📚 Year 12 Cambridge Computer Science: High-Frequency Topics and Common Pitfalls | Year 12 剑桥计算机：高频考点与易错题分析
In the Cambridge AS Computer Science syllabus (9618), students encounter a broad range of theoretical and practical challenges. This article identifies the topics that frequently appear in examinations and highlights the common errors that students make. Mastering these areas will not only boost exam confidence but also build a solid foundation for A2 studies.
在剑桥 AS 计算机科学课程（9618）中，学生面临着广泛的理论与实践挑战。本文梳理了考试中反复出现的高频考点，并剖析了学生最常犯的错误。掌握这些内容，既能提升应试信心，也为 A2 阶段打下坚实基础。
1. Data Representation | 数据表示
Data representation covers binary, hexadecimal, two’s complement, floating-point numbers, character sets, sound sampling, and bitmap images. These subtopics appear almost every session. Many students memorise conversion methods but overlook the underlying concepts such as overflow, sampling resolution, and the impact of bit depth on file size and quality.
数据表示涵盖二进制、十六进制、补码、浮点数、字符集、声音采样以及位图图像。这些子主题几乎每次考试都会出现。许多学生记住了转换方法，却忽略了基本概念，例如溢出、采样分辨率以及位深度对文件大小和质量的影响。
- Binary addition often causes carry errors; candidates forget that an overflow flag is set when the result exceeds the available bits.
- 二进制加法常产生进位错误；考生常常忘记当结果超出可用位数时会设置溢出标志。
- Misinterpreting normalised floating-point representation leads to incorrect mantissa-exponent pairs. Always check the implied leading bit.
- 误解规范化浮点表示会导致尾数和指数对错误。务必检查隐含的前导位。
- ze: File size = sample rate × sample resolution × duration (seconds) × number of channels.
- 在声音表示中，许多人混淆采样率与采样精度，无法正确计算文件大小：文件大小 = 采样率 × 采样精度 × 时长（秒） × 声道数。
- For images, the relationship between colour depth and the number of colours (2ⁿ) is often reversed. A common mistake is using 256 for 8-bit colour depth but failing to recognise that the index refers to a palette entry, not the actual colour value.
- 对于图像，颜色深度与颜色数量 (2ⁿ) 的关系常被颠倒。常见错误是知道 8 位色深对应 256 色，却不理解索引指向调色板条目，而非实际颜色值。
2. Communication and Internet Technologies | 通信与互联网技术
This section focuses on the TCP/IP protocol stack, packet switching, client-server model, and web technologies. High-frequency exam questions ask students to compare circuit switching with packet switching, and to explain the role of each layer in TCP/IP.
本节聚焦于 TCP/IP 协议栈、分组交换、客户端-服务器模型和 Web 技术。高频考题要求学生比较电路交换与分组交换，并解释 TCP/IP 各层的作用。
- A typical pitfall: stating that ‘packet switching is faster’ without qualification. Packet switching reduces latency for bursty data but can introduce jitter and delays due to routing overhead.
- 典型陷阱：不加限定地声称“分组交换更快”。分组交换对于突发数据可降低延迟，但由于路由开销，可能引入抖动和延迟。
- Confusing the transport layer with the network layer. The transport layer (TCP) ensures reliable delivery; the network layer (IP) handles addressing and routing.
- 混淆传输层与网络层。传输层 (TCP) 确保可靠交付；网络层 (IP) 负责寻址与路由。
- Many candidates cannot correctly describe the role of DNS. It translates domain names to IP addresses, but also caches results; failing to mention caching loses marks.
- 许多考生不能正确描述 DNS 的作用。它将域名转换为 IP 地址，同时还会缓存结果；遗漏缓存这一点会失分。
- When drawing packet-switched diagrams, ensure hop-by-hop transmission is shown, not a single direct path. Each packet may take different routes.
