AS CAIE Computer Science: High-Frequency Topics and Common Mistakes | AS CAIE 计算机：高频考点与易错题分析

📚 AS CAIE Computer Science: High-Frequency Topics and Common Mistakes | AS CAIE 计算机：高频考点与易错题分析

To score well on the AS CAIE Computer Science 9618 paper, it is not enough to simply memorise facts. You need to understand exactly where marks are gained and lost. This article walks through the high-frequency topics that appear in nearly every examination session and highlights the common mistakes that even well-prepared candidates make. Use it as a checklist alongside your past-paper practice.

要想在 AS CAIE 计算机科学（9618）考试中取得高分，光靠死记硬背远远不够。你必须清楚地知道分数从何而来、又丢在何处。本文梳理了几乎每场考试都会出现的高频考点，并重点指出即便准备充分的考生也常犯的错误。请将其当作一份备考清单，配合历年真题练习使用。

1. Data Representation: Binary and Hexadecimal Conversions | 数据表示：二进制与十六进制转换

The ability to convert rapidly between binary, denary and hexadecimal is tested in every paper, often in the first question. You must also handle binary addition and subtraction accurately. A common error is forgetting to add binary digits in groups of four when converting to hexadecimal, especially when the most significant group has fewer than four bits. Practise padding leftmost bits with zeros until you do it automatically.

每次考试几乎都会考查二进制、十进制和十六进制之间的快速转换，往往出现在第一题。考生还必须准确完成二进制加法与减法。一个常见错误是转换到十六进制时忘记按四位一组分组，尤其是最高有效组不足四位时。务必养成左侧补零的习惯，直到能够不假思索地完成。

Another subtle pitfall is binary subtraction using the borrowing method when the minuend digit is 0. Candidates frequently borrow from a non-adjacent column incorrectly, leading to cascading errors. Drawing a clear table and checking with denary validation can save marks.

另一个细微的陷阱是在被减数位为 0 时采用借位减法。考生经常错误地从非相邻列借位，导致连锁错误。画一张清晰的借位表格并用十进制验算，能有效保住分数。

2. Two’s Complement and Overflow | 补码与溢出

Two’s complement is the standard method for representing negative integers in binary. Candidates must know how to form the two’s complement of a binary number (invert all bits and add 1₂) and how to recognise overflow. A very common mistake is confusing the sign bit with the magnitude when interpreting a two’s complement number. For an 8-bit representation, the range is -128 to +127; any result outside this band indicates overflow.

补码是表示负整数的标准方法。考生必须掌握如何求一个二进制数的补码（按位取反后加 1₂），以及如何识别溢出。一个很常见的错误是在解读补码数时将符号位与数值部分混淆。在 8 位表示下，取值范围是 -128 到 +127；任何超出该范围的结果都表明发生了溢出。

Many candidates mishandle the addition of two negative numbers. For example, adding 10001010₂ (-118) and 10010100₂ (-108) yields 1 00011110₂. The extra bit is discarded, giving 00011110₂ (+30), which is clearly wrong. Overflow has occurred. The rule: overflow happens when carry into the sign bit and carry out of the sign bit are different. This is a favourite examiner trick.

许多考生处理两个负数相加时容易出错。例如，将 10001010₂ (-118) 与 10010100₂ (-108) 相加得到 1 00011110₂，多余位被丢弃后剩下 00011110₂ (+30)，显然错误。此时发生了溢出。判断规则是：进入符号位的进位与离开符号位的进位不同时，溢出发生。这是考官最爱考的点。

3. Normalised Floating-Point Representation | 规格化浮点表示

AS CAIE often asks you to normalise a floating-point number given a certain mantissa and exponent size. The key rule is that the mantissa must start with 0.1 for positive numbers and 1.0 for negative numbers. A typical mistake is stopping after shifting the binary point without adjusting the exponent accordingly, or shifting in the wrong direction. If you move the point left, increase the exponent; if you move it right, decrease the exponent.

