📚 Year 12 CIE Computer Science: Core Topics in a Nutshell | 12年级 CIE 计算机科学：核心知识点梳理
Mastering the AS Level Computer Science syllabus (9618) requires a solid grip on a wide range of fundamentals – from binary representations to network protocols, from processor architecture to algorithm design. This article distils the core topics for Year 12 learners, presenting each key idea in a clear, exam-focused way. Whether you need a quick revision guide or a structured overview to spot knowledge gaps, the following sections will help you build confidence and precision.
掌握AS阶段计算机科学(9618)课程，需要牢牢把握广泛的基础知识——从二进制表示到网络协议，从处理器架构到算法设计。本文提炼了12年级的核心主题，以清晰、贴合考试的方式呈现每一个关键概念。无论你需要快速复习指南，还是需要结构化的概览来发现知识漏洞，以下各节都将帮你建立信心与准确度。
1. Data Representation: Numbers and Beyond | 数据表示：数字与更广泛的编码
Every piece of data inside a computer is ultimately a pattern of bits. The syllabus expects you to work fluently with binary, denary, hexadecimal, and binary-coded decimal (BCD). You must be able to convert between these number systems and explain why hex is a convenient shorthand for binary, especially when representing memory addresses and colour codes.
计算机内部每一份数据最终都是比特的排列。课程要求你能熟练处理二进制、十进制、十六进制以及BCD码。你必须能在这些进制之间相互转换，并能解释为何十六进制是二进制的便捷简写，特别是用于表示内存地址和颜色编码时。
- Bit, nibble, byte, kibibyte (KiB), mebibyte (MiB) – know the exact powers: 1 KiB = 2¹⁰ bytes, 1 MiB = 2²⁰ bytes.
- 比特、半字节、字节、千比字节(KiB)、兆比字节(MiB)——要记清精确的幂次：1 KiB = 2¹⁰ 字节，1 MiB = 2²⁰ 字节。
- Two’s complement representation for signed integers allows the same addition circuitry to handle subtraction.
- 用补码表示有符号整数，能使加法电路同时处理减法。
For real numbers, you meet floating-point representation in the form mantissa × 2exponent. You need to normalise mantissas to maximise precision and understand the trade-off between range and accuracy – the limited number of bits means most real values are mere approximations.
对于实数，我们会遇到尾数 × 2指数形式的浮点表示。你需要对尾数进行规范化以最大化精度，并理解范围与精度之间的权衡——有限的比特位意味着大多数实数值只是近似值。
2. Representing Text, Sound, and Images | 文本、声音与图像的表示
Text is encoded using character sets: ASCII (7-bit or extended 8-bit) and Unicode (UTF-8, UTF-16). Be clear about how ASCII wastes space for international scripts and how Unicode solves this, while UTF-8 retains backward compatibility with ASCII.
文本通过字符集进行编码：ASCII（7位或扩展8位）和Unicode（UTF-8、UTF-16）。要清楚ASCII在国际文字上浪费空间的问题，以及Unicode如何解决，同时UTF-8保持了与ASCII的向后兼容。
Sound is captured by sampling the amplitude of an analogue wave at regular intervals (sample rate) and quantising each sample into a binary number (sample resolution). The size of an uncompressed audio file is calculated as:
le size (bits) = sample rate × sample resolution × duration (seconds) × number of channels
声音通过定期采样模拟波幅（采样率）并将每个样本量化为二进制数字（采样精度）来捕获。未压缩音频文件大小计算方式为：
文件大小(比特) = 采样率 × 采样精度 × 时长(秒) × 声道数
Images break down into pixels, each assigned a colour value. A black-and-white bitmap uses 1 bit per pixel; a 24-bit colour depth allows over 16 million colours. Vector graphics store drawing commands (circles, lines) rather than individual pixels, making them scalable without loss of quality. The file size of a bitmap image is:
size (bits) = width × height × colour depth
图像分解为像素，每个像素赋予一个颜色值。黑白位图每个像素用1位；24位色深则可表示超过1600万种颜色。矢量图形存储绘图命令（圆、线）而非单个像素，因此可以无质量损失地缩放。位图图像的文件大小为：
大小(比特) = 宽度 × 高度 × 色深
3. Compression Techniques | 压缩技术
Lossless compression uses patterns to reduce file size without losing any data. Run-length encoding (RLE) replaces repeated values with a count and the value, which works well on simple graphics but poorly on noisy photographs. Huffman coding assigns shorter codes to more frequent symbols, building a binary tree to ensure unique prefixes.
