AIGC标识 软件开发英语学习教材

软件开发英语学习教材(30 课时)

适用人群:软件从业者、计算机专业学生、想进外企/大厂的英语学习者
内容体系:围绕「软件开发 → 测试 → 部署 → 运维」四大主线,共 30 个课时,每个课时一个真实主题
每课结构:英文阅读(附中文对照翻译)→ 重点词汇(音标/词性/释义/例句/频度)→ 重点短语(音标/释义/例句/频度)
频度说明:★★★ 极高频(工作天天见)|★★ 高频(经常出现)|★ 常用(值得掌握)
学习建议:① 每课先读英文原文猜词义,再看中文对照核对;② 优先记忆 ★★★ 词汇;③ 大声朗读例句并尝试自己造句;④ 按"当天 → 第2天 → 第7天 → 第30天"间隔复习


课程目录(Syllabus)

课时 主题(中文) Topic(English) 核心内容
1 软件开发全流程总览 From Code to Production 开发/测试/部署/运维总览 + 55 个核心词汇
2 需求分析与产品规划 Requirements & Product Planning 用户故事、待办清单、冲刺
3 敏捷开发与 Scrum Agile Development & Scrum 站会、回顾、燃尽图
4 版本控制与 Git 实操 Version Control with Git 仓库、推送、拉取、冲突
5 代码评审与团队协作 Code Review & Collaboration 评审、批准、结对编程
6 代码质量与重构 Code Quality & Refactoring 可维护性、技术债务、坏味道
7 软件架构与设计模式 Architecture & Design Patterns 模块化、解耦、分层
8 单元测试与测试金字塔 Unit Testing & Test Pyramid 断言、模拟、测试用例
9 集成与端到端测试 Integration & E2E Testing 接口测试、冒烟测试、无头测试
10 自动化测试框架 Automation Testing Frameworks 测试脚本、并行、质量门禁
11 持续集成(CI) Continuous Integration 触发、构建服务器、主干开发
12 持续交付与部署(CD) Continuous Delivery & Deployment 发布频率、前置时间、功能开关
13 构建与产物管理 Build & Artifact Management 编译、打包、制品库、发布说明
14 环境与配置管理 Environment & Configuration 环境变量、密钥、配置漂移
15 容器化与 Docker Containerization & Docker 镜像、分层、卷、运行时
16 容器编排与 Kubernetes Orchestration & Kubernetes 集群、节点、Pod、负载均衡
17 微服务架构 Microservices Architecture 单体、网关、熔断器、服务发现
18 部署策略 Deployment Strategies 蓝绿部署、金丝雀、零停机
19 基础设施即代码(IaC) Infrastructure as Code 声明式、幂等、资源编排
20 监控与指标 Monitoring & Metrics 仪表盘、阈值、黄金信号
21 日志与可观测性 Logging & Observability 链路追踪、遥测、日志聚合
22 告警与值班 Alerting & On-call 升级、寻呼机、值班轮换
23 事件管理与故障处理 Incident Management & Troubleshooting 分诊、缓解、事后复盘
24 站点可靠性工程(SRE) Site Reliability Engineering SLO、SLA、错误预算、混沌工程
25 性能优化与延迟 Performance & Latency 瓶颈、吞吐量、压测、缓存
26 安全与 DevSecOps Security & DevSecOps 漏洞、加密、零信任、审计
27 云原生与云计算 Cloud Native & Cloud Computing 虚拟机、无服务器、弹性、可用区
28 数据库与数据管理 Databases & Data Management 查询、迁移、备份、索引
29 团队协作与沟通 Team Collaboration & Communication 同步、交接、文档、行动项
30 技术面试与职业英语 Technical Interview & Career English 面试、简历、晋升、入职

第 1 课 软件开发全流程总览

Lesson 1: From Code to Production

学习目标:掌握软件工程全流程(开发→测试→部署→运维)的框架性词汇,建立后续 29 课的"词汇地图"。

📖 英文阅读(Reading)

原文:

Most people imagine software as lines of code on a screen. In reality, writing code is only one step in a much longer journey. Every serious application passes through a lifecycle that includes development, testing, deployment, and ongoing operations. Understanding this full process is essential for anyone starting a career in the software industry — and it is also a great way to learn practical English.

