📚 Common Mistakes and Corrections for Year 12 CIE Computer Science | Year 12 CIE 计算机科学常见误区与纠正方法
Many Year 12 students following the CIE Computer Science syllabus develop misunderstandings that can cost valuable marks. This article targets the most frequent conceptual errors in topics ranging from data representation and hardware to programming and networking, and provides clear corrections with practical examples.
许多学习 CIE 计算机科学课程的 Year 12 学生都会形成一些可能导致失分的误解。本文针对数据表示、硬件、编程和网络等主题中最常见的概念性错误，并提供清晰的纠正方法和实际示例。
1. Two’s Complement Sign Extension | 二进制补码的符号扩展
A common mistake is padding a negative two’s complement number with zeros when increasing the number of bits, treating it like a positive binary value. For example, representing –5 in 8 bits as 11111011, students might incorrectly extend it to 10 bits as 0011111011.
一个常见错误是在增加二进制位数时用零填充负数的补码表示，就像处理正数一样。例如，8 位的 –5 表示为 11111011，学生可能会错误地将其扩展到 10 位变成 0011111011。
The correct method is sign extension: copy the most significant bit (the sign bit) into all new higher-order bits. So –5 in 10-bit two’s complement is 1111111011. This preserves both the sign and the magnitude. For positive numbers, zero extension works, but for negatives, sign extension is mandatory.
正确的方法是符号扩展：将最高有效位（符号位）复制到所有新增的高位。因此，10 位补码的 –5 是 1111111011。这既保留了符号也保留了数值。对于正数，零扩展可行；但对于负数，必须进行符号扩展。
Confusion also arises with the range: an n-bit two’s complement number can represent –2ⁿ⁻¹ to 2ⁿ⁻¹ – 1, not –2ⁿ⁻¹ – 1 to 2ⁿ⁻¹. This off-by-one error often appears in exam answers.
范围方面也存在混淆：n 位补码能表示的范围是 –2ⁿ⁻¹ 到 2ⁿ⁻¹ – 1，而不是 –2ⁿ⁻¹ – 1 到 2ⁿ⁻¹。这种差一错误经常出现在考试答案中。
2. Floating-Point Precision and Normalisation | 浮点精度与规格化
Students often believe that floating-point representations store real numbers exactly, or that increasing the mantissa bits always improves accuracy without considering the trade-off with the exponent range.
学生通常认为浮点表示能精确存储实数，或者增加尾数位总能提高精度，却没有考虑与指数范围的权衡。
Floating-point numbers are approximations. The precision is determined by the mantissa bits, while the range is set by the exponent bits. For example, in a 16-bit representation with a 10-bit mantissa and 6-bit exponent, the smallest positive normalised number is not zero but something like 0.1₂ × 2⁻³¹. Normalisation requires that the mantissa starts with 01 (for positive) or 10 (for negative) in binary, ensuring maximum precision. A denormalised number loses significant leading zeros.
浮点数是近似值。精度由尾数位数决定，而范围由指数位数决定。例如，在一个 10 位尾数、6 位指数的 16 位表示中，最小的正规格化数不是零，而是类似 0.1₂ × 2⁻³¹ 的值。规格化要求尾数以二进制 01（正数）或 10（负数）开头，从而保证最大精度。非规格化数会丢失有效的前导零。
Another error is forgetting that the exponent is typically stored in excess notation (bias), not as a two’s complement number. The bias for k exponent bits is 2ᵏ⁻¹ – 1. Understanding the bias is crucial for converting between stored and actual exponent values.
另一个错误是忘记指数通常以 excess 表示法（偏置）存储，而不是补码形式。k 位指数的偏置值为 2ᵏ⁻¹ – 1。理解偏置对于在存储值和实际指数值之间进行转换至关重要。
3. Logic Gate Simplification and Karnaugh Maps | 逻辑门化简与卡诺图
A persistent mistake is confusing the NAND and NOR gate outputs as being the simple inverse of AND and OR respectively, but then applying De Morgan’s laws incorrectly. For instance, students may write A NAND B as NOT A AND NOT B, which is wrong.
一个顽固的错误是将 NAND 和 NOR 门的输出分别视为 AND 和 OR 的简单取反，但随后又错误地应用德摩根定律。例如，学生可能将 A NAND B 写作 NOT A AND NOT B，这是错误的。
By De Morgan’s theorem, A NAND B = NOT A OR NOT B, and A NOR B = NOT A AND NOT B. Mistakes in these transformations ruin logic simplification. In Karnaugh maps, a frequent error is grouping cells that are not powers of two, or failing to wrap around edges (top-bottom, left-right), missing valid groupings of 1s that would lead to simpler expressions.
根据德摩根定理，A NAND B = NOT A OR NOT B，A NOR B = NOT A AND NOT B。这些转换中的错误会破坏逻辑化简。在卡诺图中，常见的错误是对非 2 的幂次个数的单元格进行分组，或者没有利用边界环绕（上下、左右），从而遗漏了可以产生更简单表达式的有效 1 的分组。
When drawing a logic circuit from a Boolean expression, students sometimes draw an OR gate where an AND gate is required, or vice versa, because they misinterpret the precedence of operators. Remember that AND has higher precedence than OR in standard Boolean algebra.
在根据布尔表达式绘制逻辑电路时，学生有时会在需要 AND 门的地方画了 OR 门，反之亦然，因为他们误解了运算符的优先级。记住，在标准布尔代数中，AND 的优先级高于 OR。
4. File Organisation and Database Normalisation | 文件组织与数据库规范化
It is easy to mix up serial, sequential, indexed sequential, and random (direct) file access methods. Students often claim that serial files allow fast record retrieval when they in fact force a linear search from the beginning.
很容易混淆串行、顺序、索引顺序和随机（直接）文件存取方法。学生常说串行文件可以快速检索记录，但实际上它强制从头开始线性搜索。
Serial: records stored in the order they arrive, with no ordering, so searching requires examining every record. Sequential: records are ordered by a key field, enabling binary search on a sorted file but still physical order matters. Indexed sequential: adds an index for faster random access while preserving sequential order. Direct/random: a hashing algorithm computes the address from the key, allowing one-step access. Each has its use case and trade-offs.
串行：记录按到达顺序存储，无排序，搜索需要检查每条记录。顺序：记录按键字段排序，可以在排序文件上进行二分查找，但物理顺序仍然重要。索引顺序：添加索引以加快随机访问速度，同时保持顺序。直接/随机：哈希算法根据键计算地址，允许一步访问。每种方法都有其应用场景和权衡。
In databases, normalisation is frequently misunderstood. The goal is not purely to split tables into as many as possible. Students often stop at First Normal Form (1NF) without resolving partial dependencies, or they try to achieve Third Normal Form (3NF) but introduce update anomalies by not identifying transitive dependencies correctly.
在数据库中，规范化经常被误解。目标并不是单纯地将表拆分成尽可能多的表。学生通常只达到第一范式（1NF），而没有解决部分依赖问题；或者他们试图达到第三范式（3NF），但由于没有正确识别传递依赖而引入了更新异常。
A typical mistake: in a table with columns (OrderID, ProductID, ProductName, CustomerName), ProductName depends on ProductID, not on the whole primary key (OrderID, ProductID), so it should be moved to a Products table. Not removing such partial dependencies violates 2NF.
一个典型错误：在包含 (OrderID, ProductID, ProductName, CustomerName) 列的表里，ProductName 依赖于 ProductID，而不是整个主键 (OrderID, ProductID)，因此它应该被移到一个 Products 表中。不移除这种部分依赖就违反了第二范式。
5. Pseudocode Loop Conditions and Variable Updates | 伪代码循环条件与变量更新
When writing pseudocode for CIE papers, many students make off-by-one errors in loops. A WHILE loop condition like WHILE count <= 10 when the intended number of iterations is 10, but they accidentally use count < 10 or start count at 0 without adjusting the end condition.
在为 CIE 卷子编写伪代码时，许多学生在循环中犯差一错误。例如，当预期迭代 10 次时，一个 WHILE 循环的条件可能是 WHILE count <= 10，但他们不小心用了 count < 10 或者将 count 初始化为 0 却没有调整结束条件。
For REPEAT…UNTIL loops, the mistake is thinking that the loop exits as soon as the condition becomes true before the end of the loop body; actually, the condition is checked only at the end of the iteration. So if the body changes a variable to make the condition true, the loop still finishes the current pass.