无损压缩利用模式减小文件尺寸而不丢失数据。游程编码(RLE)用计数值和值本身替代重复的值，对简单图形效果很好，但对噪点多的照片表现不佳。霍夫曼编码为频率更高的符号分配更短的编码，并构建二叉树来保证无歧义前缀。
Lossy compression discards some data that the human senses are unlikely to notice. JPEG reduces image file size by splitting the picture into blocks, applying a discrete cosine transform, quantising, and then encoding. MP3 audio compression removes frequencies outside the range of typical hearing and uses perceptual noise shaping. Always link the method to the nature of the data when explaining the choice of compression in an exam.
有损压缩丢弃人感官不易察觉的数据。JPEG通过分块、离散余弦变换、量化然后编码来减小图像文件大小。MP3音频压缩去除典型听觉范围外的频率，并利用感知噪声整形。在考试中解释为何选择某种压缩时，一定要将方法与数据的特性联系起来。
4. Processor Architecture and the Fetch–Execute Cycle | 处理器架构与取指–执行周期
The von Neumann architecture stores both instructions and data in the same memory space, which simplifies hardware but creates a bottleneck. A processor contains special-purpose registers: Program Counter (PC), Memory Address Register (MAR), Memory Data Register (MDR), Current Instruction Register (CIR), and Accumulator (ACC). The Control Unit (CU) orchestrates the cycle, and the ALU performs arithmetic and logic operations.
冯·诺依曼架构将指令和数据存放在同一内存空间，简化了硬件但造成了瓶颈。处理器包含专用寄存器：程序计数器(PC)、内存地址寄存器(MAR)、内存数据寄存器(MDR)、当前指令寄存器(CIR)以及累加器(ACC)。控制单元(CU)协调整个周期，ALU执行算术和逻辑运算。
The fetch–decode–execute cycle forms the heartbeat of any CPU. During fetch, the address in PC is copied to MAR, a read signal is sent, the instruction arrives in MDR, is moved to CIR, and PC is incremented. Decode interprets the opcode and operand. Execute may involve loading data from memory, performing a calculation in the ALU, or writing a result back. You must be able to describe this sequence precisely, showing how buses (address, data, control) are used at each step.
取指–解码–执行周期构成任何CPU的心跳。在取指阶段，PC中的地址复制到MAR，发出读信号，指令到达MDR，移入CIR，然后PC递增。解码阶段解释操作码和操作数。执行阶段可能包括从内存加载数据、在ALU中执行计算或将结果写回。你必须能精确描述该顺序，展示每一步使用了哪些总线（地址总线、数据总线、控制总线）。
5. Memory and Storage Hierarchy | 内存和存储层次
Computers use a hierarchy to balance speed and cost. Registers are the fastest but fewest. Cache (SRAM) sits between the CPU and main memory, exploiting locality of reference. RAM (DRAM) is volatile and holds currently running programs and data; ROM is non-volatile and stores firmware, such as the bootstrap loader.
计算机使用分层结构来平衡速度与成本。寄存器最快但数量最少。缓存(SRAM)位于CPU和主存之间，利用了局部性原理。RAM(DRAM)是易失性的，保存当前运行的程序和数据；ROM是非易失性的，存储固件，如引导加载程序。
Secondary storage devices provide persistent storage. Hard disk drives (HDDs) use magnetic platters and moving heads; solid-state drives (SSDs) rely on NAND flash memory, offering faster access but higher cost per gigabyte. Optical media (CD, DVD, Blu-ray) are read by laser. When comparing storage devices, reference capacity, speed, portability, durability, and cost – exam questions often ask you to justify a choice for a given scenario, such as backing up a school server or distributing a movie.
辅助存储设备提供持久存储。硬盘驱动器(HDD)使用磁性盘片和移动磁头；固态硬盘(SSD)依赖NAND闪存，提供更快访问速度但每GB成本更高。光学介质（CD、DVD、蓝光）通过激光读取。在比较存储设备时，要引用容量、速度、便携性、耐用性和成本——考题常要求你为特定场景（如备份学校服务器或发行一部电影）论证选择。
6. Operating Systems and Language Translators | 操作系统与语言翻译器
An operating system (OS) manages hardware resources and provides a user interface. Key functions include memory management (paging, segmentation, virtual memory), process scheduling (round-robin, priority-based, short job first), file management, and handling interrupts. The OS also provides a platform for applications through system calls and API.
