📚 Year 13 AQA Computer Science: Exam Techniques and Marking Criteria | AQA A-level计算机科学：答题技巧与评分标准
Excelling in the Year 13 AQA A-Level Computer Science exams (7517) requires more than just knowing the theory; it demands a precise understanding of how marks are allocated across assessment objectives, how to interpret command words, and how to structure your answers to match the examiner’s expectations. This guide distils the official mark schemes, examiner reports, and proven classroom strategies into practical techniques you can apply immediately in both the on-screen programming paper and the written theory paper.
在AQA A-level计算机科学（7517）考试中拔得头筹，光掌握理论远远不够。你必须准确理解分数如何在各评估目标间分配，如何解读命令词，以及如何组织答案才能满足阅卷者的期待。本指南将官方评分标准、考官报告及课堂实战经验提炼为可立即运用的实用技巧，覆盖上机编程卷与笔试理论卷。
1. Understanding Assessment Objectives (AOs) | 深入理解评估目标
| AO1: Demonstrate knowledge and understanding of facts, concepts, principles and terminology. | AO1：展示对事实、概念、原理和术语的知识与理解。 |
| AO2: Apply knowledge and understanding to problems and scenarios, including programming. | AO2：将知识与理解应用于问题和情境，包括编程任务。 |
| AO3: Design, develop, test and refine programs and systems; analyse problems using computational thinking. | AO3：设计、开发、测试并优化程序与系统；运用计算思维分析问题。 |
| AO4: Evaluate systems, solutions and the impact of computing, making reasoned judgements. | AO4：评估系统、解决方案以及计算技术的影响，做出合理的判断。 |
Paper 1 (on-screen programming) targets mainly AO2 and AO3, while Paper 2 (written theory) assesses all four AOs, with a strong emphasis on AO1 and AO4 in the extended response questions. Knowing which AO is being tested by each question is the first step to shaping a high-scoring answer.
Paper 1（上机编程）主要考核AO2和AO3，而Paper 2（笔试理论）则覆盖全部四个评估目标，并在扩展作答题目中特别侧重AO1和AO4。认清每道题所考核的AO，是塑造高分答案的第一步。
2. Decoding Command Words | 解码命令词
| State / Identify / Name – Provide a brief fact or term, no explanation needed. | 陈述 / 识别 / 命名 – 给出简短的事实或术语，无需解释。 |
| Describe – Give a detailed account of characteristics or steps, still without explaining reasons. | 描述 – 详细说明特征或步骤，同样无需解释原因。 |
| Explain – Make something clear by giving reasons or causes; ‘because’ is your friend here. | 解释 – 通过给出理由或原因把问题说清楚；在这里多用“因为”。 |
| Compare / Contrast – Identify similarities and differences; use ‘whereas’, ‘however’, ‘on the other hand’. | 比较 / 对比 – 找出相似点与不同点；使用“而”、“然而”、“另一方面”等词语。 |
| Evaluate / Assess / Discuss – Present a balanced argument, weighing up pros and cons, and arrive at a justified conclusion. | 评估 / 评价 / 讨论 – 呈现平衡的论点，权衡利弊，并得出有理据的结论。 |
| Complete / Trace – Follow an algorithm step by step, updating a trace table exactly as the code would execute. | 完成 / 追踪 – 逐步执行算法，完全按照代码执行的方式更新追踪表。 |
Examiners repeatedly report that many marks are lost simply because candidates misinterpret ‘explain’ as ‘describe’ or omit a balanced discussion when asked to ‘evaluate’. Highlight the command word in the question before you start writing.
考官一次次指出，许多失分仅仅因为考生将“解释”误解为“描述”，或在要求“评价”时遗漏了平衡的讨论。下笔前，先在题目中圈出命令词。
3. Paper 1 (7517/1) On-screen Programming: Key Techniques | Paper 1 上机编程考试关键技巧
The on-screen exam provides skeleton code, usually in Python, which you must modify, debug or extend. Before typing anything, spend five minutes reading the entire skeleton code and its comments. Identify the purpose of each function and the global variables it uses.
