📚 Year 13 WJEC Computer Science: High-Frequency Topics and Common Pitfalls | Year 13 WJEC 计算机：高频考点与易错题分析
For Year 13 students following the WJEC Computer Science specification, certain topics consistently appear in examinations and regularly cause misunderstandings. This revision article identifies the most frequently tested areas across the A2 programme, from abstract data structures and recursion to database normalisation and finite state machines. Each section presents the core concepts that are often examined, followed by a careful analysis of the common mistakes candidates make and advice on how to avoid them. Whether you are preparing for the Unit 3 and Unit 4 written papers or reinforcing your practical programming skills, a secure grasp of these high-frequency topics will significantly boost your confidence and final grade.
对于正在学习 WJEC 计算机科学课程的高三学生来说，有一些考点每年都出现在试卷中，也反复让考生犯错。本文梳理了 A2 阶段最高频的考查领域——涵盖抽象数据结构、递归、数据库范式、有限状态机等主题。每个小节都先提炼出考试中常见的关键知识点，然后重点剖析大家最容易出错的地方，并给出避免失分的建议。无论你是在备战单元三和单元四的理论笔试，还是在强化编程实践技能，牢牢掌握这些高频考点都能让你的信心和最终成绩迈上一个台阶。
1. Abstract Data Types: Stacks, Queues and Linked Lists | 抽象数据类型：栈、队列与链表
The WJEC specification demands that students understand the behaviour and implementation of stacks (LIFO), queues (FIFO), and dynamic structures such as linked lists. Exams frequently ask candidates to trace operations on these structures, write pseudocode for push/pop and enqueue/dequeue, and compare the advantages of linked lists over arrays for insertions and deletions in the middle of a collection.
WJEC 考纲要求学生理解栈（后进先出）、队列（先进先出）以及链表这类动态结构的行为和实现。考试经常会让考生手工模拟对栈、队列的操作过程，编写入栈/出栈、入队/出队的伪代码，并比较链表相比数组在集合中间位置插入和删除元素时的优势。
One common mistake is forgetting to update both the head pointer and, in a linked list implementation of a queue, the tail pointer when the first element is added to an empty queue. Students also lose marks by not checking for overflow (stack/queue full) or underflow (stack/queue empty) conditions before performing operations. When writing pseudocode for linked lists, failing to maintain correct next-pointer references during insertion leads to broken chains and lost marks.
一个常见错误是忘记同时更新头指针和尾指针——在用链表实现队列时，往空队列中添加第一个元素后，必须同时修改头尾指针。学生也经常因为没有在执行操作前检查溢出（栈/队列已满）或下溢（栈/队列为空）而丢分。在编写链表伪代码时，插入操作中若未能保持正确的 next 指针引用顺序，就会导致链条断裂并因此失分。
2. Recursion and Its Pitfalls | 递归及其隐藏陷阱
Recursion is a high-frequency topic that combines algorithmic thinking with a thorough understanding of the call stack. An effective recursive solution must have a base case that is guaranteed to be reached and a recursive case that moves the problem towards that base case. Typical exam questions ask students to trace recursive functions, write recursive solutions for problems such as factorial, Fibonacci, and tree traversals, and discuss the memory implications of deep recursion.
递归是高频考点，它把算法思维与对调用栈的透彻理解结合在一起。一个有效的递归解法必须具有一个保证能够到达的基线条件，以及一个能让问题逐步逼近基线条件的递归条件。典型的考题会要求考生追踪递归函数的执行过程，编写阶乘、斐波那契数列、树的遍历等问题的递归解法，并讨论深度递归带来的内存消耗问题。
A repeatedly observed pitfall is omitting the base case altogether, which results in infinite recursion and a stack overflow error. Even when the base case is present, candidates sometimes write a recursive step that does not reduce the problem size (e.g., calling the function with the same arguments). Another subtle error arises with Fibonacci: the naive recursive version has exponential time complexity O(2ⁿ), and students fail to recognise that the repeated recalculation of subproblems can be eliminated with memoisation or iteration. In tracing exercises, confusion about the order in which stacked calls return often leads to incorrectly written output sequences.