操作系统(OS)管理硬件资源并提供用户界面。关键功能包括内存管理（分页、分段、虚拟内存）、进程调度（轮转、基于优先级、短作业优先）、文件管理以及处理中断。OS还通过系统调用和API为应用程序提供平台。
Language translators convert human-readable code into machine code. An assembler translates assembly language on a one-to-one basis. A compiler translates an entire high-level source file into object code before execution, generating a standalone executable. An interpreter translates and executes line by line, which is slower in repeated execution but easier for debugging. Many modern environments (e.g., Java) use a hybrid approach: compile to intermediate bytecode, then interpret or JIT-compile on the target machine.
语言翻译器将人类可读代码转化为机器码。汇编器将汇编语言一对一翻译。编译器在执行前将整个高级源文件翻译为目标代码，生成独立的可执行文件。解释器逐行翻译并执行，在重复执行时较慢但便于调试。许多现代环境（如Java）使用混合方式：先编译为中间字节码，然后在目标机上解释或即时编译。
7. Networking Fundamentals | 网络基础
Networks are classified by their scale: LANs (small geographic area, often use Ethernet and Wi-Fi), WANs (connect LANs across cities or countries), and the internet as a global WAN. A typical router connects LANs and forwards packets based on IP addresses, while a switch operates at the data link layer, forwarding frames using MAC addresses on a LAN.
网络按规模分类：LAN（小地域，常用以太网和Wi-Fi）、WAN（跨城市或国家连接LAN），互联网则是全球范围的WAN。典型的路由器连接LAN并依据IP地址转发数据包，而交换机工作在数据链路层，在LAN内使用MAC地址转发帧。
The TCP/IP protocol stack provides layered communication. The application layer (HTTP, FTP, SMTP) generates user data; transport layer (TCP) breaks data into segments, adds port numbers, and ensures reliable delivery via acknowledgements and retransmissions; the internet layer (IP) adds source and destination IP addresses and routes packets; the link layer handles the physical transmission. You must be able to trace what happens at each layer when, for example, a browser requests a web page.
TCP/IP协议栈提供了分层的通信。应用层（HTTP、FTP、SMTP）生成用户数据；传输层（TCP）将数据分割为段，添加端口号，并通过确认和重传来保证可靠交付；网络层（IP）添加源和目的IP地址并路由数据包；链路层负责物理传输。你必须能追踪每一层发生的事，例如浏览器请求网页时。
8. Security, Privacy, and Data Integrity | 安全、隐私与数据完整性
Threats come in many forms: malware (viruses, worms, Trojans), phishing, denial-of-service attacks, and infiltration. Encryption transforms plaintext into ciphertext using a key; symmetric encryption (e.g., AES) uses the same key for both, while asymmetric encryption (e.g., RSA) uses a public key for encryption and a private key for decryption, enabling secure key exchange and digital signatures.
威胁形式多样：恶意软件（病毒、蠕虫、木马）、网络钓鱼、拒绝服务攻击和渗透。加密使用密钥将明文转换为密文；对称加密（如AES）加密解密用同一密钥，而非对称加密（如RSA）用公钥加密、私钥解密，可实现安全的密钥交换和数字签名。
Data integrity techniques ensure that data remains accurate and unauthorised changes are detectable. Parity bits add one extra bit to make the number of 1s either even or odd; checksums compute a numeric sum of a data block; hash functions produce a fixed-length digest – any change to the input yields an entirely different hash. Digital certificates and SSL/TLS combine hashing with asymmetric encryption to authenticate websites and create secure sessions.
数据完整性技术确保数据保持准确，未经授权的改动能被检测到。奇偶校验位额外添加一位令1的个数为偶数或奇数；校验和计算数据块的数值总和；哈希函数产生固定长度的摘要——任何输入变化都会生成完全不同的哈希。数字证书与SSL/TLS则结合哈希与非对称加密来验证网站身份并建立安全会话。
9. Databases and Structured Data | 数据库与结构化数据
A relational database organises data into tables (relations) where each row is a record and each column an attribute. A primary key uniquely identifies each record; foreign keys link tables together, enforcing referential integrity. The syllabus expects you to be comfortable with entity-relationship diagrams (showing one-to-one, one-to-many, and many-to-many relationships) and to normalise data up to third normal form (3NF) to minimise redundancy and anomalies.
