📚 A-Level WJEC Computer Science: Common Misconceptions | A-Level WJEC计算机科学：常见误区
In WJEC A-Level Computer Science, many topics appear straightforward but hide subtle traps that can cost marks. This article dissects ten common misconceptions, explains the correct underlying principles, and provides concise examples to help you avoid mistakes in the exam. Each section pairs an English explanation with its Chinese equivalent so you can reinforce your understanding bilingually.
在WJEC A-Level计算机科学中，许多主题看似简单却隐藏着导致失分的陷阱。本文剖析十个常见误区，解释正确的底层原理，并给出简明示例，帮助你避免考试中的错误。每个小节均采用中英对照，让你能双语强化理解。
1. Binary and Two’s Complement | 二进制与补码
Many students assume that to find the decimal value of a two’s complement negative number they simply invert the bits and add 1, then attach a minus sign. This works for obtaining the magnitude, but it does not explain the true weighting. In an n-bit two’s complement representation, the most significant bit (MSB) has a negative weight of -2n-1. For an 8-bit pattern, the MSB is worth -128, while the remaining bits contribute positive values as in standard binary. Treating the MSB as a sign flag without understanding the weighting leads to errors when adding or subtracting directly in binary.
许多学生认为，要找出一个补码负数的十进制值，只需将位取反加1，然后加上负号。这种方法能得到绝对值，但无法解释真正的权重。在n位补码表示中，最高有效位(MSB)具有负权重 -2n-1。对于8位模式，MSB值-128，其余位如同标准二进制一样贡献正值。不理解权重仅把MSB当作符号标志，会导致直接进行二进制加减时的错误。
A second pitfall is confusing arithmetic shifts with logical shifts. An arithmetic right shift preserves the sign bit (copies the MSB), thus implementing division by 2 with rounding towards negative infinity for negative numbers. A logical right shift fills the vacated positions with zeros, which is appropriate for unsigned integers but destroys the value of a signed two’s complement number. In the exam, always check whether the data is signed and select the correct shift operation.
第二个陷阱是混淆算术移位与逻辑移位。算术右移保留符号位（复制MSB），因此对负数实现除以2并向负无穷方向舍入。逻辑右移用零填充空出的位，这适用于无符号整数，但会破坏有符号补码数的值。考试中务必检查数据是否有符号，并选择正确的移位操作。
Decimal value = -b7 × 27 + b6 × 26 + b5 × 25 + … + b0 × 20
2. Floating Point Representation | 浮点数表示
Students often believe that floating-point numbers can represent real numbers exactly. In reality, binary floating-point formats (e.g., IEEE 754 single precision) store numbers as sign × mantissa × 2exponent – bias. The mantissa is normalised so that the leading 1 is implicitly assumed, which means only a fixed number of fractional bits are stored. Many decimal fractions, such as 0.1, become recurring binary fractions and cannot be represented precisely, leading to rounding errors. Understanding normalisation and the hidden bit is essential for predicting the range and precision limitations.
学生常误以为浮点数能精确表示实数。实际上，二进制浮点格式（如IEEE 754单精度）将数字存储为 符号 × 尾数 × 2指数 – 偏置。尾数被规范化，隐含前导1，这意味着只存储固定数量的分数位。许多十进制小数，如0.1，会变成二进制循环小数，无法精确表示，从而导致舍入误差。理解规范化和隐含位对于预测范围和精度限制至关重要。
Another common error involves the exponent bias. In single precision, the bias is 127, so an exponent field of 120 actually represents 120 – 127 = -7. Candidates sometimes subtract the wrong bias or forget that the exponent is stored in excess notation. When converting back to denary, always apply the correct bias and account for the implicit 1 in the mantissa.
另一个常见错误涉及指数偏置。单精度中偏置为127，因此指数字段120实际代表120 – 127 = -7。考生有时减去错误的偏置或忘记指数是以移码方式存储的。当转换回十进制时，始终应用正确的偏置并考虑尾数中隐含的1。
3. Logic Gates and Boolean Algebra | 逻辑门与布尔代数
A widespread misconception is that NAND and NOR gates are just secondary combinations, not universal gates. In fact, both NAND and NOR are functionally complete, meaning any Boolean function can be implemented using only NAND gates (or only NOR gates). For instance, a NOT gate is a NAND with both inputs tied together; an AND is a NAND followed by a NOT. Recognising this universality is frequently tested in circuit simplification questions.