反复出现的一个陷阱是完全遗漏基线条件，导致无限递归和堆栈溢出错误。即使写下了基线条件，有时考生写的递归步骤并未缩小问题规模（比如用相同的参数调用自身）。在斐波那契数列中还有一个隐蔽的错误：朴素递归版本具有指数级时间复杂度 O(2ⁿ)，学生常常意识不到可以通过记忆化或迭代来消除子问题的重复计算。在追踪练习中，对堆叠调用返回顺序的混淆，经常导致写出错误的输出序列。
3. Sorting Algorithms: Quicksort vs Mergesort | 排序算法：快速排序与归并排序
WJEC examinations expect candidates to be able to trace the step-by-step execution of standard sorting algorithms, analyse their time and space complexities in the best, average and worst cases, and discuss their suitability for different data sets. Among comparison-based sorts, Quicksort (average O(n log n), worst O(n²)) and Mergesort (always O(n log n) but with additional O(n) space) receive particular attention.
WJEC 考试要求考生能够逐步追踪标准排序算法的执行过程，分析它们在最好、平均和最坏情况下的时间与空间复杂度，并讨论它们对不同数据集的适用性。在基于比较的排序算法中，快速排序（平均 O(n log n)，最坏 O(n²)）和归并排序（始终 O(n log n)，但需要 O(n) 额外空间）尤其受到关注。
Mistakes frequently occur when students manually choose the pivot in Quicksort: many exam scripts show candidates using the first element as pivot without considering the impact on an already sorted list, which triggers the worst-case O(n²) behaviour. When computing the space complexity of Mergesort, candidates often forget to count the auxiliary array needed for merging, stating space complexity as O(1) instead of O(n). Another classic error is confusing the execution order of the two recursive calls in Mergesort, drawing a merge tree that merges unsorted halves before they have been recursively sorted.
常见错误出现于手动为快速排序选择基准值时：很多答卷显示考生总是选用第一个元素作为基准，而没有考虑它对已经有序的列表会造成何种影响，这将触发最坏情况的 O(n²) 行为。在计算归并排序的空间复杂度时，考生经常忘记统计合并时所需的辅助数组，错误地把空间复杂度写为 O(1)，而非 O(n)。另一个经典错误是混淆归并排序中两次递归调用的执行顺序，画出的归并树中两半子数组在尚未被递归排序之前就直接合并了。
4. Object-Oriented Programming: Inheritance and Polymorphism | 面向对象编程：继承与多态
Object-oriented concepts form a significant part of both the theory paper and the practical programming project. Students must be able to define encapsulation, inheritance and polymorphism with clear examples, explain how subclasses inherit public and protected members, and describe the mechanism of method overriding and dynamic binding. Exam questions may ask for UML class diagrams that illustrate generalisation-specialisation relationships.
面向对象的概念在理论笔试和实践编程项目中都占有重要分量。学生必须能用清晰的例子定义封装、继承和多态，解释子类如何继承 public 和 protected 成员，并描述方法重写和动态绑定的机制。考题可能会要求画出展示泛化-特化关系的 UML 类图。
A widespread error stems from confusing method overriding (redefining a superclass method in the subclass with the same signature) and method overloading (multiple methods with the same name but different parameter lists). Overloading is compile-time polymorphism, while overriding is run-time polymorphism; mixing these leads to incorrect explanations of dynamic dispatch. In diagram questions, candidates draw arrows from the superclass to the subclass, but inheritance should point the opposite way (subclass towards superclass), or they forget the hollow triangle notation entirely. Access modifiers also trip up learners: protected members are accessible within subclasses but not freely outside the hierarchy, and many scripts treat protected as identical to public.