关系数据库将数据组织为表（关系），其中每行为一条记录，每列为一个属性。主键唯一标识每条记录；外键将表关联起来，并强制引用完整性。课程要求你能够熟练绘制实体关系图（显示一对一、一对多、多对多关系）并将数据规范化至第三范式(3NF)，以减少冗余和异常。
SQL (Structured Query Language) is used to define and manipulate data. Key statements include SELECT with filters (WHERE), sorting (ORDER BY), grouping (GROUP BY and HAVING), and joining tables (INNER JOIN, LEFT JOIN). You may be given a table and asked to write a query that produces a specific result set – always test your logic on paper with the sample data given.
SQL（结构化查询语言）用于定义和操作数据。关键语句包括带筛选的SELECT（WHERE）、排序（ORDER BY）、分组（GROUP BY和HAVING）以及表连接（INNER JOIN、LEFT JOIN）。考试可能会提供表格要求你编写查询来生成特定结果集——务必在纸上用样本数据检验你的逻辑。
10. Algorithm Design and Pseudocode | 算法设计与伪代码
Algorithmic thinking is central to problem solving. You must be able to write structured pseudocode following the CIE style guide: variable declarations, input/output, assignment (←), conditional statements (IF…THEN…ELSE…ENDIF), and loops (FOR…NEXT, WHILE…ENDWHILE, REPEAT…UNTIL). Clarity and logical structure matter more than syntax perfection, but consistent indentation is essential for readability.
算法思维是解决问题的核心。你必须能按CIE风格指南编写结构化伪代码：变量声明、输入/输出、赋值(←)、条件语句（IF…THEN…ELSE…ENDIF）、循环（FOR…NEXT、WHILE…ENDWHILE、REPEAT…UNTIL）。清晰度和逻辑结构比完美语法更重要，但一致的缩进对可读性至关重要。
Standard algorithms appear repeatedly. A linear search scans each element until a match is found; a binary search works on a sorted array by repeatedly halving the search interval. Bubble sort repeatedly swaps adjacent out-of-order pairs; insertion sort builds the sorted portion one element at a time. You must be able to trace these algorithms, state their time complexity in terms of the number of comparisons, and identify when one is more appropriate than another.
标准算法会反复出现。线性搜索依次扫描每个元素直到找到匹配项；二分搜索在有序数组中通过不断将搜索区间减半来工作。冒泡排序反复交换相邻的乱序对；插入排序一次一个元素地构建有序部分。你必须能追踪这些算法，根据比较次数说明其时间复杂度，并判断何时一种算法比另一种更合适。
11. Programming Fundamentals and Data Structures | 编程基础与数据结构
Beyond the syntax, you need to understand structured programming concepts: sequence, selection, iteration, and modularisation using functions/procedures. Parameters can be passed by value (a copy) or by reference (the original can be altered). A well-designed program is broken into cohesive modules that are easier to test and maintain.
除了语法，你还需要理解结构化编程概念：顺序、选择、迭代，以及使用函数/过程的模块化。参数可以按值传递（传递副本）或按引用传递（原始值可被修改）。设计良好的程序被分解为内聚的模块，更易于测试和维护。
The one-dimensional array is the most common composite data type. Indexed from zero (or one – check the context), arrays store elements under a single identifier and allow random access. You should be able to declare an array, iterate through it with a loop, search for an element, and update it. More advanced AS questions may introduce a two-dimensional array to model a grid or board, so practise addressing elements as MyArray[row_index][column_index].
一维数组是最常见的复合数据类型。从零（或一——要依据上下文）开始索引，数组将元素存储在一个标识符下并允许随机访问。你应该能够声明数组、用循环遍历它、搜索元素并更新它。更进阶的AS题目可能引入二维数组来模拟网格或棋盘，因此要练习用MyArray[row_index][column_index]的方式访问元素。
12. Ethical, Legal, and Ownership Issues | 道德、法律与所有权议题
Computing does not exist in a vacuum. The CIE syllabus expects you to discuss the impact of technology on society. This includes issues of privacy (data collection, profiling), computer misuse (hacking, cyberbullying), intellectual property (software copyright, patents, open-source licences), and the digital divide. In essay-style questions, always structure your answer to present balanced arguments, give concrete examples (such as the role of AI in recruitment or surveillance), and link back to the scenario in the question.
计算并非孤立存在。CIE课程要求你讨论技术对社会的影响。这包括隐私问题（数据收集、用户画像）、计算机滥用（黑客、网络欺凌）、知识产权（软件版权、专利、开源许可）以及数字鸿沟。在论述型题目中，始终使答案结构平衡，给出具体例子（如AI在招聘或监控中的作用），并紧扣题目中的场景。
Published by TutorHao | CIE Computer Science (9618) AS-Level 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