📚 Year 13 WJEC Computer Science: Core Knowledge Review | Year 13 WJEC 计算机：核心知识点梳理
This article provides a structured review of the essential topics in the Year 13 WJEC Computer Science syllabus. It covers data representation, computer architecture, operating systems, networking, data structures, algorithms, object‑oriented programming, databases, the theory of computation, and emerging technologies. Mastering these areas is crucial for the A2 units and builds on AS foundations.
本文系统梳理了 Year 13 WJEC 计算机科学课程的核心知识点，涵盖数据表示、计算机体系结构、操作系统、网络、数据结构、算法、面向对象编程、数据库、计算理论以及新兴技术。掌握这些内容对 A2 单元考试至关重要，并巩固 AS 阶段的基础。
1. Data Representation and Compression | 数据表示与压缩
Floating‑point numbers are stored using the IEEE 754 standard, which splits a value into sign, exponent, and mantissa fields. Normalising the mantissa guarantees a unique representation and maximises precision by removing leading zeros.
浮点数采用 IEEE 754 标准存储，将数值分为符号、指数和尾数三个部分。对尾数进行规范化可以消除前导零，确保唯一表示并最大程度提高精度。
Character data is encoded with standard systems such as ASCII (7‑bit) and Unicode (UTF‑8, UTF‑16). Unicode supports a vast range of global scripts, while ASCII remains compact but limited to English characters and control codes.
字符数据使用标准编码系统，例如 ASCII（7位）和 Unicode（UTF‑8、UTF‑16）。Unicode 支持全球大量文字，而 ASCII 虽然紧凑，但仅限于英文字符和控制码。
Lossless compression (e.g. run‑length encoding, Huffman coding) reduces file size without losing data, relying on statistical redundancy. Lossy compression (e.g. JPEG, MP3) sacrifices some fidelity to achieve higher compression ratios suitable for images and audio.
无损压缩（如游程编码、哈夫曼编码）通过利用统计冗余缩小文件体积而不丢失数据。有损压缩（如 JPEG、MP3）牺牲一定保真度，以换取更适合图像和音频的更高压缩比。
Symmetric encryption uses a single shared key (e.g. AES), while asymmetric encryption employs a public‑private key pair (e.g. RSA). Hashing produces a fixed‑length digest for integrity checks but cannot be reversed.
对称加密使用单一共享密钥（如 AES），非对称加密则使用公私钥对（如 RSA）。散列算法产生固定长度摘要用于完整性校验，且不可逆。
2. Computer Architecture and Assembly Language | 计算机体系结构与汇编语言
The processor consists of the control unit, ALU, registers (PC, MAR, MDR, CIR, accumulator), and buses (address, data, control). The fetch‑decode‑execute cycle describes how instructions are processed step by step.
处理器由控制单元、ALU、寄存器（PC、MAR、MDR、CIR、累加器）和总线（地址、数据、控制）组成。取指‑解码‑执行周期描述了指令逐步处理的过程。
Pipelining improves throughput by overlapping the execution of multiple instructions. Hazards (data, structural, control) can cause stalls, which are mitigated by forwarding, branch prediction, or interlock mechanisms.
流水线技术通过重叠多条指令的执行来提高吞吐量。数据冒险、结构冒险和控制冒险可能导致停顿，可通过转发、分支预测或互锁机制予以缓解。
RISC architectures use a small, highly optimised instruction set with fixed‑length instructions, while CISC processors support complex, variable‑length instructions that can perform multi‑step operations in one.
RISC 架构使用精简、高度优化的固定长度指令集，而 CISC 处理器支持复杂的变长指令，能够在一条指令中完成多步操作。
Assembly language instructions use mnemonics (LDR, STR, ADD, B) and addressing modes such as immediate, direct, indirect, indexed, and register indirect. These give programmers low‑level control over memory and I/O.