一个普遍的错误源于混淆方法重写（在子类中用相同签名重新定义超类方法）和方法重载（多个同名但参数列表不同的方法）。重载是编译时多态，重写是运行时多态；把二者搞混就会导致对动态分派的解释不正确。在图形题中，考生有时会从超类画箭头指向子类，但继承关系应该方向相反（子类指向超类），或者完全忘记空心三角形符号。访问修饰符也让许多学生栽跟头：protected 成员可以在子类中访问，但不能在继承层次之外自由使用，而很多答卷把 protected 等同于 public 来对待。
5. Assembly Language and Machine Code | 汇编语言与机器码
At Year 13, students extend their understanding of the fetch-decode-execute cycle into low-level programming with the WJEC assembly language instruction set. Typical tasks include writing short assembly programs that perform arithmetic, use branching (CMP, B, BEQ, BNE, BLT) and employ indexed addressing for array traversal. The relationship between assembly mnemonics, opcodes, and the underlying machine code is often tested through trace-table exercises.
在高三阶段，学生会将取指-解码-执行周期的理解延伸到低层编程，使用 WJEC 汇编语言指令集。典型任务包括编写完成算术运算、使用分支指令（CMP、B、BEQ、BNE、BLT）以及利用变址寻址遍历数组的简短汇编程序。汇编助记符、操作码与底层机器码之间的关系常常通过跟踪表练习来考查。
two equal numbers sets the zero flag (Z=1), but many candidates forget this and incorrectly predict the outcome of a subsequent BEQ branch. When writing assembly, students often confuse immediate addressing (using # lue) and direct addressing (using a memory location), which completely changes the program' ng to a subroutine, making it impossible to resume execution after the subroutine returns.
在跟踪表题目中，一个常犯的错误是执行 CMP 指令后没有恰当地更新状态标志。例如，两个相等的数相减会置零标志位 (Z=1)，但许多考生会忘记这一点，从而错误地预测随后 BEQ 分支的行为。在编写汇编程序时，学生经常混淆立即寻址（用 # 表示常数值）与直接寻址（使用内存地址），这会彻底改变程序的行为。另一个陷阱是跳转到子程序前没有保存返回地址，导致子程序返回后无法继续恢复执行。
6. Database Normalisation to Third Normal Form | 数据库范式化至第三范式
Normalisation is a persistently examined topic. Candidates are expected to work through a given set of unnormalised data, remove repeating groups to achieve 1NF, eliminate partial dependencies for 2NF, and then remove transitive dependencies to reach 3NF. They must justify each decomposition step by identifying the primary key and the dependencies that violate the normal form in question.
数据库范式化是一个经久不衰的考点。考生需要针对给定的未范式化数据集，通过去除重复组来达到第一范式，消除部分函数依赖以达到第二范式，再消除传递依赖以达到第三范式。他们必须在每一步分解中，明确指出主键以及违反当前范式的依赖关系，来说明分解的理由。
The most common conceptual error is misidentifying the primary key and thus misjudging partial dependencies. A partial dependency exists when a non-key attribute depends on only part of a composite primary key, not the whole key. Students often erroneously treat any attribute that is not the primary key as transitively dependent on the key, leading to premature decomposition. In 3NF, confusion between a true transitive dependency (non-key attribute depends on another non-key attribute) and a direct dependency on the primary key results in splitting tables that are already in 3NF and losing marks for over-normalisation. Numerous scripts also present decomposed tables without clearly showing the relations with their primary and foreign keys underlined.