- 绘制分组交换示意图时，确保画出逐跳传输，而非单一直接路径。每个分组可能走不同路径。
3. Hardware | 硬件
Input, output, memory, storage, and logic gates form the core of hardware questions. Common exam scenarios ask students to justify a choice of device for a given context, or to analyse simple logic circuits and derive truth tables.
输入、输出、内存、存储器以及逻辑门构成了硬件问题的核心。常见考题场景要求学生为给定情境选择设备并说明理由，或分析简单逻辑电路并推导真值表。
- Mixing up RAM and ROM in embedded systems. A microcontroller’s firmware is stored in ROM, while temporary data resides in RAM. Students often assign the opposite roles.
- 在嵌入式系统中混淆 RAM 与 ROM。微控制器的固件存储在 ROM 中，而临时数据驻留在 RAM 里。学生常将两者的角色颠倒。
- For logic gates, a common error is misreading the symbols for NAND and NOR, or incorrectly cascading outputs. Always build an intermediate column in truth tables.
- 对于逻辑门，常见错误是误读 NAND 与 NOR 的符号，或级联输出错误。始终在真值表中构建中间列。
- Storage device questions often require comparison of durability, speed, portability, and cost per GB. Omitting a quantitative reasoning loses marks.
- 存储器设备问题通常要求比较耐久性、速度、便携性和每 GB 成本。缺少量化推理会导致失分。
- SSD vs HDD: state that SSDs have no moving parts (faster access, lower power), but do not forget that HDDs are still cheaper per TB.
- SSD 与 HDD：应指出 SSD 无活动部件（访问更快、功耗更低），但别忘了 HDD 每 TB 成本仍然更低。
4. Processor Fundamentals | 处理器基础
The fetch-decode-execute cycle, CPU registers, assembly language, and addressing modes are repeated topics. Students must be able to trace through cycles and describe the role of each special register.
取指-解码-执行周期、CPU 寄存器、汇编语言以及寻址模式是反复出现的主题。学生必须能够跟踪周期并描述每个特殊寄存器的作用。
- Confusing the program counter (PC) with the memory address register (MAR). The PC holds the address of the next instruction; the MAR holds the address for the current memory access.
- 混淆程序计数器 (PC) 与存储器地址寄存器 (MAR)。PC 存放下一条指令的地址；MAR 存放当前内存访问的地址。
- In the decode stage, the Control Unit interprets the opcode; many think decoding happens in the ALU – it does not.
- 在解码阶段，控制单元解释操作码；许多人认为解码发生在 ALU 中——并没有。
- Assembly language mnemonics (e.g., LDR, STR, ADD) and immediate/direct/indexed addressing models are common. Forgetting that index registers add a displacement to a base address is a typical mistake.
- 汇编语言助记符（如 LDR, STR, ADD）以及立即/直接/索引寻址模式是常见考点。忘记索引寄存器会将偏移量加到基址上是一个典型错误。
- When tracing machine code execution, always update the PC after fetch, not during decode. Incorrect ordering of register transfers causes cascading errors.
- 在跟踪机器码执行时，务必在取指后更新 PC，而不是在解码期间。寄存器传输顺序不正确会导致连锁错误。
5. System Software | 系统软件
This topic includes operating systems, utility software, library programs, and language translators. High-frequency questions centre on the functions of an OS, the difference between compiler/interpreter, and the use of assemblers.
本主题包括操作系统、实用程序、库程序以及语言翻译器。高频问题聚焦于操作系统的功能、编译器与解释器的差异，以及汇编器的使用。
- Describing memory management: ‘allocates memory to processes’ is insufficient; add ‘paging, segmentation, virtual memory’ and how it prevents one process from accessing another’s space.
- 描述内存管理：“为进程分配内存”是不够的；应补充“分页、分段、虚拟内存”以及如何防止一个进程访问另一个进程的空间。
- Compiler vs. interpreter: an interpreter translates and executes line-by-line, a compiler produces an executable. Students often forget that interpreted code runs slower but is easier to debug.
