SE Exam 1 Cheat Sheet

软件工程考试核心概念速查

1️⃣ Software Development Process

软件开发过程

Software process organizes development activities into phases.

软件开发过程将开发活动结构化为多个阶段。

Waterfall Model 瀑布模型

Sequential phase development.

顺序阶段式开发。

✅ Pros 优点

Simple & structured
简单清晰、易管理
Good for stable requirements
适合需求稳定项目
Clear documentation
文档规范完整

❌ Cons 缺点

Hard to accommodate change
难以适应需求变化
Late defect discovery
缺陷发现较晚
Expensive rework
返工成本高

Exam Tip: rigid process, late feedback

Spiral Model 螺旋模型

Risk-driven iterative development.

以风险驱动的迭代开发。

✅ Pros

Focus on risk reduction
优先降低风险
Early risk detection
早期发现风险
Iterative improvement
迭代优化

❌ Cons

Complex management
管理复杂
Expensiv e
成本较高
Requires expertise
需要经验丰富团队

Exam Tip: risk-first development

Agile Development 敏捷开发

Iterative, adaptive, customer collaboration.

迭代开发，快速响应变化。

✅ Pros

Adapts to change
适应需求变化
Continuous feedback
持续反馈
Faster value delivery
快速交付价值

❌ Cons

Less predictability
可预测性较低
Requires customer involvement
需要客户持续参与
Hard to scale
大型项目难扩展

Exam Tip: embraces uncertainty & change

Extreme Programming (XP)

Agile practice emphasizing engineering discipline.

强调工程实践的敏捷方法。

✅ Pros

High code quality (TDD & pair programming)
TDD与结对编程提高质量
Rapid feedback
快速反馈
Strong collaboration
团队协作强

❌ Cons

Intensive teamwork required
协作要求高
Hard for large teams
不适合大型团队

2️⃣ Planning & Scheduling

项目规划与进度

Importance of Planning

Benefits

reduces cost later
降低后期成本
improves efficiency
提高效率
prevents integration failure
避免集成失败

Poor planning leads to:

scope creep 范围膨胀
late defects 缺陷晚发现
integration failures 集成失败

Milestones vs Deliverables

Milestones

internal progress checkpoints
内部进度检查点

Deliverables

outputs delivered to customer
交付给客户的成果

❗ “80% complete” is NOT a milestone

Scheduling Reality

Last 10% of work may take 40% of total time.

最后10%可能耗费40%的时间。

原因：

integration problems
edge cases
polishing & debugging

3️⃣ Risk & Uncertainty

风险与不确定性

Risk management = identify → assess → prioritize → mitigate

风险管理流程：识别 → 评估 → 排序 → 降低

Common Risks

staff turnover 人员流失
performance issues 性能问题
new competitors 竞争产品
technology failure 技术风险

Effective Risk Management

address risks early
estimate likelihood & impact
focus on major risks
prepare contingency plans

尽早处理风险、评估概率与影响、关注关键风险、制定应急方案。

Zero-Risk Bias 零风险偏见

Tendency to eliminate one risk completely instead of minimizing overall risk.

倾向完全消除某一风险，而不是降低整体风险。

Weber’s Law 韦伯定律

Problems obvious in small systems are harder to detect in large systems.

小系统明显的问题，在大型系统中更难察觉。

4️⃣ Measurement & Metrics

软件度量与指标

Measurement reduces uncertainty.

度量用于减少不确定性。

Lines of Code (LOC)

✅ Pros

easy to measure
易测量
simple comparison
易比较

❌ Cons

affected by language/style
受语言与格式影响
more LOC ≠ productivity
行数多不代表效率高

Cyclomatic Complexity 圈复杂度

Measures number of independent execution paths.

衡量独立执行路径数量。

✅ Pros

estimates testing effort
估算测试复杂度
identifies complex code
识别复杂代码

❌ Cons

ignores data complexity
忽略数据复杂度
not full quality measure
不能全面衡量质量

Maintainability Index (MI)

0–100 scale indicating maintainability.

数值越高越易维护。

Derived from:

LOC
Cyclomatic complexity
Halstead volume

Halstead Volume

Measures program size via operators & operands.

通过操作符与操作数衡量程序规模。

Measurement Biases 度量偏差

Streetlight Effect 路灯效应

Measuring what is easy instead of what matters.