一个普遍的误区是认为NAND和NOR门仅是次级组合，不是通用门。实际上，NAND和NOR都是功能完备的，即仅用NAND门（或仅用NOR门）就能实现任何布尔函数。例如，将NAND的两个输入连接就构成了非门；AND门可由一个NAND后接一个非门组成。认识到这种通用性经常在电路简化题中被考察。
Boolean algebra errors often stem from misapplying De Morgan’s theorems. The correct forms are (A · B)’ = A’ + B’ and (A + B)’ = A’ · B’. Students frequently miss the change of operator and the inversion of individual variables. When simplifying expressions like (A’ + B’)’, it is tempting to write A + B directly, but careful stepwise application of De Morgan avoids mistakes. Also, the XOR gate (A ⊕ B = A’B + AB’) is sometimes confused with XNOR; note the equivalence operator.
布尔代数错误常源于错误应用德摩根定理。正确形式为 (A · B)’ = A’ + B’ 和 (A + B)’ = A’ · B’。学生常漏掉运算符的改变和单个变量的取反。化简如 (A’ + B’)’ 的表达式时，直接写 A + B 似乎诱人，但逐步应用德摩根定理可避免错误。此外，异或门 (A⊕B = A’B + AB’) 有时与同或门混淆；注意等价运算符。
4. Data Structures: Arrays vs Linked Lists | 数据结构：数组与链表
It is common to think that arrays are always faster than linked lists because of contiguous memory. While arrays provide O(1) random access, insertion and deletion at arbitrary positions require shifting elements, costing O(n) time. A linked list, by contrast, allows O(1) insertion/deletion once the position is known (via a pointer), but searching for an element is O(n) since it must be traversed sequentially. Deciding which structure to use must consider the dominant operations.
人们普遍认为数组因内存连续而总比链表快。虽然数组提供O(1)的随机访问，但在任意位置插入和删除需要移动元素，耗时O(n)。相反地，链表一旦通过指针已知位置，插入/删除为O(1)，但搜索元素必须顺序遍历，时间复杂度为O(n)。选择哪种结构必须考虑主要操作。
Another misunderstanding concerns dynamic arrays. A dynamic array (like an ArrayList) appears to grow without penalty, but resizing involves allocating a new larger array and copying all elements, an occasional O(n) cost that is amortised over many insertions. Students often overlook this hidden cost when comparing with linked lists. Also, the memory overhead of storing pointers in a linked list means arrays can be more memory-efficient for a fixed number of elements.
另一个误解涉及动态数组。动态数组（如ArrayList）似乎可以无代价地增长，但扩容涉及分配一个更大的新数组并复制所有元素，这是偶尔出现的O(n)成本，平摊到多次插入上。学生在与链表比较时常常忽略这种隐藏成本。此外，链表中存储指针的内存开销意味着对于固定数量元素，数组可能更节省内存。
5. Sorting and Searching Algorithms | 排序与查找算法
Some learners assume bubble sort is completely useless and always O(n²). Bubble sort actually has a best-case time complexity of O(n) when the input is already sorted (with a flag to detect no swaps). This characteristic makes it useful where data is almost sorted. However, insertion sort generally outperforms bubble sort in practice for small or partially sorted data, and its best case is also O(n). Selection sort always performs O(n²) comparisons regardless of input, making it inefficient on any data.
有些学生认为冒泡排序完全无用且总是O(n²)。实际上，当输入已排序时（通过标志检测无交换），冒泡排序的最佳情况时间复杂度为O(n)。这一特性使它在数据近乎有序时有用。然而，插入排序对于小型或部分有序数据通常比冒泡排序表现更佳，其最佳情况也是O(n)。选择排序在任何输入下都进行O(n²)次比较，因此在任何数据上效率都低。
Binary search is often overestimated. It requires a sorted array and random access; its O(log n) time is excellent, but the initial sorting cost must be considered. If the data is unsorted and searches are infrequent, linear search O(n) may be more appropriate. Another nuance is that binary search can be implemented iteratively to avoid stack overhead, and the mid calculation should use mid = low + (high - low)/2 to prevent overflow in fixed-bit environments.