Development. The journey begins with planning. Teams collect requirements from customers and other stakeholders, decide which features to build, and set a rough timeline. Modern teams usually work in short iterations, delivering small pieces of value every two weeks or so, instead of waiting for one giant release at the end of the project.

During development, developers manage their code with a version control system such as Git. They create a separate branch for each new task, save their progress with commits, and later merge their work back into the main codebase. Before a merge, the changes usually go through a code review: other developers read the code, look for problems, and suggest improvements. A small mistake in the code is called a bug, and finding and fixing bugs is a normal part of daily work. Teams also refactor code to keep it clean, and they manage dependencies — the external libraries the software relies on — very carefully, because one broken dependency can affect the whole project.

Testing. No code should reach users before it is tested. Quality assurance (QA) is the discipline of making sure software meets quality standards. Testing is done at several levels. Unit tests check individual functions in isolation. Integration tests verify that different components work together correctly. End-to-end tests walk through real user scenarios, from logging in to checking out a shopping cart, to validate the system as a whole.

Teams also measure test coverage — the percentage of code that is actually exercised by tests. The goal is to catch defects early, because fixing a problem during development is far cheaper than fixing it in production. Regression testing makes sure that new changes do not break features that already work. Many teams rely on automated testing: a test suite runs automatically on every change, so every commit is checked within minutes.

Deployment. Once the code passes testing, the team deploys it. The automated path from code to users is called a pipeline. Modern pipelines are built on continuous integration (CI) and continuous delivery (CD). CI builds and tests every code change automatically; CD packages the approved code and keeps it ready for release. Together they make delivery fast, frequent, and repeatable.

The software moves through several environments. Developers work in a local environment; a staging environment mimics production for final checks; and production is where real users access the service. Each stage of the pipeline produces an artifact, such as a package or a container image. Containers like Docker bundle the application with everything it needs, so it runs the same way everywhere. If a release causes problems, teams can do a rollback and switch back to the previous version within minutes. Automation is the heart of this process: manual steps are slow and error-prone, so teams automate everything they can.

Operations. Deployment is not the end of the story. Once software goes live, operations teams keep it running and maintain it day after day. They monitor systems around the clock, watching metrics such as CPU usage, latency, and error rates, and reading logs to understand what the software is doing. Monitoring tools raise alerts when something looks wrong, for example when a service becomes slow or starts failing.

When a real problem occurs, it becomes an incident. Teams respond quickly, limit the outage and downtime, and later investigate the root cause so the same problem never happens again. Engineers take turns being on-call: they carry a phone or pager and must respond to emergencies at any hour. Strong troubleshooting skills — finding and fixing the true cause of a failure — are among the most valuable abilities in the industry.

The ultimate goals of operations are reliability, availability, and scalability. Reliability means the software works correctly; availability means it is almost always reachable, with high uptime; scalability means it can handle growth by adding more resources. Feedback from monitoring and users flows back into planning, closing the loop. The software improves through continuous iteration and continuous improvement: every cycle makes the product a little better than the last, guided by a short feedback loop.

Conclusion. Development, testing, deployment, and operations are not separate silos — they are one continuous cycle. Mastering this vocabulary will not only help you read technical documents with confidence, but also give you a mental map of how modern software really works.

中文翻译:

大多数人以为软件就是屏幕上的代码。实际上,写代码只是漫长旅程中的一步。每个正式应用都要经历一个生命周期:包括开发、部署和持续运维。理解这个完整流程,对任何想进入软件行业的人来说都至关重要——同时它也是学习实用英语的好途径。

开发。 旅程始于规划。团队从客户和其他利益相关者那里收集需求,决定要构建哪些功能,并设定大致的时间线。现代团队通常以短迭代的方式工作,每两周左右交付一小部分价值,而不是等到项目结束才一次性发布一个巨大的版本。