测量容易的，而非重要的。

McNamara Fallacy

Relying only on quantitative metrics.

只依赖量化指标做决策。

Correlation ≠ Causation

Confounding Variables

Hidden factors influencing results.

混杂变量影响结果。

5️⃣ Quality Assurance & Testing

质量保证与测试

QA focuses on preventing defects through process.

QA 通过流程改进预防缺陷。

Testing executes the program to find defects.

测试通过运行程序发现缺陷。

QA vs Testing

QA = prevention & process
QA 是预防与流程

Testing = execution & detection
Testing 是执行与发现

Testing Truth

Testing can reveal bugs but cannot prove absence.

测试能发现错误，但不能证明没有错误。

原因：

infinite input space
halting problem

External vs Internal Quality

External

correctness
reliability
security
no crashes

用户体验质量

Internal

readability
maintainability
structure

开发者视角质量

Testing Types

Unit Testing

Tests individual components.

优点：

fast & repeatable
easy debugging

Integration Testing

Tests modules working together.

验证模块协同工作。

Regression Testing

Ensures changes don’t break old features.

防止旧功能被破坏。

Test-Driven Development (TDD)

write failing test
implement code
pass test
refactor

优点：

better design
fewer bugs

Mocking

Simulates dependencies.

模拟依赖组件。

Used when:

APIs unavailable
failures hard to trigger

Fuzz Testing

Random inputs to detect crashes & vulnerabilities.

随机输入发现崩溃与安全漏洞。

Penetration Testing

Simulates attacks to evaluate security.

模拟攻击评估安全性。

6️⃣ Test Coverage & Test Quality

Coverage measures how much code tests execute.

覆盖率衡量测试执行范围。

Line Coverage

executed lines / total lines

Branch Coverage

tests true & false paths

Branch coverage > line coverage

Coverage Limitations

Coverage ≠ correctness
覆盖率不保证正确性

Mutation Testing

Insert defects to evaluate test strength.

Mutation score = killed mutants / total

优点：

evaluates test effectiveness

缺点：

expensive
equivalent mutants

7️⃣ Test Inputs & Oracles

Test case includes:

input
oracle
comparator

测试用例包含：输入 + 预期输出 + 比较机制

Oracle Problem

Determining correct output is difficult.

确定正确输出困难。

Implicit oracles:

no crash
no infinite loop

Path Explosion

Paths = 2^N decisions
路径指数增长。

Invariants

Properties always true.

始终成立的性质。

可作为测试判定依据。

8️⃣ Code Review & Inspection

Static QA without execution.

不运行程序的质量保证方法。

Code Review

Peer review of changes.

同伴评审代码。

Goals

find defects
improve readability
knowledge sharing

Code Inspection

Formal, structured team review.

正式结构化评审。

优点：

thorough detection

缺点：

expensive

Why Not Testing Alone?

Testing cannot evaluate:

maintainability
design quality
documentation

9️⃣ Dynamic Analysis

动态分析

Runs instrumented program to collect runtime data.

运行插桩程序收集运行时信息。

Uses

performance profiling
memory usage
race detection
security vulnerabilities

Race Condition

Concurrent access without synchronization.

并发访问共享资源导致错误。

Taint Tracking

Tracks untrusted input influence.

追踪不可信输入传播。

Limitations

performance overhead
false positives
behavior change (Heisenbugs)

🔟 Static & Dataflow Analysis

静态与数据流分析

Static analysis examines code without execution.

静态分析在不运行程序的情况下分析代码。

Why Important

Testing cannot cover all paths.

测试无法覆盖所有路径。

Static analysis reasons about all possible executions.

Dataflow Analysis

Tracks abstract values through program flow.

跟踪变量在程序中的传播。

AST & CFG

AST → syntax structure
CFG → execution paths

Conservative Analysis

Prefer unknown over incorrect results.

宁可不确定，也不能错误判断。

Limit

Many properties are undecidable (Rice’s theorem).