对于 REPEAT…UNTIL 循环，错误在于认为当条件在循环体结束前变为真时循环就会退出；实际上，条件仅在迭代结束时检查。因此，如果循环体改变了某个变量使得条件为真，该次循环仍会完成当前这一遍。
Variable updates inside loops are another source of confusion. Students sometimes forget to increment a counter inside a WHILE loop, leading to infinite loops, or they update the counter before the processing logic, skipping the first or last element of an array.
循环内部的变量更新是另一个混淆点。学生有时忘记在 WHILE 循环内增加计数器的值，导致无限循环；或者他们在处理逻辑之前更新计数器，从而跳过数组的第一个或最后一个元素。
6. Data Transmission and Error Checking | 数据传输与错误校验
There is a tendency to assume that parity checks can correct errors. Parity bits (even or odd) can only detect an odd number of bit errors. If two bits are flipped, the parity remains the same, so the error goes undetected. Moreover, parity provides no correction mechanism – ARQ (Automatic Repeat reQuest) must request retransmission.
人们倾向于认为奇偶校验可以纠正错误。奇偶校验位（偶校验或奇校验）只能检测到奇数个比特错误。如果有两个比特翻转，奇偶性保持不变，因此错误被忽略。此外，奇偶校验没有纠正机制——必须通过 ARQ（自动重传请求）请求重传。
Checksums and CRC (Cyclic Redundancy Check) are more robust, but students confuse them with hash functions. A checksum is used for error detection in communication, not for security. CRC can detect burst errors effectively, but it does not authenticate the sender.
校验和与 CRC（循环冗余校验）更健壮，但学生常将其与哈希函数混淆。校验和用于通信中的错误检测，而不是为了安全。CRC 能有效检测突发错误，但不能验证发送方身份。
In the context of serial and parallel transmission, a common error is stating that parallel is always faster. While parallel sends multiple bits simultaneously, it suffers from skew (bits arriving at different times) over long distances, making it less reliable. Serial transmission is preferred for longer distances due to fewer physical wires and simpler synchronization.
在串行和并行传输的情境下，一个常见错误是声称并行总是更快。虽然并行同时发送多个比特，但它会在长距离上受到偏差（各比特到达时间不同）的影响，使其较不可靠。串行传输因物理线缆更少和同步更简单而成为长距离通信的首选。
7. Object-Oriented Programming Concepts | 面向对象编程概念
Students frequently misuse the terms class, object, instantiation, and inheritance. A class is a blueprint; an object is an instance of a class. Instantiation is the process of creating an object from a class. However, a wrong statement like “an object inherits from another object” shows a fundamental misunderstanding: inheritance happens between classes, not objects.
学生经常误用类、对象、实例化和继承这些术语。类是蓝图；对象是类的实例。实例化是从类创建对象的过程。然而，像“一个对象从另一个对象继承”这样的错误说法显示出了一个根本性的误解：继承发生在类之间，而不是对象之间。
Polymorphism is also often defined incorrectly. It is not simply method overloading. In the context of CIE, polymorphism refers to the ability of objects of different classes to respond to the same message (method call) in their own way, typically implemented via method overriding in a class hierarchy. A typical error is to say that polymorphism means “the same method name in different unrelated classes,” which is actually just coincidence, not true polymorphic behaviour.
多态性也常常被错误定义。它不仅仅是方法重载。在 CIE 的语境中，多态性指的是不同类的对象以自己的方式响应相同的消息（方法调用），通常通过类层次结构中的方法重写来实现。一个典型错误是说多态性意味着“在不同且不相干的类中使用相同的方法名”，这实际上只是巧合，并非真正的多态行为。
8. Memory and Storage Hierarchy | 存储器与存储层次结构
A frequent misconception is confusing ROM (Read-Only Memory) with secondary storage like a hard disk. ROM is non-volatile memory used primarily to store firmware (e.g., BIOS), but it is still directly addressable by the CPU and resides in the primary memory space. In contrast, secondary storage is much slower and requires I/O operations to access.
一个常见的误解是将 ROM（只读存储器）与硬盘这样的二级存储混淆。ROM 是非易失性存储器，主要用于存储固件（如 BIOS），但它仍然可以被 CPU 直接寻址，并位于主存储器空间中。相比之下，二级存储器速度慢得多，并且需要 I/O 操作才能访问。
Also, RAM is not the only type of primary memory; cache memory is a smaller, faster type of SRAM located closer to the CPU. Students often omit cache when describing the fetch-execute cycle, forgetting that the CPU first checks cache for instructions and data before accessing RAM.