开发期间,开发者用 Git 等版本控制系统管理代码。他们为每个新任务创建独立分支,通过提交保存进度,最后把工作合并回主代码库。合并之前,变更通常要经过代码评审:其他开发者阅读代码、查找问题、提出改进建议。代码中的小错误叫bug(缺陷),发现并修复 bug 是日常工作的常态。团队还会重构代码保持整洁,并非常谨慎地管理依赖——即软件所依赖的外部库——因为一个损坏的依赖可能影响整个项目。

测试。 任何代码在到达用户之前都必须经过测试。质量保证(QA)是确保软件达到质量标准的学科。测试分为多个层次:单元测试独立检查单个函数;集成测试验证不同组件能否正确协同;端到端测试走查真实用户场景(从登录到购物车结算),从整体上验证系统。

团队还会衡量测试覆盖率——即被测试真正执行到的代码百分比。目标是尽早发现缺陷,因为在开发阶段修复问题的成本远低于在生产环境修复。回归测试确保新变更不会破坏已经正常工作的功能。许多团队依赖自动化测试:测试套件在每次变更时自动运行,因此每次提交都能在几分钟内得到检查。

部署。 代码通过测试后,团队就进行部署。从代码到用户的自动化路径叫流水线(pipeline)。现代流水线建立在持续集成(CI)和持续交付(CD)之上:CI 自动构建并测试每一次代码变更;CD 打包已批准的代码,使其随时可以发布。两者结合,让交付又快、又频繁、又可重复。

软件会经过几个环境:开发者工作在本地环境;预发布环境(staging)模拟生产环境做最后检查;生产环境(production)是真实用户访问服务的地方。流水线的每个阶段都会产出构建产物(artifact),比如一个软件包或容器镜像。Docker 这类容器把应用和它所需的一切打包在一起,所以它在任何地方运行结果都一样。如果发布出了问题,团队可以回滚(rollback),几分钟内切回上一个版本。自动化是这个流程的核心:手动步骤又慢又容易出错,所以团队会尽可能把所有环节自动化。

运维。 部署不是故事的终点。软件上线后,运维团队让它持续运行,并日复一日地维护它。他们全天候监控系统,观察 CPU 使用率、延迟、错误率等指标,通过阅读日志了解软件在做什么。当某个环节异常(比如服务变慢或开始报错)时,监控工具会发出告警。

当真实问题发生时,它就成为一个事件(incident)。团队快速响应,限制中断(outage)和停机时间(downtime),之后调查根本原因(root cause),确保同样的问题不再发生。工程师轮流值班(on-call):他们随身带着手机或寻呼机,任何时刻都必须响应紧急情况。强大的故障排查(troubleshooting)能力——找到并修复故障的真正原因——是行业中最有价值的能力之一。

运维的终极目标是可靠性(reliability)、可用性(availability)和可扩展性(scalability)。可靠性意味着软件正确运行;可用性意味着服务几乎始终可达,保持很高的正常运行时间(uptime);可扩展性意味着可以通过增加资源来应对增长。来自监控和用户的反馈流回规划环节,形成闭环。软件通过持续的迭代和持续改进不断进步:在反馈回路的引导下,每个周期都让产品比上一个更好。

结论。 开发、测试、部署和运维不是各自孤立的筒仓(silo),而是一个连续的循环。掌握这套词汇,不仅能让你自信地阅读技术文档,还会为你绘制一张"现代软件如何真正运转"的认知地图。

📚 重点词汇(Key Vocabulary)