二分查找常被高估。它需要有序数组和随机访问；其O(log n)时间确实优秀，但必须考虑初始排序成本。如果数据无序且搜索不频繁，线性查找O(n)可能更合适。另一个细节是二分查找可以迭代实现以避免栈开销，且中值计算应使用 mid = low + (high - low)/2 以防止固定位环境下的溢出。
6. Object-Oriented Programming Concepts | 面向对象编程概念
Encapsulation is frequently reduced to simply making attributes private and providing getters/setters. True encapsulation hides internal state and behaviour so that objects interact through well-defined interfaces, allowing internal changes without affecting external code. If setters just blindly assign values, they break encapsulation by exposing the internal data structure. Always validate and apply business logic inside methods.
封装经常被简化为仅将属性设为私有并提供getter/setter。真正的封装隐藏内部状态和行为，使对象通过明确定义的接口交互，允许内部变更不影响外部代码。如果setter只是盲目赋值，它们通过暴露内部数据结构破坏了封装。务必在方法内部验证并应用业务逻辑。
Inheritance is mistaken for code reuse alone. In OOP, inheritance establishes an “is-a” relationship. Overusing inheritance for code sharing can lead to deep, rigid hierarchies. Composition (“has-a”) is often preferable because it offers greater flexibility. Another error lies in confusing method overloading (compile-time polymorphism) with overriding (runtime polymorphism). The latter relies on dynamic binding, where the actual object’s method is called regardless of the reference type. Missing the @Override annotation or incorrect signatures can reintroduce subtle bugs.
继承被误解为只是代码复用。在面向对象编程中，继承建立的是”是一个“关系。过度使用继承进行代码共享会导致深层、僵化的层次结构。组合（“有一个”）通常更优，因为它提供更大灵活性。另一个错误在于混淆方法重载（编译时多态）与重写（运行时多态）。后者依赖动态绑定，即调用的是实际对象的方法，与引用类型无关。缺少@Override注解或签名不正确都可能重新引入细微的bug。
7. Networking: IP Addressing and Protocols | 网络：IP地址与协议
A classic misunderstanding is that an IP address uniquely identifies a device forever. IP addresses are often dynamically assigned via DHCP and can change. Moreover, a device can have multiple interfaces, each with its own IP. The concept of a subnet mask is also problematic: it is not simply a string of 255s. The mask defines the network and host portions; the logical AND between the IP and mask yields the network address. Errors occur when students calculate the number of usable hosts as 2host bits rather than 2host bits – 2, forgetting the network and broadcast addresses.
一个经典的误解是认为IP地址永远唯一标识一个设备。IP地址通常通过DHCP动态分配且可能变化。而且，一个设备可以有多个接口，每个接口有自己的IP。子网掩码的概念也常出问题：它不只是一串255。掩码定义网络部分和主机部分；IP和掩码的逻辑与得出网络地址。当学生计算可用主机数为2主机位数而不是2主机位数 – 2时，就忘了网络地址和广播地址。
Regarding the transport layer, many candidates assume UDP provides reliable communication because it appears simpler. UDP is connectionless and offers no guarantees of delivery, order, or error correction. It is suitable for real-time applications where speed matters more than reliability. TCP, conversely, uses a three-way handshake, acknowledgements, and sequence numbers to ensure reliable, ordered delivery. Confusing the responsibilities of layers in the OSI model (e.g., thinking that routing occurs at the data link layer) is another frequent mistake.
关于传输层，许多考生因为UDP看似更简单就认为它提供可靠通信。UDP是无连接的，不保证传递、顺序或纠错。它适用于速度重于可靠性的实时应用。相反地，TCP使用三次握手、确认和序列号来确保可靠的有序传递。混淆OSI模型中各层的职责（例如认为路由发生在数据链路层）是另一个常见错误。
8. Operating Systems: Processes and Memory | 操作系统：进程与内存
It is common to equate a process with a program. A program is a passive set of instructions on disk; a process is an active instance in memory, with its own address space, registers, and execution context. Threads are lightweight units within a process that share the same address space, making inter-thread communication efficient but also introducing concurrency issues such as race conditions.