此外，RAM 并不是唯一的主存类型；高速缓冲存储器（cache）是一种更小、更快的 SRAM，位于更靠近 CPU 的位置。学生在描述取指-执行周期时常常忽略缓存，忘记了 CPU 在访问 RAM 之前会先检查缓存中是否有指令和数据。
Another point of confusion is virtual memory vs. physical RAM. Virtual memory uses a portion of the hard drive as an extension of RAM, but this is not the same as simply adding more RAM. Paging and swapping are the mechanisms; thrashing occurs when the system spends more time swapping pages than executing processes, dramatically slowing performance.
另一个混淆点是虚拟内存和物理 RAM。虚拟内存使用部分硬盘作为 RAM 的扩展，但这与简单地增加 RAM 不同。分页和交换是它的机制；当系统花费在交换页面上的时间多于执行进程的时间时，就会发生“抖动”，性能急剧下降。
9. Network Protocols and Layers | 网络协议与分层模型
It is incorrect to think that protocols at one layer can directly see or alter the data of another layer, or that encryption at the application layer makes the entire packet encrypted at every layer. Each layer encapsulates the data from the layer above. For example, TLS (Transport Layer Security) operates at the transport layer and encrypts the payload, but the TCP and IP headers remain in plaintext for routing.
错误地认为某一层的协议可以直接看到或更改另一层的数据，或者认为应用层的加密会使整个数据包在每一层都被加密，这些都是不正确的。每一层都会封装来自上一层的数据。例如，TLS（传输层安全协议）在传输层运行并加密有效载荷，但 TCP 和 IP 报头仍保持明文以进行路由。
Many students struggle to place protocols in the correct TCP/IP layers. A typical mistake is putting HTTP into the transport layer, or confusing the roles of TCP and IP. TCP (Transmission Control Protocol) provides reliable, ordered delivery; IP (Internet Protocol) handles addressing and routing.
许多学生难以将协议归入正确的 TCP/IP 层。一个典型错误是把 HTTP 归入传输层，或者混淆 TCP 和 IP 的职责。TCP（传输控制协议）提供可靠的、有序的交付；IP（互联网协议）处理寻址和路由。
Another frequent confusion: DHCP and DNS. DHCP assigns IP addresses dynamically; DNS translates domain names to IP addresses. Saying “DHCP translates domain names” is a serious error that reveals a lack of understanding of network configuration.
另一个常见混淆：DHCP 和 DNS。DHCP 动态分配 IP 地址；DNS 将域名翻译为 IP 地址。说“DHCP 翻译域名”是一个严重错误，说明对网络配置缺乏理解。
10. Binary Coded Decimal (BCD) vs. Pure Binary | 二-十进制编码 (BCD) 与纯二进制
Students often use BCD when asked for binary representation, or treat BCD addition as identical to binary addition. BCD represents each decimal digit with a 4-bit binary code (0000 to 1001). Thus, 25 in BCD is 0010 0101, not 11001 (which is the binary for 25).
学生在被问到二进制表示时经常使用 BCD，或者将 BCD 加法视为与二进制加法完全相同。BCD 用 4 位二进制码（0000 到 1001）表示每个十进制数字。因此，25 的 BCD 码是 0010 0101，而不是 11001（后者是 25 的纯二进制）。
When adding BCD numbers, if a 4-bit group exceeds 9 (1001) or generates a carry, we must add 6 (0110) to correct it to a valid BCD digit and produce the correct carry. Ignoring this correction step leads to wrong results. This is a very common source of error in exams that test BCD arithmetic.
在 BCD 加法中，如果某个 4 位组超过 9（1001）或产生进位，必须加上 6（0110）来将其修正为有效的 BCD 数字并产生正确的进位。忽略这个修正步骤会导致错误结果。这在考查 BCD 算术的考试中是一个非常常见的丢分点。
BCD is used in applications where decimal display is critical, like digital clocks or calculators, because conversion to pure binary for arithmetic and back would introduce rounding errors. Knowing why BCD is used, not just how, is essential for higher-mark questions.
BCD 被用于十进制显示至关重要的场合，比如数字钟表或计算器，因为若转换为纯二进制进行算术再转回会引入舍入误差。知道为什么使用 BCD，而不仅仅是怎样用，对于高分值题目至关重要。
Published by TutorHao | CIE 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