最常见的概念性错误是错误地识别主键，从而导致误判部分函数依赖。部分函数依赖是指某个非键属性仅依赖于组合主键的一部分，而非整个主键。学生经常错误地把任何不是主键的属性都当作对主键的传递依赖，从而导致不成熟的分解。在第三范式阶段，混淆真正的传递依赖（非键属性依赖于另一个非键属性）与对主键的直接依赖，会导致拆分本来已经处于 3NF 的表，因过度范式化而失分。大量答卷还在呈现分解后的表时，没有清晰地标出下划线以示主键和外键。
7. Regular Expressions and Finite State Machines | 正则表达式与有限状态机
The theory of computation segment in the WJEC A2 specification covers regular expressions, deterministic and non-deterministic finite automata (DFA and NFA), and their equivalence. Exam questions may ask students to derive a regular expression from a given FSM, design a state transition diagram to recognise a particular language, or convert between a regular expression and a DFA.
WJEC A2 规范中的计算理论部分涵盖了正则表达式、确定性与非确定性有限自动机（DFA 和 NFA）以及它们的等价性。考题可能会要求学生从给定的 FSM 中推导出正则表达式，设计状态转换图以识别某一特定语言，或者在正则表达式和 DFA 之间进行转换。
A typical source of error is handling the empty string in regular expressions. When translating an FSM that accepts the empty string, candidates frequently forget that the initial state must also be marked as an accepting state, or they write a regular expression that fails to match ε. In state diagram questions, missing transitions cause the automaton to become stuck for valid strings, creating a non-deterministic behaviour that is often not intended. Another area where marks are lost is in the conversion from NFA to DFA: the state explosion is sometimes mishandled, with candidates drawing incomplete compound states, thereby losing the language accepted by the original NFA.
一个典型错误来源是在正则表达式中处理空串。当转换一个接受空串的 FSM 时，考生经常忘记还必须将初始状态也标记为接受状态，或者写出无法匹配 ε 的正则表达式。在状态图题目中，缺失的转移会导致自动机在遇到有效字符串时卡住，从而产生一种通常并非有意的非确定性行为。另一个容易失分的领域是从 NFA 到 DFA 的转换：状态爆炸有时被不当处理，考生画出的复合状态不完整，因而丧失了原 NFA 所接受的语言。
8. Networking: TCP/IP Protocol Suite and Network Security | 网络：TCP/IP 协议族与网络安全
The networking topic in WJEC requires detailed knowledge of the TCP/IP four-layer model (Application, Transport, Internet, Network Access), the roles of key protocols such as HTTP, HTTPS, FTP, SMTP, TCP, UDP, IP, and the secure alternatives that use encryption. Students should also be able to explain the handshake process in TCP, the purpose of the three-way handshake, and the difference between packet switching and circuit switching.
WJEC 的网络主题要求详细了解 TCP/IP 四层模型（应用层、传输层、互联网层、网络接入层），掌握关键协议（如 HTTP、HTTPS、FTP、SMTP、TCP、UDP、IP）的作用及其使用加密的安全版本。学生还应能够解释 TCP 的握手过程、三次握手的目的，以及包交换与电路交换之间的区别。
One of the most persistent errors is misplacing protocols into the wrong layers. For example, placing HTTP in the Transport layer or TCP in the Application layer reveals a fundamental misunderstanding of encapsulation and the service model. Another common slip concerns the TCP three-way handshake: candidates memorise SYN, SYN-ACK, ACK but fail to explain that the sequence numbers are synchronised and that this process establishes a full-duplex reliable connection. In security questions, confusing the roles of symmetric and asymmetric encryption in protocols like HTTPS (asymmetric for key exchange, symmetric for bulk data) is a frequent cause of lost marks. Network Access layer details such as MAC addresses and the role of switches versus routers are also commonly reversed.