上机考试提供骨架代码（通常为Python），你需要修改、调试或扩展。在动手输入任何代码之前，花五分钟通读整个骨架代码及其注释。弄清每个函数的作用及其使用的全局变量。
Every change you make must be purposeful and match the question’s instructions exactly. If the question asks to ‘add validation to ensure the user enters a positive integer’, do not add unnecessary features like formatting the output. Marks are allocated for precise functionality, not for extra code.
你所做的每一处改动都必须目的明确且完全符合题目要求。如果题目要求“添加验证以确保用户输入正整数”，就不要添加输出格式等无关功能。分数只给精准的功能，而不是给额外的代码。
When debugging, use a trace table on the paper provided. Manually step through the suspect code section, recording variable values. This systematic approach will reveal off-by-one errors, incorrect loop conditions and logical flaws far more reliably than random trial and error.
调试时，利用提供的草稿纸制作一个追踪表。手动逐步执行可疑的代码段，记录变量值。这种系统方法能远比盲目试错更可靠地发现差一错误、错误的循环条件以及逻辑缺陷。
Remember that Paper 1 questions often embed AO3 design marks for clear, maintainable code. Use meaningful variable names, add concise comments where instructed, and keep your indentation perfect. Examiners can deduct marks if your code becomes messy or unreadable.
请记住，Paper 1的题目常常会因代码清晰、可维护而涉及AO3设计分。使用有意义的变量名，按指示添加简洁的注释，并保持完美的缩进。如果你的代码变得混乱或不可读，考官有权扣分。
4. Paper 2 (7517/2) Written Theory: Tackling Extended Responses | Paper 2 笔试理论：应对扩展作答
Paper 2 features 9-mark and 12-mark extended writing questions that assess AO4 (evaluation). The mark scheme typically rewards a balance of points, explicit linkages, and a substantiated conclusion. A common structure is PEEL: Point, Evidence, Explanation, Link.
Paper 2包含9分和12分的扩展写作题，考核AO4（评价）。评分方案通常奖励平衡的论点、清晰的关联和有依据的结论。常见的结构为PEEL：观点、证据、解释、联系。
For example, if asked to evaluate the use of cloud storage for a school, your answer might open with a clear point about reduced hardware cost, support it with specific evidence (no local servers needed, pay-as-you-go model), explain the benefit, and then link to a counterpoint about data security. The final paragraph must present your overall judgement, directly addressing the scenario.
例如，若要求评价学校使用云存储的方案，你的答案可以首先明确提出降低硬件成本的观点，用具体证据支撑（无需本地服务器、按用量付费），解释其好处，然后联系到数据安全方面的相反观点。最后一段必须给出你的总体判断，并直接针对题目情境。
Underline the specific context words in the question, such as ‘a small primary school’ or ‘a global bank’, because evaluative marks are tied to that context. Generic answers that ignore the context are capped at a lower band.
在题中标出特定的情境词汇，如“一所小型小学”或“一家跨国银行”，因为评价分是与该情境挂钩的。忽视情境的泛泛而答只能被限制在较低分数段。
5. Mastering Algorithm Trace Tables | 精通算法跟踪表
Trace table questions are a staple of both papers. They demand absolute accuracy. Consider this typical loop: total ← 0; FOR i ← 1 TO 3: total ← total + i * i; ENDFOR. You would build a table with columns for i and total, updating on every iteration.
跟踪表题型是两卷的常客，要求绝对准确。考虑典型的循环结构：total ← 0; FOR i ← 1 TO 3: total ← total + i * i; ENDFOR。你将建立一个包含i和total等列的表格，并在每次迭代时更新。
Marks are awarded for completely correct rows. A single misplaced value in a later iteration can invalidate all subsequent rows. To avoid this, double-check the initial values before the loop, and trace each instruction exactly as an interpreter would, not as you wish it would work.
只有完全正确的行才能得分。如果某次迭代中出现一个错误的值，可能导致后续所有行无效。为避免这一点，请仔细检查循环前的初始值，并完全按照解释器执行的方式跟踪每条指令，而不是按照你希望的方式。
When the algorithm involves arrays or subroutines, extend the table with columns for each array index or parameter change. Use a small, neat layout and always label your columns. If the question provides a pre-printed trace table, never alter its structure.