- 编译器与解释器：解释器逐行翻译并执行，编译器生成可执行文件。学生常忘记解释型代码运行较慢但更易于调试。
- Utility software (disk defragmenter, backup, encryption, compression) is tested via scenario-based questions. For disk defragmentation, explicitly mention it rearranges fragmented files to contiguous blocks for faster access.
- 实用程序（磁盘碎片整理、备份、加密、压缩）通过情景题考查。对于磁盘碎片整理，要明确提及它将碎片文件重新排列为连续块以加快访问。
- Linkers and loaders are often confused. A linker combines object modules; a loader places the executable in memory ready to run.
- 链接器与加载器常被混淆。链接器将目标模块组合起来；加载器将可执行文件放入内存准备运行。
6. Security, Privacy and Data Integrity | 安全、隐私与数据完整性
Security topics cover encryption (symmetric vs asymmetric), malware, authentication, firewalls, and data validation. Verifiable examples gain marks: describe AES for symmetric, RSA for asymmetric.
安全主题涵盖加密（对称与非对称）、恶意软件、身份验证、防火墙以及数据验证。可验证的实例能得分：描述 AES 用于对称加密，RSA 用于非对称加密。
- Misunderstanding public key cryptography: the public key encrypts, the private key decrypts – not the reverse. Signing uses private key for encryption of the hash.
- 误解公钥加密：公钥加密，私钥解密——而不是反过来。数字签名使用私钥对哈希值进行加密。
- Malware categories (virus, worm, Trojan, spyware) need distinct descriptions: a worm self-replicates across a network, a Trojan disguises as legitimate software.
- 恶意软件类别（病毒、蠕虫、木马、间谍软件）需要明确区分：蠕虫在网络中自我复制，木马伪装成合法软件。
- Data integrity vs. data validity: validity checks (range, format, presence) prevent entry errors; integrity (hashing, parity) ensures data remains uncorrupted. Students frequently equate them.
- 数据完整性与数据有效性：有效性检查（范围、格式、存在性）防止输入错误；完整性（哈希、奇偶校验）确保数据保持未损坏状态。学生经常将二者等同。
7. Ethics and Ownership | 道德与所有权
Ethical questions probe copyright, software licensing (open source vs proprietary), artificial intelligence, and the digital divide. This section is often undervalued by candidates who provide shallow answers.
道德问题涉及版权、软件许可（开源与专有）、人工智能以及数字鸿沟。这部分常被低估，考生往往给出浅显的回答。
- Open source does not mean ‘free of cost’; it means source code is publicly accessible and can be modified/distributed under a license. Must mention the four freedoms.
- 开源并不意味“免费”；它指源代码可公开访问，并可根据许可证进行修改/分发。必须提及四大自由。
- AI ethical concerns: bias in training data, job displacement, accountability. A common weakness is repeating “it replaces humans” without elaborating on specific risks.
- 人工智能伦理问题：训练数据中的偏见、工作岗位流失、责任归属。常见的弱点是反复说“它取代人类”，却不详述具体风险。
- Digital divide answers must reference both infrastructure (access to devices/internet) and skills (education). Mentioning socio-economic factors earns extra marks.
- 数字鸿沟的答案必须同时提及基础设施（设备/互联网接入）和技能（教育）。提到社会经济因素可获得额外加分。
8. Databases | 数据库
Flat-file vs. relational databases, entity-relationship diagrams, normalisation up to third normal form (3NF), and SQL are high-weight topics. Structured answers are essential.
平面文件与关系数据库、实体关系图、规范化的第三范式 (3NF) 以及 SQL 是高分值话题。结构化的回答至关重要。
- In flat-file systems, data redundancy and update anomalies are common. Students can state the problem but rarely illustrate with a concrete example (e.g., changing a customer address in multiple records).