汇编语言指令使用助记符（LDR、STR、ADD、B）及多种寻址模式，如立即寻址、直接寻址、间接寻址、变址寻址和寄存器间接寻址，为程序员提供对内存和 I/O 的底层控制。
3. Operating Systems and Memory Management | 操作系统与内存管理
The operating system acts as an interface between hardware and user, handling resource management, process scheduling, memory allocation, file systems, and user authentication.
操作系统充当硬件与用户之间的接口，负责资源管理、进程调度、内存分配、文件系统以及用户身份验证。
Interrupts are signals that alert the CPU to high‑priority events; the ISR (interrupt service routine) executes while the current context is saved. Polling and vectored interrupts are alternative handling strategies.
中断是提醒 CPU 注意高优先级事件的信号；中断服务例程（ISR）执行时当前上下文会被保存。轮询和向量化中断是两种不同的处理策略。
Memory management techniques include paging, segmentation, and virtual memory. Pages are fixed‑size blocks mapped to frames in physical memory, allowing programs to run even if they exceed RAM size.
内存管理技术包括分页、分段和虚拟内存。页是固定大小块，映射到物理内存中的帧，使得程序即使超过 RAM 容量也能运行。
Disk scheduling algorithms (FCFS, SSTF, SCAN, C‑SCAN) reduce seek time. A file allocation table (FAT) or indexed allocation records where file blocks reside on disk.
磁盘调度算法（FCFS、SSTF、SCAN、C‑SCAN）可缩短寻道时间。文件分配表（FAT）或索引分配记录文件块在磁盘上的位置。
4. Communication and Network Protocols | 通信与网络协议
The TCP/IP protocol stack comprises four layers: application (HTTP, FTP, SMTP), transport (TCP, UDP), internet (IP), and network access (Ethernet, Wi‑Fi). Each layer encapsulates data with its own header.
TCP/IP 协议栈分为四层：应用层（HTTP、FTP、SMTP）、传输层（TCP、UDP）、互联网层（IP）和网络接入层（以太网、Wi‑Fi）。每一层用各自的头部对数据进行封装。
IP addresses (IPv4 32‑bit, IPv6 128‑bit) uniquely identify devices. Subnet masks define network and host portions. Network Address Translation (NAT) allows multiple private IPs to share one public address.
IP 地址（IPv4 32位，IPv6 128位）唯一标识设备。子网掩码定义网络地址和主机地址部分。网络地址转换（NAT）使多个私有 IP 共享一个公网地址。
The Domain Name System (DNS) resolves human‑readable domain names to IP addresses through a hierarchical, distributed database. Recursive and iterative queries speed up resolution.
域名系统（DNS）通过分层分布式数据库将人类可读的域名解析为 IP 地址。递归查询和迭代查询可加快解析过程。
Firewalls filter packets based on rules, while encryption (SSL/TLS) secures data in transit. Digital signatures provide authentication and non‑repudiation using hashing and asymmetric keys.
防火墙根据规则过滤数据包，加密（SSL/TLS）保护传输中的数据。数字签名利用散列和非对称密钥提供身份认证和不可否认性。
5. Data Structures and Abstract Data Types | 数据结构与抽象数据类型
Static arrays have fixed size and allow O(1) indexed access. Dynamic structures such as linked lists grow and shrink at runtime, but element access requires O(n) traversal unless a tail pointer is maintained.
静态数组大小固定，支持 O(1) 的索引访问。链表等动态结构可在运行时增长和收缩，但访问元素需要 O(n) 的遍历，除非维护了尾指针。
A stack (LIFO) supports push and pop; a queue (FIFO) supports enqueue and dequeue. Both can be implemented with arrays or linked lists and are fundamental for recursion and scheduling.
栈（LIFO）支持压入和弹出；队列（FIFO）支持入队和出队。两者都可用数组或链表实现，是递归和调度的基础。
Binary search trees order nodes so that left children are smaller and right children larger, enabling O(log n) search when balanced. Tree traversals include pre‑order, in‑order, and post‑order.