本课收录软件开发领域最核心的 55 个高频词(全教材基础词汇库,后续课时在此基础上扩展新词)。

一、开发 Development(15 词)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
lifecycle /ˈlaɪfsaɪkl/ n. 生命周期 The software lifecycle includes development, testing, and maintenance. 软件生命周期包括开发、测试和维护。 ★★
requirement /rɪˈkwaɪəmənt/ n. 需求 The client added a new requirement for the login page. 客户为登录页增加了一个新需求。 ★★★
stakeholder /ˈsteɪkhəʊldə(r)/ n. 利益相关者 We discussed the new feature with all stakeholders. 我们与所有利益相关者讨论了新功能。 ★
feature /ˈfiːtʃə(r)/ n. 功能特性 This release adds a dark mode feature. 这个版本新增了深色模式功能。 ★★★
timeline /ˈtaɪmlaɪn/ n. 时间线、进度表 The project timeline was extended by two weeks. 项目时间线延长了两周。 ★
iteration /ˌɪtəˈreɪʃn/ n. 迭代 Each iteration improves the product a little. 每次迭代都让产品进步一点。 ★★★
version control /ˈvɜːʃn kənˈtrəʊl/ n. 版本控制 Git is the most popular version control system. Git 是最流行的版本控制系统。 ★★
branch /brɑːntʃ/ n. 分支 She created a new branch for the payment module. 她为支付模块创建了一个新分支。 ★★★
commit /kəˈmɪt/ v. 提交(代码) Remember to commit your changes before leaving. 离开前记得提交你的更改。 ★★★
merge /mɜːdʒ/ v. 合并 He merged his branch into the main codebase. 他把自己的分支合并进主代码库。 ★★★
code review /kəʊd rɪˈvjuː/ n. 代码评审 Every pull request must pass a code review. 每个合并请求都必须通过代码评审。 ★★★
bug /bʌɡ/ n. 缺陷、程序错误 The team fixed a critical bug in the search function. 团队修复了搜索功能中的一个严重缺陷。 ★★★
refactor /ˌriːˈfæktə(r)/ v. 重构 We refactored the module to make it easier to read. 我们重构了这个模块使其更易读。 ★★
dependency /dɪˈpendənsi/ n. 依赖(库) This library has too many dependencies. 这个库有太多依赖。 ★★
maintain /meɪnˈteɪn/ v. 维护 A dedicated team maintains the system after launch. 上线后由专门团队维护系统。 ★★

二、测试 Testing(10 词)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
quality assurance (QA) /ˈkwɒləti əˈʃʊərəns/ n. 质量保证 QA ensures the product meets quality standards. 质量保证确保产品达到质量标准。 ★★
unit test /ˈjuːnɪt test/ n. 单元测试 Unit tests check a single function in isolation. 单元测试独立检查单个函数。 ★★
integration test /ˌɪntɪˈɡreɪʃn test/ n. 集成测试 Integration tests verify that components work together. 集成测试验证各组件能否协同工作。 ★★
end-to-end test /ˌend tə ˈend test/ n. 端到端测试 End-to-end tests simulate real user behavior. 端到端测试模拟真实用户行为。 ★★
scenario /səˈnɑːriəʊ/ n. 场景 The test covers the most common user scenarios. 测试覆盖最常见的用户场景。 ★
validate /ˈvælɪdeɪt/ v. 验证 We validate every input before saving it. 保存前我们会验证每一个输入。 ★
coverage /ˈkʌvərɪdʒ/ n. 测试覆盖率 Our test coverage reached 85 percent. 我们的测试覆盖率达到了 85%。 ★★
defect /ˈdiːfekt/ n. 缺陷 The defect was found during the final review. 该缺陷在最终审查时被发现。 ★★
regression /rɪˈɡreʃn/ n. 回归(缺陷) The update caused a regression in the login flow. 这次更新导致登录流程出现回归问题。 ★★
automated testing /ˈɔːtəmeɪtɪd ˈtestɪŋ/ n. 自动化测试 Automated testing saves hours of manual work. 自动化测试节省了大量人工时间。 ★★

三、部署 Deployment(13 词)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
deploy /dɪˈplɔɪ/ v. 部署 We deploy the new version every Friday. 我们每周五部署新版本。 ★★★
deployment /dɪˈplɔɪmənt/ n. 部署 The deployment went smoothly with no downtime. 这次部署很顺利,没有停机。 ★★★
release /rɪˈliːs/ v./n. 发布、版本 The team announced the release of version 2.0. 团队宣布发布 2.0 版本。 ★★★
pipeline /ˈpaɪplaɪn/ n. 流水线 The CI pipeline runs tests on every commit. CI 流水线对每次提交运行测试。 ★★★
continuous integration (CI) /kənˈtɪnjuəs ˌɪntɪˈɡreɪʃn/ n. 持续集成 Continuous integration catches errors early. 持续集成能及早发现错误。 ★★
continuous delivery (CD) /kənˈtