📚 Pre-U AQA Computer Science: Core Knowledge Areas | Pre-U AQA 计算机核心知识点梳理
The Pre-U Computer Science qualification from AQA bridges the gap between A-level and first-year undergraduate study, demanding both depth and breadth across theoretical and applied topics. This article distils the core knowledge areas into a structured revision guide, covering hardware, data, networking, algorithms, and the wider consequences of computing.
AQA 的 Pre-U 计算机科学课程衔接 A-level 与大学一年级，要求考生在理论与应用方面兼具深度和广度。本文将核心知识点提炼成系统的复习指南，涵盖硬件、数据、网络、算法以及计算的社会影响。
1. Computer Systems Architecture | 计算机系统架构
The von Neumann architecture is the foundational model for most computers, characterised by a single shared memory for both instructions and data, and a sequential fetch-decode-execute cycle. Alternatives like Harvard architecture separate instruction and data memories to enable parallel access, which is common in embedded systems.
冯·诺依曼架构是大多数计算机的基础模型，其特点是采用单一共享存储器存放指令和数据，并遵循顺序的取指-译码-执行周期。另一种哈佛架构将指令与数据存储器分开，可实现并行访问，常用于嵌入式系统。
Inside the CPU, the Control Unit (CU) directs operations by interpreting instructions, while the Arithmetic Logic Unit (ALU) performs arithmetic and logical functions. Key registers include the Program Counter (PC) holding the next instruction address, the Memory Address Register (MAR) and Memory Data Register (MDR) interfacing with memory, and the Current Instruction Register (CIR) storing the fetched instruction. Accumulator and status registers support execution.
CPU 内部，控制单元 (CU) 通过译码指令来指挥操作，算术逻辑单元 (ALU) 执行算术与逻辑功能。关键寄存器包括：存放下一指令地址的程序计数器 (PC)、与存储器接口的存储器地址寄存器 (MAR) 和存储器数据寄存器 (MDR)、存储被取指令的当前指令寄存器 (CIR)。累加器和状态寄存器辅助执行。
The fetch-decode-execute cycle proceeds as follows: the PC value is placed on the address bus and memory is read, the instruction is loaded into the CIR, the CU decodes it, and the required operation is performed, possibly involving the ALU or further memory accesses. The PC is then updated to point to the next instruction.
取指-译码-执行周期过程如下：PC 的值送上地址总线并读取存储器，指令装入 CIR，控制单元译码后执行所需的操作，可能涉及 ALU 或进一步访存，然后 PC 更新指向下一条指令。
Modern processors boost performance through pipelining, where multiple instructions overlap their execution stages; superscalar architectures that issue several instructions per clock cycle; and a memory hierarchy of registers, cache (L1, L2, L3), RAM, and secondary storage. Cache exploits temporal and spatial locality to reduce average memory access time.
现代处理器通过流水线（多条指令各阶段重叠）、超标量架构（每时钟周期发射多条指令）以及寄存器、缓存 (L1, L2, L3)、RAM 和二级存储构成的内存层次来提升性能。缓存利用时间局部性与空间局部性降低平均访存时间。
The address bus carries memory locations, the data bus transfers data, and the control bus carries timing and command signals. The width of the address bus determines the maximum addressable memory (e.g., a 32-bit address bus can address 2³² locations). Bus arbitration and interrupts manage device access.
地址总线传送存储地址，数据总线传送数据，控制总线承载时序与命令信号。地址总线的宽度决定了最大可寻址内存（如 32 位地址总线可寻址 2³² 个单元）。总线仲裁与中断管理设备的访问。
2. Data Representation | 数据表示
All data in a computer is represented in binary. Unsigned integers are stored directly, while signed integers typically use two’s complement, where the most significant bit represents a negative weight. The two’s complement of a number is found by inverting bits and adding 1; it simplifies arithmetic as addition and subtraction can use the same hardware.
计算机中所有数据均用二进制表示。无符号整数直接存储，有符号整数通常采用补码（二补数）表示，其最高位代表负权。一个数的补码通过按位取反再加 1 得到；补码简化了运算，因为加减法可复用同一硬件。
Real numbers are often represented in floating-point format based on the IEEE 754 standard: sign, exponent, and mantissa. Normalised form maximises precision. Programs must be aware of rounding errors and limited precision, as 0.1 cannot be represented exactly in binary floating point.
实数常采用 IEEE 754 标准的浮点格式：符号位、指数和尾数。规格化形式可最大化精度。程序必须留意舍入误差和有限精度，因为 0.1 无法用二进制浮点精确表示。
Characters are encoded using schemes such as ASCII (7-bit, 128 characters) and Unicode (UTF-8, UTF-16). Unicode provides a unique code point for characters across global scripts, supporting multilingual text. UTF-8 is backward compatible with ASCII and variable length.
字符通过编码方案表示，如 ASCII（7 位，128 个字符）和 Unicode (UTF-8, UTF-16)。Unicode 为全球文字提供唯一码位，支持多语言文本。UTF-8 与 ASCII 后向兼容且变长。
Images are represented as bitmaps (rasters) of pixels, each pixel having a colour depth (bits per pixel) defining the number of available colours. Vector graphics describe images through geometric primitives, allowing lossless scaling. Sound is sampled at a given sample rate and bit depth; the Nyquist theorem states the sampling rate must be at least twice the highest frequency to avoid aliasing.
图像表示为像素位图（光栅），每个像素的色深（每像素位数）定义了可用的颜色数。矢量图形通过几何图元描述图像，可无损缩放。声音以给定的采样率和位深度进行采样；奈奎斯特定理指出，采样率须至少为最高频率的两倍以避免混叠。
3. System Software | 系统软件
An operating system (OS) acts as an intermediary between users/hardware, managing resources: process scheduling, memory management (paging, segmentation, virtual memory), file systems, I/O handling, and security. Kernel mode versus user mode protects critical operations.
操作系统 (OS) 作为用户与硬件之间的中介，管理各项资源：进程调度、内存管理（分页、分段、虚拟内存）、文件系统、I/O 处理与安全保护。内核态与用户态隔离保护关键操作。
Translation software includes assemblers (assembly language to machine code), compilers (high-level language to machine code, producing an executable file), and interpreters (executing source code line by line without producing a standalone executable). Just-in-time (JIT) compilers combine interpretation with compilation for runtime performance.
翻译软件包括汇编器（汇编语言→机器码）、编译器（高级语言→机器码，生成可执行文件）和解释器（逐行执行源代码，不生成独立可执行文件）。即时 (JIT) 编译结合解释与编译以提升运行时性能。
The compilation process involves lexical analysis, syntax analysis (parsing, building a parse tree or abstract syntax tree), semantic analysis, code generation, and optimisation. Linkers combine object files into a single executable; loaders place programs into memory for execution. Virtual machines like the Java Virtual Machine (JVM) allow platform independence by executing bytecode.
编译过程包括词法分析、语法分析（解析，构造语法树或抽象语法树）、语义分析、代码生成和优化。链接器将多个目标文件合并成单一可执行文件；装载器将程序载入内存执行。像 Java 虚拟机 (JVM) 这样的虚拟机通过执行字节码实现平台无关性。
4. Computer Networks and Security | 计算机网络与安全
Networks are often modelled using the TCP/IP stack: Application (HTTP, FTP, SMTP), Transport (TCP, UDP), Internet (IP), and Network Access/link layers. TCP provides reliable, connection-oriented delivery with flow control and error correction; UDP is connectionless and faster, suited for streaming.
网络常使用 TCP/IP 模型进行分层：应用层 (HTTP, FTP, SMTP)、传输层 (TCP, UDP)、网际层 (IP) 和网络接口/链路层。TCP 提供可靠的面向连接传输，具备流量控制和纠错；UDP 是无连接的，速度更快，适合流媒体。
IP addressing (IPv4, IPv6) routes packets across the internet; DNS translates domain names to IP addresses. NAT allows private addresses to share public IPs. Subnetting divides networks for efficiency.
IP 寻址 (IPv4, IPv6) 在互联网中路由数据包；DNS 将域名解析为 IP 地址。NAT 允许多个私有地址共享公网 IP。子网化将网络分隔以提高效率。
Network security threats include malware (viruses, worms, trojans), phishing, denial-of-service (DoS/DDoS) attacks, and man-in-the-middle interceptions. Defences involve firewalls (packet filter, stateful inspection), encryption, authentication, access control, and intrusion detection systems.
网络安全威胁包括恶意软件（病毒、蠕虫、木马）、网络钓鱼、拒绝服务攻击 (DoS/DDoS) 和中间人截取。防御手段包括防火墙（包过滤、状态检测）、加密、身份认证、访问控制以及入侵检测系统。
Symmetric encryption (e.g., AES) uses a shared secret key; asymmetric encryption (public-key cryptography, e.g., RSA) uses a public/private key pair. Digital signatures and certificates provide integrity and non-repudiation. Hashing (e.g., SHA-256) produces a fixed-size digest for data verification.
对称加密（如 AES）使用共享密钥；非对称加密（公钥加密，如 RSA）使用公钥/私钥对。数字签名和数字证书提供完整性和不可否认性。哈希算法（如 SHA-256）生成固定长度摘要用于数据验证。
5. Algorithms and Problem Solving | 算法与问题解决
An algorithm is a finite, well-defined sequence of steps to solve a problem. Key search algorithms include linear search (O(n)) and binary search (O(log n), requiring sorted data). Sorting algorithms: bubble sort (O(n²)), insertion sort (O(n²), but adaptive), merge sort (O(n log n), stable, divide-and-conquer), and quicksort (average O(n log n), worst O(n²) if poor pivot chosen, in-place).
算法是解决问题的一组有限且明确定义的步骤。核心搜索算法包括线性搜索 (O(n)) 和二分搜索 (O(log n)，要求数据有序)。排序算法：冒泡排序 (O(n²))、插入排序 (O(n²) 但自适应)、归并排序 (O(n log n)，稳定，分治) 和快速排序（平均 O(n log n)，最坏 O(n²) 若枢轴选择不佳，原地排序）。
Big O notation expresses upper-bound time/space complexity in terms of input size n. Common classes: O(1) constant, O(log n) logarithmic, O(n) linear, O(n log n) linearithmic, O(n²) quadratic, O(2ⁿ) exponential. Understanding this allows informed choices about algorithm suitability.
大 O 表示法用输入规模 n 来表示上界时间/空间复杂度。常见类别：O(1) 常数，O(log n) 对数，O(n) 线性，O(n log n) 线性对数，O(n²) 平方，O(2ⁿ) 指数。理解这一点有助于合理选择算法。
Problem-solving techniques include abstraction (removing unnecessary detail to focus on essentials), decomposition (breaking a problem into sub-problems), and pattern recognition. Recursion uses a function that calls itself with a base case to prevent infinite recursion. Backtracking explores candidate solutions incrementally, abandoning partial candidates that cannot lead to a valid solution (e.g., maze solving, N-Queens). Dynamic programming stores intermediate results to avoid recomputation.
问题解决技术包括抽象（移除不必要的细节以聚焦本质）、分解（将问题拆分成子问题）和模式识别。递归使用自调用的函数，并设置基准情形防止无限递归。回溯法逐步尝试候选解，一旦部分候选解不可能导出有效解便舍弃（如迷宫求解、N-皇后问题）。动态规划存储中间结果以避免重复计算。
6. Programming Concepts | 编程概念
Programming paradigms shape how solutions are expressed. Procedural programming organises code into procedures/functions that operate on data; imperative languages like C follow this. Object-oriented programming (OOP) encapsulates data and behaviour into objects using classes, supporting inheritance, polymorphism, and abstraction. Functional programming treats computation as the evaluation of mathematical functions, avoiding side effects and mutable state; first-class functions and recursion are essential.
编程范式塑造解决方案的表达方式。过程式编程将代码组织成操作数据的过程/函数；C 等命令式语言遵循此范式。面向对象编程 (OOP) 通过类将数据和行为封装为对象，支持继承、多态和抽象。函数式编程将计算视为数学函数的求值，避免副作用与可变状态；一等函数和递归至关重要。
Data types include primitive types (integer, float, boolean, char) and composite types (arrays, strings, records/structs). Type systems can be static/dynamic and strong/weak. Parameter passing mechanisms: pass-by-value copies the argument; pass-by-reference provides an alias to the original variable. Scope defines the visibility of variables (local, global, block).
数据类型包括原语类型（整型、浮点型、布尔型、字符）和复合类型（数组、字符串、记录/结构体）。类型系统可分为静态/动态和强/弱类型。参数传递机制：按值传递复制实参；按引用传递提供原变量的别名。作用域定义变量的可见性（局部、全局、块级）。
Program construction relies on simple constructs: sequence, selection (if-else, switch), iteration (for, while). Robust programs handle exceptions and validate inputs. Testing encompasses white-box (unit, integration, covering paths) and black-box (equivalence partitioning, boundary value analysis). Alpha/beta testing and acceptance tests complete the cycle.
程序构建基于简单结构：顺序、选择 (if-else, switch)、循环 (for, while)。健壮的程序处理异常并验证输入。测试涵盖白盒测试（单元测试、集成测试，覆盖路径）和黑盒测试（等价类划分、边界值分析）。Alpha/Beta 测试和验收测试完成周期。
7. Data Structures | 数据结构
Data structures organise data for efficient access and modification. Static arrays store elements of the same type contiguously; random access is O(1), but insertion/deletion is O(n) due to shifting. Dynamic arrays (e.g., ArrayList) resize automatically, amortising cost.
数据结构组织数据以实现高效访问与修改。静态数组将同类型元素连续存储；随机访问为 O(1)，但插入/删除因移位为 O(n)。动态数组（如 ArrayList）自动调整大小，均摊了成本。
A linked list consists of nodes, each with data and a pointer to the next node; insertion/deletion is O(1) at known positions, but search is O(n). Stacks follow Last-In-First-Out (LIFO) and support push and pop; queues are First-In-First-Out (FIFO). Both can be implemented with arrays or linked lists. Priority queues extract highest priority element first, often using a heap.
链表由节点构成，每个节点包含数据与指向下一节点的指针；在已知位置插入/删除为 O(1)，但搜索为 O(n)。栈遵循后进先出 (LIFO)，支持压入和弹出；队列是先进先出 (FIFO)。两者均可用数组或链表实现。优先队列优先取出最高优先级元素，通常用堆实现。
Trees are hierarchical: a binary tree has at most two children; a binary search tree (BST) maintains order left < root < right, enabling O(log n) search if balanced. Heaps (min/max) are complete binary trees satisfying the heap property, used in heap sort and priority queues. Graphs (vertices, edges) can be directed/undirected, weighted, and represented by adjacency matrices or adjacency lists. Hash tables use a hash function to map keys to buckets, giving expected O(1) search; collisions are resolved by chaining or open addressing.
树是层次结构：二叉树至多有两个子节点；二叉搜索树 (BST) 维持左 < 根 < 右的次序，若平衡则搜索为 O(log n)。堆（最小堆/最大堆）是满足堆性质的完全二叉树，用于堆排序和优先队列。图（顶点、边）可分为有向/无向、加权，用邻接矩阵或邻接表表示。哈希表使用哈希函数将键映射到桶，期望搜索 O(1)；冲突通过链地址法或开放寻址解决。
8. Databases | 数据库
The relational model organises data into tables (relations) with rows (tuples) and columns (attributes). A primary key uniquely identifies each tuple; foreign keys define relationships between tables, enforcing referential integrity. Entity-relationship diagrams model data before implementation.
关系模型将数据组织成表（关系），包含行（元组）和列（属性）。主键唯一标识每个元组；外键定义表间关系，强制引用完整性。实体-关系图在实现前对数据建模。
SQL is the standard language for relational databases. Data Definition Language (DDL) commands (CREATE, ALTER, DROP) define the schema. Data Manipulation Language (DML) commands (SELECT, INSERT, UPDATE, DELETE) query and modify data. Joins combine tables: INNER JOIN returns matching rows; LEFT/RIGHT OUTER joins include unmatched rows. Aggregation functions (COUNT, SUM, AVG) and GROUP BY with HAVING summarise data.