当算法涉及数组或子程序时，要为每个数组索引或参数变化扩展表格列。使用小巧整洁的布局，并始终标明列标题。如果题目提供了预印的追踪表，千万不要改变它的结构。
6. Regular Languages, FSM and BNF | 正则语言、有限状态机与巴科斯-诺尔范式
When drawing a finite state machine, clarity is paramount. Use circles for states, label the start state with an incoming arrow, and clearly indicate accepting states with a double circle. Transitions should be labelled with input symbols, using ‘0,1’ if multiple inputs cause the same transition.
绘制有限状态机时，清晰度至关重要。用圆圈表示状态，用入向箭头标记起始状态，用双圈明确标示接受状态。转换要用输入符号标记，若多个输入导致同一转换则标注为“0,1”。
Examiners look for correctness in the number of states and the completeness of transitions. A missing transition for a valid input symbol loses marks instantly. Always test your FSM mentally with a few accepted and rejected strings to verify it behaves as intended.
考官看重状态数和转换的完备性。缺少某个有效输入符号的转换会立刻失分。务必在心里用几个可接受和应被拒绝的字符串测试你的FSM，验证其行为是否与预期一致。
For BNF questions, precision is everything. Use angle brackets for non-terminals, the ::= symbol for definitions, and vertical bars for alternatives. For instance, a simple digit rule: <digit> ::= 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9. Do not forget to define all non-terminals used, and avoid any ambiguity that could allow invalid strings.
对于BNF问题，准确性就是一切。用尖括号括起非终结符，使用 ::= 表示定义，用竖线表示选择。例如，一个简单的数字规则：<digit> ::= 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9。不要忘记定义所有用到的非终结符，并避免任何可能导致无效字符串的二义性。
7. Tackling System Software and Architecture Comparison Questions | 解答系统软件与体系结构比较题
Comparison questions, such as ‘Compare RISC and CISC architectures’, require a point-by-point juxtaposition. A single paragraph listing all RISC features and then all CISC features will score poorly. Use a paired-sentence approach: ‘RISC uses a fixed-length instruction format, whereas CISC instructions vary in length.’
比较类题目，如“比较RISC与CISC体系结构”，需要逐点并列。若用一段单独列出RISC的所有特点，再用一段列出CISC的特点，得分会很低。应采用成句比较的方式：“RISC使用固定长度指令格式，而CISC指令长度可变。”
Marks are split between providing a factual difference and explicitly contrasting the two. Use discourse markers like ‘however’, ‘on the contrary’, and ‘in contrast’ to make the comparison explicit. Aim to present at least three distinct contrasted points for a 6-mark question.
分数分配在提供事实差异与明确进行对比之间。使用“然而”、“相反地”、“对比之下”等语篇标记词，使对比显化。对于6分题，争取至少给出三组明确的对比点。
When comparing scheduling algorithms, remember to ground your answer in the assumptions of the scenario (interactive vs batch processing). Avoid vague statements like ‘it is faster’. Instead, state that ‘Round Robin provides better response time for interactive users at the expense of higher context-switch overhead’.
比较调度算法时，记得将答案建立在情境假设之上（交互式处理与批处理）。避免使用“它更快”这类模糊表述。相反，应阐明“轮转法可为交互用户提供更短的响应时间，代价是更高的上下文切换开销”。
8. Networking and Security: Structuring Your Answers | 网络与安全：答卷结构
For questions that ask you to explain how data is transmitted using the TCP/IP model, always begin from the application layer and work downwards. Name each layer, describe its key protocol, and explain the encapsulation that occurs. Use a stepwise structure: ‘At the Application layer, HTTP formats the request… The Transport layer then encapsulates this into segments…’
对于要求解释数据如何通过TCP/IP模型传输的题目，务必从应用层开始向下描述。说出每一层的名称，描述其关键协议，并解释发生的封装过程。采用逐步结构：“在应用
Published by TutorHao | Year 13 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