二叉搜索树的节点排列使得左子节点较小、右子节点较大，平衡时可以实现 O(log n) 的查找。树的遍历包括前序、中序和后序遍历。
Hash tables map keys to indices via a hash function. Collisions are resolved by chaining or open addressing. With a good hash function, average lookup is O(1).
哈希表通过哈希函数将键映射到索引。冲突通过链接法或开放寻址解决。良好的哈希函数能使平均查找复杂度达到 O(1)。
6. Algorithm Complexity and Graph Algorithms | 算法复杂度与图算法
Big‑O notation describes the upper bound of an algorithm’s time or space complexity. Common classes include O(1), O(log n), O(n), O(n log n), O(n²), and O(2ⁿ).
大 O 表示法描述算法时间复杂度或空间复杂度的上界。常见类别有 O(1)、O(log n)、O(n)、O(n log n)、O(n²) 和 O(2ⁿ)。
Linear search scans each item (O(n)); binary search requires sorted data and achieves O(log n). Quicksort and merge sort both run in O(n log n) on average, but quicksort can degrade to O(n²) if pivot selection is poor.
线性搜索逐一扫描每个项（O(n)）；二分搜索要求数据有序，复杂度为 O(log n)。快速排序和归并排序平均运行时间为 O(n log n)，但快速排序若枢轴选择不当可能退化为 O(n²)。
Graphs can be represented by adjacency matrices (O(V²) space) or adjacency lists (O(V+E) space). Dijkstra’s algorithm finds the shortest path in a weighted graph with non‑negative edges; it has O((V+E) log V) complexity when using a priority queue.
图可以用邻接矩阵（O(V²) 空间）或邻接表（O(V+E) 空间）表示。Dijkstra 算法可以在边权非负的加权图中找到最短路径，使用优先队列时复杂度为 O((V+E) log V)。
The A* algorithm improves on Dijkstra by adding a heuristic, allowing faster path‑finding in applications like GPS navigation. Decision problems are grouped into P (polynomial) and NP (non‑deterministic polynomial) classes; NP‑hard problems have no known polynomial solution.
A* 算法在 Dijkstra 基础上增加了启发式函数，能够在 GPS 导航等应用中更快寻路。判定问题分为 P（多项式）和 NP（非确定多项式）类；NP 难问题目前没有已知的多项式解法。
7. Object‑Oriented Programming Principles | 面向对象编程原则
Encapsulation binds data (attributes) and methods within a class, restricting direct access via visibility modifiers (private, public, protected). This shields an object’s internal state and reduces side effects.
封装将数据（属性）和方法绑定在类内部，并通过可见性修饰符（private、public、protected）限制直接访问，从而保护对象的内部状态并减少副作用。
Inheritance allows a subclass to reuse and extend the properties and behaviours of a superclass, promoting code reuse and hierarchical classification.
继承允许子类重用并扩展父类的属性和行为，促进了代码重用和层次化分类。
Polymorphism lets objects of different classes respond to the same method name differently. Overriding achieves runtime polymorphism; overloading provides compile‑time polymorphism with different parameter signatures.
多态使得不同类的对象能够以不同方式响应相同的方法名。方法重写实现运行时多态；方法重载通过不同的参数签名实现编译时多态。
Abstract classes and interfaces define contracts without full implementation. Common design patterns such as Singleton (single instance) and Factory (object creation without specifying exact class) encapsulate best practices.
抽象类和接口定义契约而不提供完整实现。常见的设计模式，如单例模式（唯一实例）和工厂模式（不指定具体类即创建对象），封装了最佳实践。
8. Relational Databases and Normalisation | 关系数据库与规范化
A relational database organises data into tables (relations) linked by primary and foreign keys. SQL (Structured Query Language) is used to define, manipulate, and query data with statements such as SELECT, INSERT, UPDATE, DELETE, and JOIN.