常见误區是将进程等同于程序。程序是磁盘上被动的指令集；进程是内存中活跃的实例，拥有自己的地址空间、寄存器和执行上下文。线程是进程内的轻量级单元，共享相同地址空间，使得线程间通信高效但也引入了竞争条件等并发问题。
Virtual memory is often described as “using disk as RAM”, but that oversimplification leads to the misconception that adding RAM eliminates paging. Virtual memory extends the logical address space by mapping virtual pages to physical frames, with inactive pages stored on disk. When a needed page is not in physical memory, a page fault occurs, and the page is loaded from disk. Even with abundant RAM, page faults can occur during initial loading. Scheduling policies (Round Robin, priority-based) and disk scheduling are also areas where students misjudge the effects of starvation and overhead.
虚拟内存常被描述为“把磁盘当RAM用”，但这种过度简化导致一种误解：增加RAM就能消除分页。虚拟内存通过将虚拟页面映射到物理帧来扩展逻辑地址空间，不活跃的页面存储在磁盘上。当所需的页不在物理内存时，发生缺页错误，页从磁盘加载。即便有充足的RAM，初始加载时仍可能出现缺页。调度策略（轮转、基于优先级）和磁盘调度也是学生误判饥饿和开销影响的领域。
9. Programming: Parameters and Scope | 编程：参数与作用域
A persistent error concerns pass-by-value versus pass-by-reference. In languages like Java, all parameters are passed by value, but when an object reference is passed, the value copied is the reference (memory address). This means the method can modify the object’s state, but if it reassigns the reference to a new object, the caller’s reference remains unchanged. True pass-by-reference (as in C++ with &) would allow the method to change the caller’s reference itself. Exam questions often test this distinction with code traces.
一个顽固的错误涉及传值与传引用。在Java等语言中，所有参数都是按值传递的，但当传递对象引用时，复制的是引用的值（内存地址）。这意味着方法可以修改对象的状态，但如果将引用重新指向新对象，调用者的引用保持不变。真正的传引用（如C++中的&）允许方法更改调用者的引用本身。考题常通过代码追踪来测试这一区别。
Scope and lifetime are mixed up as well. Variables declared inside a block have local scope, limiting their visibility to that block, but their lifetime may extend if captured by closures or if they are static. Global variables (or class-level fields) have broader visibility but should be used sparingly to avoid side effects. Constants declared with final (or const in some languages) prevent reassignment, but the underlying object can still be mutated if it is mutable – a subtle but important point.
作用域和生命周期也常被混淆。块内声明的变量具有局部作用域，限其可见性于该块内，但若被闭包捕获或是静态的，其生命周期可能延长。全局变量（或类级别字段）具有更广可见性，但应谨慎使用以避免副作用。用final（或某些语言中的const）声明的常量阻止重新赋值，但如果底层对象是可变的，对象本身仍可改变——这是一个细微但重要的点。
10. Algorithm Complexity and Big O | 算法复杂度与大O表示法
One of the most harmful oversimplifications is claiming “O(1) is always the fastest”. Big O notation describes the growth rate of an algorithm’s time or space requirements as the input size n increases, ignoring constants and lower-order terms. An O(n) algorithm with a tiny constant factor can outperform an O(1) algorithm with a huge constant overhead for all practical n. The notation is asymptotic, useful for comparing scalability, not raw speed.
最有害的过度简化之一是声称“O(1)总是最快”。大O表示法描述当输入规模n增大时，算法时间或空间需求增长的速率，忽略了常数和低阶项。对于所有实际n值，一个常数因子极小的O(n)算法可能优于一个常数开销巨大的O(1)算法。该表示法是渐进的，用于比较可扩展性，而非原始速度。
Another frequent blunder is misidentifying nested loop complexity. Two nested loops each iterating n times result in O(n²). If the inner loop executes log n times, the product becomes O(n log n). But when one loop depends on a different variable m, the complexity is O(n*m). Students also forget that recursive algorithms require solving a recurrence; for example, the merge sort recurrence T(n) = 2T(n/2) + n solves to O(n log n). Space complexity matters equally, especially for recursive functions that use stack space proportional to recursion depth.
另一个常见失误是错误识别嵌套循环的复杂度。两个各迭代n次的嵌套循环结果是O(n²)。如果内层循环执行log n次，乘积变为O(n log n)。但当一个循环依赖于不同的变量m时，复杂度为O(n*m)。学生还忘记递归算法需要求解递推式；例如归并排序递推式T(n)=2T(n/2)+n解得O(n log n)。空间复杂度同样重要，特别是递归函数使用的栈空间与递归深度成正比。
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