最顽固的错误之一是把协议放到了错误的层次中。例如，把 HTTP 放在传输层，或把 TCP 放在应用层，暴露出对封装和服务模型的根本性误解。另一个常见疏漏与 TCP 三次握手有关：考生能够机械地背出 SYN、SYN-ACK、ACK，却没有解释序列号是如何同步的，以及这一过程如何建立起全双工的可靠连接。在安全问题中，混淆对称和非对称加密在 HTTPS 等协议中的角色（非对称用于密钥交换，对称用于批量数据加密）是导致失分的常见原因。网络接入层的细节，如 MAC 地址以及交换机和路由器各自的作用，也经常被弄反。
9. Algorithm Design: Greedy, Divide-and-Conquer, and Dynamic Programming | 算法设计：贪心、分治与动态规划
WJEC Unit 3 and the practical non-exam assessment emphasise algorithm design paradigms. Students need to recognise scenarios where a greedy approach may fail to produce an optimal solution (e.g., the knapsack problem), understand how divide-and-conquer algorithms such as binary search work, and appreciate when overlapping subproblems warrant a dynamic programming approach. They must be able to compare the computational complexity of different strategies for the same problem.
WJEC 单元三和课外的实践评估都强调算法设计范式。学生需要识别贪心法可能无法产生最优解的场景（例如背包问题），理解像二分查找这样的分治算法是如何工作的，并判断何时由于重叠子问题而应使用动态规划。他们必须能够比较不同策略针对同一问题的计算复杂度。
A common analytical error is claiming that a greedy algorithm always finds the optimal solution. While greedy strategies are optimal for problems that exhibit optimal substructure and the greedy-choice property (like Dijkstra's shortest path with non-negative weights), applying them to the coin change problem with certain denominations can yield suboptimal results. In dynamic programming, students often construct a table but fill it in the wrong order, resulting in uninitialised values being used. Another mistake is failing to reconstruct the actual solution from a DP table, providing only the optimal value without the corresponding sequence of decisions. With divide-and-conquer, many scripts misjudge the base case in binary search, causing off-by-one errors or infinite loops when the target is at the boundaries of the search space.
一个常见的分析错误是声称贪心算法总能找到最优解。虽然贪心策略对于展现出最优子结构和贪心选择性质的问题（如带有非负权重的 Dijkstra 最短路径）是最优的，但将其应用于某些面额的硬币找零问题却可能得出次优结果。在动态规划中，学生往往构建出了表格，却以错误的顺序填充，导致使用了未初始化的值。另一个错误是未能从 DP 表格中重构出实际的解决方案，只给出了最优值，而缺少相应的决策序列。在分治算法中，很多答卷误判了二分查找的基线条件，当目标值位于搜索空间边界时，会导致“差一”错误或无限循环。
10. System Software: OS Roles and Compiler Stages | 系统软件：操作系统的角色与编译器阶段
Students are examined on the purposes and functions of operating systems (memory management, process scheduling, I/O management, file systems, and security) and the stages of compilation: lexical analysis, syntax analysis, semantic analysis, code generation and code optimisation. They should be able to explain how syntax analysis builds a parse tree or abstract syntax tree, and identify the types of errors detected at each stage.
考试会考查操作系统的目的和功能（内存管理、进程调度、I/O 管理、文件系统和安全）以及编译过程的各个阶段：词法分析、语法分析、语义分析、代码生成和代码优化。学生应能解释语法分析如何构建语法分析树或抽象语法树，并识别各阶段检测出的错误类型。
Confusion between the roles of the linker and loader is a frequent mark-loser: linking combines object code with library routines to produce an executable, while loading transfers the executable into memory and prepares it for execution. In the compilation pipeline, students often attribute type incompatibility errors to the syntax analysis phase instead of semantic analysis. Another lingering mistake is mixing up virtual memory with the concept of RAM expansion; virtual memory uses a portion of secondary storage to extend the apparent size of main memory but relies on paging or segmentation, and its overuse results in thrashing, not a larger physical RAM.
