 10, 11) systematically and then compute intermediate columns step by step. This method reduces errors in final output columns.
对于逻辑门题目，你必须能够完成真值表，并根据布尔表达式画出电路。使用标准的 AND、OR、NOT、NAND、NOR 和 XOR 符号。当构建有两个输入 A 和 B 的真值表时，系统地写出全部四种组合（00, 01, 10, 11），然后一步一步计算中间的列。这种方法可以降低最终输出列的错误率。
Remember that a Boolean expression such as (A AND B) OR (NOT A) can be simplified, but you do not need to simplify unless asked. If you are asked to design a circuit with the fewest gates, apply Boolean algebra identities or use a Karnaugh map to minimise the expression first. Always label the output of each gate clearly, and double-check that the number of gates matches your simplified expression.
请记住，像 (A AND B) OR (NOT A) 这样的布尔表达式可以化简，但除非题目要求，否则你不需要这样做。如果要求用最少的门电路设计，应首先应用布尔代数恒等式或使用卡诺图来化简表达式。始终清晰地标注每个门的输出，并再次核对门的数量是否与化简后的表达式相符。
When a question provides a circuit and asks for the Boolean expression, work backward from the output. Write the expression for the final gate's inputs, then substitute the sub‑expressions of preceding gates. Use parentheses to preserve order. A well‑structured expression is easier to evaluate and leaves fewer chances for transcription mistakes.
当题目给出电路并要求写出布尔表达式时，从输出端反向推导。先写出最后一级门输入端的表达式，然后代入前级门的子表达式。使用括号保持运算顺序。结构良好的表达式更容易求值，也减少了抄写错误的机会。
6. Data Representation and Number Conversions | 数据表示与数制转换
8, 4, 2, 1) above each bit; this simple visual aid prevents place value mistakes. For binary to hexadecimal, group bits into nibbles (4‑ se that each nibble corresponds to exactly one hex digit: 1010₂ = A, 1011₂ = B, and so on.
二进制、十进制和十六进制之间的转换是必考内容。二进制转十进制时，在每位上方写下位权（……128, 64, 32, 16, 8, 4, 2, 1）；这个简单的视觉辅助可以防止位值错误。二进制转十六进制时，从右向左将位分成半字节（4 位一组）。记住每个半字节恰好对应一个十六进制数位：1010₂ = A，1011₂ = B，以此类推。
ow your carries clearly; marks are often awarded for correct carries even if the final sum is wrong due to a minor slip. In overflow questions, explain that overflow occurs when the result exceeds the allocated bit width, leading to an incorrect value being stored.
对二进制数做加法时，应用与十进制相同的进位原则：1+1 = 0 进 1，1+1+进位 = 1 进 1。清楚地写出进位；即使因小失误导致最终和出错，正确的进位通常也能得分。在溢出题中，解释溢出发生在结果超出分配的位宽时，导致存储的值不正确。
e file sizes for images: width × height × colour depth (in bits), then convert to bytes by dividing by 8. Use the exact formula and show your workings; an incorrect unit conversion (bits instead of bytes) will lose the answer mark even if the method is perfect.
在文本表示方面，复习 ASCII 和字符集的用途。要准备好解释为什么需要 Unicode 来支持多语言文本。计算图像文件大小：宽度 × 高度 × 色深（以位为单位），然后除以 8 转换为字节。使用准确的公式并展示计算过程；即使方法完美，单位换算错误（位写成字节）也会丢失答案分。
7. Networking and Hardware Recall | 网络与硬件记忆题
Theory questions on networks and hardware often ask you to 'describe' the role of components such as routers, switches, NICs, or to 'explain' the difference between LAN and WAN. Build your answers around the key functions. For a router, mention forwarding data packets between networks, using IP addresses, and maintaining routing tables. One well‑explained function earns more marks than a list of four vague terms.
关于网络和硬件的理论题经常要求你“描述”路由器、交换机、网卡等组件的作用，或者“解释”LAN 与 WAN 的区别。围绕关键功能组织答案。对于路由器，要提到在网络之间转发数据包、使用 IP 地址以及维护路由表。清晰解释一个功能比列举四个模糊术语所得的分数更高。
Use comparisons to structure higher‑mark answers. For example, when comparing packet switching and circuit switching, contrast dedicated paths with dynamic routing, setup time, and efficiency under different traffic conditions. Draw a small labelled sketch if it helps you visualise, but only if the question permits – diagrams can boost a description but should be annotated with text to be eligible for marks.
用对比来组织高分答案。例如，在比较分组交换和电路交换时，从专用路径与动态路由、建立时间以及不同流量条件下的效率等方面进行对比。如果有助于可视化，可以画一个带标注的小草图——但仅限于题目允许；图表能强化描述，但必须配有文字注解才能得分。
Hot topics like cybersecurity threats (phishing, malware, DoS) and prevention (firewalls, encryption, two‑factor authentication) appear frequently. Pair each threat with a specific countermeasure and explain how it mitigates the risk. Never just write 'use a firewall' – state that a firewall monitors incoming and outgoing traffic and blocks unauthorised access based on predefined rules, directly reducing the chance of intrusion.