AS CAIE 常要求考生对给定尾数和指数长度的浮点数进行规格化。核心规则是：对于正数，尾数必须以 0.1 开头；负数必须以 1.0 开头。常见的错误是移动了二进制小数点却没有相应调整指数，或者移动方向搞反。记住：小数点左移则指数增加，右移则指数减小。

Also, candidates sometimes forget that the exponent itself is stored in two’s complement. When calculating the actual exponent value, you must decode it correctly. Practice converting the biased or two’s complement exponent back to denary, then applying it to the mantissa shift.

此外，考生有时会忘记指数本身以补码形式存储。计算实际指数值时，必须正确译码。练习将偏移量表示的或补码形式的指数转换回十进制，再应用到尾数的移位上。

4. Logic Gates and Boolean Algebra | 逻辑门与布尔代数

Exam questions frequently require you to draw logic circuits from Boolean expressions, complete truth tables, or simplify expressions using laws like De Morgan’s theorem, distribution, and absorption. A high-frequency error is misapplying De Morgan’s laws when breaking the bar over a product or sum. Remember: (A · B)’ = A’ + B’ and (A + B)’ = A’ · B’.

考题经常要求根据布尔表达式画出逻辑电路、补全真值表或运用德摩根律、分配律、吸收律等规则化简表达式。一个高频错误是在拆解与或式的非号时错误使用德摩根律。请记住：(A · B)’ = A’ + B’，而 (A + B)’ = A’ · B’。

Another trap is drawing NAND-only or NOR-only circuits. Candidates often try to replace gates individually without reconstructing the whole expression correctly. The systematic approach is to convert the expression into a form suitable for the target gate type, then build the circuit step by step. Marks are often lost because the final circuit uses a mixture of gates.

另一个陷阱是绘制仅用与非门或或非门实现的电路。考生常常试图逐个替换逻辑门，而没有正确地重构整体表达式。系统的方法是将表达式转化为适合目标门类型的标准形式，再逐步搭建电路。很多失分就源于最终电路中出现了混合门。

5. Flip-Flops and Sequential Logic | 触发器与时序逻辑

SR and JK flip-flops are core sequential logic elements. You must be able to draw their symbol, complete timing diagrams, and explain the difference between level-triggered and edge-triggered behaviour. The most damaging mistake is misreading the clock edge on a timing diagram. CAIE frequently uses the falling edge (high-to-low) trigger, but candidates sometimes assume rising edge.

SR 和 JK 触发器是核心的时序逻辑元件。你必须能够画出它们的符号、补全时序图，并解释电平触发和边沿触发的区别。最致命的错误是在时序图上错误读取时钟边沿。CAIE 常采用下降沿（高到低）触发，但有些考生想当然地假定为上升沿触发。

The JK flip-flop’s toggling behaviour (J=1, K=1) during consecutive clock pulses causes confusion. When drawing the output Q, candidates often forget that it toggles ONLY on the active clock edge and retains its state otherwise. Always mark the exact trigger point and draw Q changing immediately after the edge.

JK 触发器在 J=1、K=1 时连续时钟脉冲下的翻转行为常常使人困惑。绘制输出 Q 时，考生经常忘记它仅在有效时钟边沿翻转，其余时间保持原状态。一定要标出确切的触发点，并在边沿后立即画出 Q 的变化。

6. The Fetch-Execute Cycle | 取指执行周期

The fetch-execute cycle, together with the role of registers such as PC, MAR, MDR, CIR, and ACC, is a guaranteed examination topic. Many candidates describe the steps but forget to mention the exact data flows or the buses involved. A full-mark answer must include the address bus, data bus, and control bus, and specify whether a register’s content is placed onto a bus or loaded from one.

取指执行周期以及 PC、MAR、MDR、CIR、ACC 等寄存器的作用是必考内容。许多考生描述了步骤，却忘记提及精确的数据通路或所涉及的总线。满分答案必须包含地址总线、数据总线和控制总线，并指明是将寄存器内容放到总线上还是从总线加载。

A common omission is stating that the PC is incremented simultaneously with the fetch of the instruction, not after the entire execution. Exam markers look for precise phrasing: “The content of PC is copied to MAR via the address bus, then PC is incremented.” Missing the “via the address bus” loses marks.