混淆链接器和加载器的角色是经常丢分的地方：链接是将目标代码与库例程组合以生成可执行文件，而加载则是将可执行文件移入内存并准备执行。在编译流水线中，学生往往将类型不兼容错误归因于语法分析阶段，而非语义分析。另一个挥之不去的错误是把虚拟内存与“扩充 RAM”的概念混为一谈；虚拟内存使用一部分外存来扩大主存的表观大小，但它依赖于分页或分段机制，过度使用会导致系统的颠簸，而不是获得更大的物理内存。
11. Binary Representation: Floating Point and Two's Complement | 二进制表示：浮点数与补码
Number representation questions are deceptively simple and appear regularly in WJEC papers. Candidates must be able to convert between signed and unsigned binary representations using two's complement, perform normalisation of binary floating-point numbers, and calculate the range and precision of floating-point representations given the number of bits for the mantissa and exponent.
数的表示类题目看似简单，却在 WJEC 试卷中频繁出现。考生必须能使用补码在带符号和无符号二进制表示之间进行转换，执行二进制浮点数的规格化，并在给定尾数和指数位数的情况下计算浮点表示的范围和精度。
With two's complement, a classic error involves sign extension: when converting a negative byte to a word-length representation, students forget to fill the higher-order bits with 1s, effectively changing the number's value. In normalisation of binary floating-point, the rule that the mantissa should have a leading 1 (for positive) or leading 0 followed by a 1 (for negative) is frequently applied incorrectly. Candidates may shift the mantissa left to normalise but forget to decrement the exponent accordingly, or they attempt to normalise a mantissa that is already in normal form, leading to unnecessary manipulation. In range calculations, the difference between the most negative and most positive representable values is often miscalculated because students neglect the representation of zero and the hidden bit in certain floating-point formats.
在补码方面，一个经典错误涉及符号扩展：当把一个负数字节转换为字长表示时，学生忘记用 1 填充高位，这实际上改变了该数的值。在二进制浮点数规格化中，尾数应具有前导 1（正数）或前导 0 后跟 1（负数）的规则常常被误用。考生可能会将尾数左移以进行规格化，却忘记相应地递减指数，或者试图对已经处于规格化形式的尾数再次规格化，导致不必要的操作。在范围计算中，最大负值与最大正值之间的差值经常被算错，原因是学生忽略了零的表示以及某些浮点格式中的隐藏位。
12. Exam Technique and Common Pitfalls to Avoid | 考试技巧与需要避免的常见失分点
Beyond content knowledge, many marks are lost through avoidable exam-technique errors. These include not reading the question command words carefully (e.g., 'describe' versus 'explain'), failing to show working in computational questions, and writing answers that do not address the exact mark allocation. In programming tasks, incomplete initialisation of variables and missing bounds checks on arrays consistently reduce scores across the cohort.
除了知识本身，许多分数是因本可避免的考试技巧错误而丢掉的。这些错误包括没有仔细读懂题目要求的关键指令词（如“描述”与“解释”的区别），在计算题中不展示计算步骤，以及写出的答案未能对应具体的分值分配。在编程类任务中，变量初始化不全和缺失数组边界检查是整届考生普遍丢分的地方。
A particularly widespread pitfall is attempting to answer the question the candidate wishes had been asked, rather than the one actually on the paper. Under time pressure, students often skim-read, latch on to a single keyword, and produce a well-rehearsed answer that does not fit the specific scenario provided. In trace-table and walkthrough exercises, missing a single step cascades into an entire block of lost marks. To mitigate this, practise methodical, line-by-line tracing and never skip intermediate states, even if they seem obvious. For essay-style questions, plan your answer with bullet points in the margin before writing, ensuring each point directly answers the question stem.
一个特别普遍的陷阱是，考生试图回答“自己希望被问到的问题”，而不是试卷上真正的问题。在时间压力下，学生会快速浏览，抓住一个关键词，然后写出一段滚瓜烂熟的答案，但这段答案并不符合题目给出的具体情景。在跟踪表和人工走查类的练习中，只要漏掉一步就会导致整块分数丢失。要避免这种情况，就需要练习系统化、逐行的跟踪，永远不要跳过中间状态，哪怕它们看起来显而易见。对于论述类问题，在落笔前用旁注要点的方式规划答案，确保每一个要点都直接回应了题干的要求。
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