关系数据库将数据组织为由主键和外键连接的表（关系）。SQL（结构化查询语言）用于定义、操作和查询数据，常用语句包括 SELECT、INSERT、UPDATE、DELETE 和 JOIN。
Normalisation reduces data redundancy and update anomalies by decomposing tables. First normal form (1NF) requires atomic values; second normal form (2NF) removes partial dependencies on a composite key; third normal form (3NF) eliminates transitive dependencies.
规范化通过分解表来减少数据冗余和更新异常。第一范式（1NF）要求原子值；第二范式（2NF）消除对组合键的部分依赖；第三范式（3NF）消除传递依赖。
Indexes speed up SELECT queries at the cost of slower writes. An atomic transaction satisfies ACID properties: Atomicity, Consistency, Isolation, Durability, ensuring data reliability in concurrent environments.
索引可加速 SELECT 查询，但会降低写入速度。原子事务满足 ACID 性质：原子性、一致性、隔离性和持久性，确保并发环境下的数据可靠性。
9. Theory of Computation and Automata | 计算理论与自动机
Finite state machines (FSMs) model systems with a finite number of states and transitions triggered by inputs. Deterministic (DFA) and non‑deterministic (NFA) automata recognise regular languages, which can also be described by regular expressions.
有限状态机（FSM）用有限的状态和输入触发的转移来建模系统。确定性（DFA）和非确定性（NFA）自动机识别正则语言，正则语言同样可以用正则表达式描述。
Context‑free grammars (CFGs) generate context‑free languages and are used to describe programming language syntax. Pushdown automata recognise these languages, adding a stack to the FSM model.
上下文无关文法（CFG）生成上下文无关语言，用于描述程序语言的语法。下推自动机通过为 FSM 模型增加一个栈来识别这些语言。
A Turing machine is an abstract computational model with an infinite tape and a read/write head; it can simulate any algorithm. The Church‑Turing thesis states that anything computable can be computed by a Turing machine.
图灵机是一种抽象计算模型，拥有无限长度的纸带和读写头，可以模拟任何算法。丘奇‑图灵论题指出，任何可计算的问题都可以由图灵机计算。
The halting problem proves that no general algorithm can determine whether an arbitrary program will halt, demonstrating fundamental limits of computation and undecidability.
停机问题证明，不存在一个通用算法能够判断任意程序是否会停止运行，这揭示了计算的根本局限性和不可判定性。
10. Emerging Trends and Ethical Concerns | 新兴趋势与伦理关切
Artificial intelligence and machine learning use algorithms to identify patterns and make decisions. Supervised learning trains on labelled data, while unsupervised learning finds hidden structures in unlabelled data.
人工智能和机器学习使用算法识别模式并做出决策。监督学习基于标注数据进行训练，无监督学习则从未标注数据中发现隐藏结构。
Big data refers to extremely large datasets analysed by distributed systems to reveal trends. Cloud computing offers scalable, on‑demand resources (IaaS, PaaS, SaaS) but raises security and privacy risks.
大数据指通过分布式系统分析以揭示趋势的超大数据集。云计算提供可弹性伸缩的按需资源（IaaS、PaaS、SaaS），但也带来安全和隐私风险。
Data protection laws such as GDPR enforce user consent, right to access, and right to erasure. Computer misuse acts criminalise unauthorised access and malware distribution.
GDPR 等数据保护法要求用户同意、赋予访问权与删除权。计算机滥用法将未经授权的访问和恶意软件传播定为刑事犯罪。
Ethical dilemmas arise around facial recognition, algorithmic bias, and intellectual property. Designers must embed fairness, transparency, and accountability into digital systems from the outset.
人脸识别、算法偏见和知识产权引发伦理困境。设计者必须从一开始就将公平性、透明性和可问责性嵌入数字系统。
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