- 在平面文件系统中，数据冗余和更新异常常见。学生能说出问题，但很少用具体实例说明（如，在多条记录中更改客户地址）。
- Normalisation errors: partial dependency exists when a non-key attribute depends on part of a composite key. Moving to 2NF must remove these. Many misidentify transitive dependency.
- 规范化错误：当非键属性依赖于组合键的一部分时，存在部分依赖。向 2NF 转换必须移除这些依赖。许多人会错误识别传递依赖。
- SQL syntax: forgetting the semicolon, mixing DDL and DML, writing JOIN conditions incorrectly. Most lost marks come from omitting the ON clause in JOIN.
- SQL 语法：忘记分号、混淆 DDL 和 DML、错误编写 JOIN 条件。大多数失分来自 JOIN 中省略 ON 子句。
9. Algorithm Design and Problem Solving | 算法设计与问题解决
Abstract data types (stack, queue, linked list, binary tree), recursion, searching and sorting algorithms form a large part of Paper 2. Pseudocode clarity and analysis of complexity are required.
抽象数据类型（栈、队列、链表、二叉树）、递归、搜索和排序算法构成了 Paper 2 的很大一部分。需要清晰的伪代码以及复杂度分析。
- Stack operations (push/pop) and the LIFO principle are straightforward, but applying them to expression evaluation (e.g., postfix) causes errors when popping order is wrong.
- 栈操作（push/pop）和 LIFO 原则较为直接，但应用于表达式求值（如后缀表达式）时，若弹出顺序错误就容易出错。
- Recursion requires a base case and a recursive case that converges. A common pitfall is infinite recursion due to missing base case or incorrect decrement.
- 递归需要基例以及能够收敛的递归情况。常见陷阱是因缺少基例或递减错误导致的无限递归。
- Binary search: many miss that the array must be sorted first. For large data, linear search may be more appropriate if data is unsorted or frequently updated.
- 二分查找：很多人忽略数组必须先排序。如果数据未排序或频繁更新，线性搜索可能更合适。
- Sorting algorithms: bubble sort O(n²) stable, merge sort O(n log n) but requires extra space. Confusing stability with performance is a typical exam trap.
- 排序算法：冒泡排序 O(n²) 且稳定，归并排序 O(n log n) 但需要额外空间。将稳定性与性能混淆是典型的考试陷阱。
10. Programming Concepts | 编程概念
Variables, data types, control structures, procedures and functions, parameter passing (by value/by reference), and file handling are tested both in pseudocode and in practical programming tasks. Cambridge expects precise understanding of scope and parameter mechanisms.
变量、数据类型、控制结构、过程与函数、参数传递（按值/按引用）以及文件处理会在伪代码和实际编程任务中考查。剑桥要求精确理解作用域和参数传递机制。
- Pass-by-value vs pass-by-reference: a classic error is assuming that a variable passed by reference still creates a local copy. Modifications affect the original.
- 按值传递与按引用传递：经典错误是认为按引用传递的变量仍会创建局部副本。实际上修改会影响原始变量。
- Variable scope: a local variable inside a function is not accessible outside; candidates often try to use the same identifier globally, causing confusion.
- 变量作用域：函数内部的局部变量在外部不可访问；考生常试图在全局使用相同的标识符，从而导致混乱。
- File handling: opening a file for reading when writing is intended leads to runtime errors. Always specify modes (READ, WRITE, APPEND) and close file handles.
- 文件处理：当意图写入时却以读取模式打开文件，会导致运行时错误。务必指定模式（READ, WRITE, APPEND）并关闭文件句柄。
- In pseudocode, ensure consistent indentation, proper use of loop structures, and that variables are declared before use. Careless syntax loses marks even if logic is correct.
- 在伪代码中，确保缩进一致、正确使用循环结构，并在使用前声明变量。语法粗心即使逻辑正确也会失分。
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