SQL 是关系数据库的标准语言。数据定义语言 (DDL) 命令 (CREATE, ALTER, DROP) 定义模式。数据操纵语言 (DML) 命令 (SELECT, INSERT, UPDATE, DELETE) 查询和修改数据。连接 (Join) 合并表：内连接返回匹配行；左/右外连接包含不匹配的行。聚合函数 (COUNT, SUM, AVG) 和 GROUP BY 配合 HAVING 汇总数据。
Normalisation reduces data redundancy and anomalies. 1NF: atomic values and no repeating groups. 2NF: 1NF plus no partial dependencies on a composite key. 3NF: 2NF plus no transitive dependencies. Transaction processing ensures ACID properties: Atomicity, Consistency, Isolation, Durability. Indexes speed up queries on specified columns, similar to a book’s index.
规范化减少数据冗余和异常。第一范式：原子值，无重复组。第二范式：满足第一范式且不存在对复合键的部分依赖。第三范式：满足第二范式且不存在传递依赖。事务处理确保 ACID 性质：原子性、一致性、隔离性、持久性。索引加速指定列的查询，类似于书的索引。
9. Theory of Computation | 计算理论
Finite state machines (FSMs) abstract systems with a finite number of states; Mealy machines produce outputs on transitions, Moore machines on states. FSMs recognise regular languages. Regular expressions define patterns for searching and lexical analysis.
有限状态机 (FSM) 抽象有限状态的系统；Mealy 机在状态转移时产生输出，Moore 机在状态上产生输出。FSM 识别正则语言。正则表达式定义用于搜索和词法分析的模式。
A Turing machine is a mathematical model of computation consisting of an infinite tape, a head that reads/writes, and a state register. It can simulate any algorithm, and the Church-Turing thesis asserts that any effectively calculable function is computable by a Turing machine. The Universal Turing Machine can simulate any other Turing machine.
图灵机是计算模型，包含无限长纸带、读写头和状态寄存器。它能模拟任何算法，丘奇-图灵论题主张任何有效可计算的函数均可由图灵机计算。通用图灵机能模拟任何其他图灵机。
The Halting Problem is undecidable: no general algorithm can determine whether an arbitrary program will halt on a given input. This proves limits of computability. Complexity classes P (solvable in polynomial time by deterministic Turing machine) and NP (verifiable in polynomial time) capture feasibility. NP-hard and NP-complete problems (e.g., Boolean satisfiability, travelling salesman decision problem) are central to the P versus NP question.
停机问题不可判定：不存在通用的算法能判定任意程序在给定输入上是否会终止。这证明了可计算性的限制。复杂度类 P（确定性图灵机多项式时间可解）和 NP（多项式时间可验证）刻画可行性。NP-难和 NP-完全问题（如布尔可满足性、旅行商判定问题）是 P 与 NP 问题的核心。
10. Ethical, Legal and Environmental Issues | 道德、法律与环境问题
Computing raises ethical concerns around privacy, surveillance, algorithmic bias, and the responsible use of artificial intelligence. Professional bodies like the BCS provide codes of conduct requiring public interest, duty of care, and professional integrity.
计算引发隐私、监控、算法偏见和负责任地使用人工智能等伦理关切。BCS 等专业机构制定行为准则，要求维护公共利益、尽职关怀和专业诚信。
Data protection legislation, such as the UK Data Protection Act 2018 and GDPR, governs the collection, storage, and processing of personal data. Individuals have rights to access, rectify, and erase their data. Computer misuse laws criminalise unauthorised access and malware distribution.
数据保护法律（如英国 2018 年《数据保护法》和 GDPR）规范个人数据的收集、存储和处理。个人有权访问、更正和删除其数据。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。
Intellectual property (IP) law protects software through copyright, patents, and licenses. Open-source licenses (GPL, MIT) permit use and modification under certain conditions. Digital rights management (DRM) controls distribution.
知识产权 (IP) 法通过版权、专利和许可证保护软件。开源许可证（GPL、MIT）允许在一定条件下使用和修改。数字版权管理 (DRM) 控制分发。
Environmental sustainability in computing addresses the energy consumption of data centres, e-waste, and the whole lifecycle of equipment. Virtualisation and cloud computing can improve efficiency, but the digital divide persists between those with and without access to technology.
计算的环境可持续性关注数据中心能耗、电子废弃物和设备的整个生命周期。虚拟化和云计算可提升效率，但享有与不享有技术接入的人群之间的数字鸿沟依然存在。
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Cancel reply