像网络安全威胁（钓鱼、恶意软件、拒绝服务攻击）和防护措施（防火墙、加密、双因素认证）这类热门主题频繁出现。将每个威胁与一项具体的应对措施配对，并解释它如何减轻风险。永远不要只写“使用防火墙”——要说明防火墙监控进出流量，并根据预设规则阻止未经授权的访问，从而直接降低入侵的可能性。
8. SQL and Database Queries | SQL 与数据库查询
SQL question marks are easily lost by forgetting the basic structure: SELECT field(s) FROM table WHERE condition ORDER BY field ASC/DESC. Write keywords in capitals to make your answer stand out to the examiner. When a question asks for 'all fields', use SELECT *. Ensure the WHERE clause matches the exact requirement – if the scenario says 'students older than 16', your condition should be WHERE Age > 16, not Age >= 16.
SQL 题的分数常常因遗忘基本结构而丢失：SELECT 字段 FROM 表 WHERE 条件 ORDER BY 字段 ASC/DESC。使用大写关键字，让考官一目了然。当题目要求“所有字段”时，使用 SELECT *。确保 WHERE 子句完全符合题目要求——如果情境是“年龄大于 16 岁的学生”，你的条件应为 WHERE Age > 16，而不是 Age >= 16。
Complex queries may involve multiple conditions joined by AND/OR, or sorting by more than one column. When ORDER BY has multiple columns, list them separated by commas, e.g., ORDER BY Surname, FirstName. If a query requires a calculation, such as finding the difference between two fields, perform it in the SELECT clause with the AS keyword to give the result column a meaningful alias. This detail pleases examiners and shows correctness.
复杂查询可能涉及用 AND/OR 连接的多个条件，或按多列排序。当 ORDER BY 使用多列时，用逗号分隔，例如 ORDER BY Surname, FirstName。如果查询需要进行计算，如求两个字段的差值，可在 SELECT 子句中使用 AS 关键字为结果列赋予一个有意义的别名。这个细节能让考官满意，并展示出正确性。
Practice drawing simple table relationships for the database design questions. Identify primary keys (often IDs) and foreign keys. When given a scenario, normalise data by splitting repeating groups into separate tables. This reduces redundancy. Explain your design choices using the correct terminology: primary key, foreign key, composite key. Such vocabulary consistently appears in high‑level answers on mark schemes.
在数据库设计题中，练习绘制简单的表关系。识别主键（通常是 ID）和外键。给定一个情境，通过将重复组分拆分成独立的表来将数据规范化，以减少冗余。用正确的术语解释你的设计选择：主键、外键、复合键。这类词汇在评分方案的高分答案中始终出现。
9. Common Mistakes and How to Avoid Them | 常见错误及避免方法
One recurring error is misreading the unit of measurement. When asked for storage in kilobytes, candidates frequently give the answer in bytes or bits. Circle the required unit in the question. Another trap is confusing left and right shifts for binary multiplication/division: a left shift multiplies by 2, a right shift divides by 2, but always check for lost bits that cause inaccurate results.
一个反复出现的错误是看错计量单位。当题目要求以千字节为单位给出存储容量时，考生常常给出以字节或位为单位的答案。在题目中将要求的单位圈出来。另一个陷阱是混淆用于二进制乘法/除法的左移和右移：左移乘以 2，右移除以 2，但要始终检查丢失的位可能导致结果不准确。
In programming and pseudocode, many marks are dropped by using assignment (= or ←) when a comparison (= or ==) is needed, or vice versa. In Cambridge pseudocode, comparison is often = a condition is required, do not accidentally assign a value. Mentally test your condition: 'if x = 5' should check equality, not set x to 5.
在编程和伪代码中，许多分数因在需要比较（= 或 ==）的地方错误地使用赋值（= 或 ←）而丢失，反之亦然。在剑桥伪代码中，比较常用 = 表示相等，但要遵从试卷中使用的风格。当需要条件判断时，不要意外地赋予一个值。在心里测试你的条件：“if x = 5” 应是检查相等，而不是将 x 设为 5。
Neglecting to validate input is another oversight. When an algorithm expects a number within a range, include a validation loop. For example, WHILE Number < 1 OR Number > 100 DO OUTPUT "Invalid, try again" INPUT Number ENDWHILE. Showing this demonstrates a thorough approach and can earn extra robustness marks.
忽视输入验证是另一个疏漏。当算法期望一个范围内的数字时，要包含一个验证循环。例如，WHILE Number < 1 OR Number > 100 DO OUTPUT "Invalid, try again" INPUT Number ENDWHILE。展示这一点体现了周密的思路，并能获得额外的健壮性分数。
10. Exam Day Strategy and Time Management | 考试日策略与时间管理
Allocate time proportionally to the marks available. In Paper 2, which involves problem‑solving, spend the first few minutes scanning all questions and identifying the ones that play to your strengths. Attempt the highest‑mark questions first while your mind is fresh, but never over‑allocate time – if you are stuck on a 6‑mark trace table after 10 minutes, move on and return to it.
按题目的分数比例分配时间。在包含问题解决的试卷二中，用头几分钟浏览所有题目，找出自己擅长的部分。在头脑清醒时优先尝试分值最高的题目，但绝不要超时——如果一道 6 分的追踪表题 10 分钟后仍无进展，跳过它，回头再处理。
In long‑answer theory questions, plan before you write. Use
Published by TutorHao | Year 11 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