许多程序性质不可判定。

⭐ EXAM CORE TRUTHS（必背结论）

Early defect detection saves cost
Testing increases confidence, not correctness
Coverage ≠ correctness
Measurement reduces uncertainty
Metrics can create bad incentives
Static + dynamic analysis complement testing
Risk = probability × impact

SE Exam 1 Cheat Sheet ​

1️⃣ Software Development Process ​

Waterfall Model 瀑布模型 ​

✅ Pros 优点 ​

❌ Cons 缺点 ​

Spiral Model 螺旋模型 ​

✅ Pros ​

❌ Cons ​

Agile Development 敏捷开发 ​

✅ Pros ​

❌ Cons ​

Extreme Programming (XP) ​

✅ Pros ​

❌ Cons ​

2️⃣ Planning & Scheduling ​

Importance of Planning ​

Benefits ​

Poor planning leads to: ​

Milestones vs Deliverables ​

Scheduling Reality ​

3️⃣ Risk & Uncertainty ​

Common Risks ​

Effective Risk Management ​

Zero-Risk Bias 零风险偏见 ​

Weber’s Law 韦伯定律 ​

4️⃣ Measurement & Metrics ​

Lines of Code (LOC) ​

✅ Pros ​

❌ Cons ​

Cyclomatic Complexity 圈复杂度 ​

✅ Pros ​

❌ Cons ​

Maintainability Index (MI) ​

Halstead Volume ​

Measurement Biases 度量偏差 ​

Streetlight Effect 路灯效应 ​

McNamara Fallacy ​

Correlation ≠ Causation ​

Confounding Variables ​

5️⃣ Quality Assurance & Testing ​

QA vs Testing ​

Testing Truth ​

External vs Internal Quality ​

External ​

Internal ​

Testing Types ​

Unit Testing ​

Integration Testing ​

Regression Testing ​

Test-Driven Development (TDD) ​

Mocking ​

Fuzz Testing ​

Penetration Testing ​

6️⃣ Test Coverage & Test Quality ​

Line Coverage ​

Branch Coverage ​

Coverage Limitations ​

Mutation Testing ​

7️⃣ Test Inputs & Oracles ​

Oracle Problem ​

Path Explosion ​

Invariants ​

8️⃣ Code Review & Inspection ​

Code Review ​

Goals ​

Code Inspection ​

Why Not Testing Alone? ​

9️⃣ Dynamic Analysis ​

Uses ​

Race Condition ​

Taint Tracking ​

Limitations ​

🔟 Static & Dataflow Analysis ​

Why Important ​

Dataflow Analysis ​

AST & CFG ​

Conservative Analysis ​

Limit ​

⭐ EXAM CORE TRUTHS（必背结论） ​

SE Exam 1 Cheat Sheet

1️⃣ Software Development Process

Waterfall Model 瀑布模型

✅ Pros 优点

❌ Cons 缺点

Spiral Model 螺旋模型

✅ Pros

❌ Cons

Agile Development 敏捷开发

✅ Pros

❌ Cons

Extreme Programming (XP)

✅ Pros

❌ Cons

2️⃣ Planning & Scheduling

Importance of Planning

Benefits

Poor planning leads to:

Milestones vs Deliverables

Scheduling Reality

3️⃣ Risk & Uncertainty

Common Risks

Effective Risk Management

Zero-Risk Bias 零风险偏见

Weber’s Law 韦伯定律

4️⃣ Measurement & Metrics

Lines of Code (LOC)

✅ Pros

❌ Cons

Cyclomatic Complexity 圈复杂度

✅ Pros

❌ Cons

Maintainability Index (MI)

Halstead Volume

Measurement Biases 度量偏差

Streetlight Effect 路灯效应

McNamara Fallacy

Correlation ≠ Causation

Confounding Variables

5️⃣ Quality Assurance & Testing

QA vs Testing

Testing Truth

External vs Internal Quality

External

Internal

Testing Types

Unit Testing

Integration Testing

Regression Testing

Test-Driven Development (TDD)

Mocking

Fuzz Testing

Penetration Testing

6️⃣ Test Coverage & Test Quality

Line Coverage

Branch Coverage

Coverage Limitations

Mutation Testing

7️⃣ Test Inputs & Oracles

Oracle Problem

Path Explosion

Invariants

8️⃣ Code Review & Inspection

Code Review

Goals

Code Inspection

Why Not Testing Alone?

9️⃣ Dynamic Analysis

Uses

Race Condition

Taint Tracking

Limitations

🔟 Static & Dataflow Analysis

Why Important

Dataflow Analysis

AST & CFG

Conservative Analysis

Limit

⭐ EXAM CORE TRUTHS（必背结论）