一个常见的遗漏点是：PC 的递增与指令的取值同时发生，而不是在整个执行之后。阅卷人看重精确表述：“PC 的内容通过地址总线复制到 MAR，然后 PC 递增。”漏掉“通过地址总线”会丢分。

7. Assembly Language and Addressing Modes | 汇编语言与寻址模式

Even though the 9618 syllabus uses a simulation IDE for assembly, the written paper tests your understanding of basic opcodes (LDM, STO, ADD, SUB, JMP, CMP, etc.) and addressing modes: immediate, direct, indirect, and indexed. A very common error is confusing immediate addressing (using #) with direct addressing. LDM #5 loads the value 5 into the accumulator; LDM 5 loads the contents of memory location 5.

尽管 9618 大纲使用模拟 IDE 进行汇编练习，但笔试仍然考查对基本操作码（LDM、STO、ADD、SUB、JMP、CMP 等）和寻址方式（立即、直接、间接、变址）的理解。一个极常见的错误是混淆立即寻址（用 # ）和直接寻址。LDM #5 是将数值 5 加载到累加器；LDM 5 是加载内存地址 5 中的内容。

Indirect addressing causes further confusion. Candidates often fail to follow the chain: the instruction contains an address A; the processor reads the content of A, which is another address B; finally, it uses the value at B. Practice tracing multilayered address resolutions, and always check whether the question expects the content of a register or a memory byte.

间接寻址更容易出错。考生常常无法跟踪寻址链：指令中含有一个地址 A；处理器读取地址 A 的内容，该内容为另一个地址 B；最后再使用地址 B 处的值。多练习追踪多层地址解析，并时刻核查题目要求的是寄存器内容还是内存字节。

8. Networks: IP Addressing and Subnetting | 网络：IP 地址与子网划分

IP addressing and subnetting are high-frequency in the communication chapter. You must distinguish between IPv4 and IPv6 formats, identify the network and host portions of an address given a subnet mask, and calculate the number of available host addresses (2ⁿ – 2). A careless mistake is to forget subtracting the network address and broadcast address from the total, resulting in 2ⁿ instead of 2ⁿ – 2.

IP 地址与子网划分是通信章节的高频考点。你必须能够区分 IPv4 和 IPv6 格式，根据子网掩码识别地址的网络部分和主机部分，并计算可用主机地址数 (2ⁿ – 2)。一个粗心的错误是忘记从总数中减去网络地址和广播地址，导致错误地给出 2ⁿ 而不是 2ⁿ – 2。

Many candidates also misapply the bitwise AND operation when finding the network address. You must perform AND on each octet individually, not on the whole decimal number. If the question asks for the broadcast address, fill all host bits with 1 after determining the network portion.

很多考生在求网络地址时还会错误地进行按位 AND 操作。必须对每个八位组分别进行 AND，而不是对整个十进制数。如果题目要求广播地址，则在确定网络部分后，将所有主机位填为 1。

9. Protocols and the TCP/IP Stack | 协议与 TCP/IP 协议栈

The TCP/IP stack and common protocols (HTTP, HTTPS, FTP, SMTP, POP3, IMAP, DNS) appear in both definition and application questions. Candidates often mix up the transport layer’s responsibilities with those of the application layer. For example, stating that HTTP ensures reliable delivery is wrong—that is TCP’s job. Understand the layer at which each protocol operates and the type of service it provides.

TCP/IP 协议栈以及常见协议（HTTP、HTTPS、FTP、SMTP、POP3、IMAP、DNS）既出现在概念题中，也出现在应用题中。考生常常混淆传输层和应用层的职责。例如，说 HTTP 确保可靠传输就是错误的——那是 TCP 的任务。要清楚地理解各个协议运行的层次及其提供的服务类型。

Another area of confusion is port numbers. The exam may ask which port is used by a service and whether it uses TCP or UDP. Even though port numbers are usually given, knowing that HTTP defaults to port 80, HTTPS to 443, SMTP to 25, and so on, saves time and improves precision.

另一个容易混淆的领域是端口号。考题可能询问某一服务使用的端口号，以及它是基于 TCP 还是 UDP。虽然端口号通常提供，但知道 HTTP 默认端口 80、HTTPS 使用 443、SMTP 使用 25 等，可以节省时间并提高答题准确性。

10. Pseudocode and Trace Tables | 伪代码与跟踪表

Pseudocode problems often involve loops, conditional statements, and arrays. The most common pitfall is failing to initialise variables before the loop or incorrectly placing the increment step, leading to infinite loops or off-by-one errors. When tracing, candidates often forget to update all variables at each iteration or misinterpret the scope of a loop.

伪代码题通常涉及循环、条件语句和数组。最常见的陷阱是循环前没有初始化变量，或者增量步骤位置不对，导致无限循环或差一错误。在进行跟踪时，考生经常忘记在每次迭代时更新所有变量，或者错误理解循环的作用域。

Exam questions sometimes deliberately use subtle syntax like WHILE ... ENDWHILE versus REPEAT ... UNTIL. A REPEAT loop always executes at least once, which can change the output dramatically. Always build a structured trace table with columns for each variable, outputs, and condition checks. This prevents you from jumping to conclusions.

考试题目有时故意使用精细的语法，例如 WHILE ... ENDWHILE 与 REPEAT ... UNTIL。REPEAT 循环至少执行一次，这会显著改变输出结果。务必构建一张结构化的跟踪表，为每个变量、输出和条件检查设立列。这样可以避免草率下结论。

11. Database and SQL Queries | 数据库与 SQL 查询

SQL is a high-scoring topic if you master the basic syntax: SELECT, FROM, WHERE, ORDER BY, GROUP BY, HAVING, and aggregate functions. A common mistake is writing WHERE conditions after GROUP BY when you want to filter groups. Aggregate filters require HAVING, not WHERE. Also, forgetting to join tables correctly when using data from multiple tables leads to Cartesian products.

SQL 是一个只要掌握基本语法就能拿高分的专题：SELECT、FROM、WHERE、ORDER BY、GROUP BY、HAVING 以及聚合函数。常见错误是想要筛选分组时，把 WHERE 条件写在 GROUP BY 之后。聚合过滤需使用 HAVING 而非 WHERE。此外，在使用多表数据时忘记正确联接表，会产生笛卡尔积。

Candidates also lose marks on simple syntax like missing quotation marks around string literals or confusing = with LIKE for pattern matching. Practise writing queries for one-to-many relationships, and always verify that your SELECT columns match those required in the specification.

考生还会在一些简单语法上丢分，例如字符串文字漏加引号，或在模式匹配时混淆 = 与 LIKE。多练习编写一对多关系的查询，并时刻核对 SELECT 的列是否与题目要求一致。

12. Security and Ethical Issues | 安全与伦理问题

Questions on security, ethics, and data protection often require descriptive, structured answers. Even though the material seems straightforward, candidates lose marks by naming a threat without explaining how it works, or by listing countermeasures without linking them to specific vulnerabilities. For example, saying “use encryption” is insufficient; you must state what type of encryption (symmetric, asymmetric) and what it protects against.

有关安全、道德和数据保护的题目通常需要描述性、结构化的答案。虽然此部分内容看似简单，考生却常因只说出威胁名称而未解释其工作原理而丢分，或者只是罗列应对措施但不与特定漏洞关联。例如，仅仅说“使用加密”是不够的；必须说明所用加密类型（对称、非对称）以及它旨在防御什么。

Another high-frequency demand is discussing the impacts of digital technology on privacy, free speech, and intellectual property. Candidates often offer generic statements rather than specific, balanced arguments. Prepare a few concrete examples—such as a company using facial recognition versus GDPR—and practise structuring a two-sided evaluation.

另一项高频要求是讨论数字技术对隐私、言论自由和知识产权的影响。考生往往给出笼统的陈述，而非具体、均衡的论点。准备几个具体案例——比如某公司使用面部识别与 GDPR 的冲突——并练习构思正反两面的评价结构。

Published by TutorHao | Computer Science Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）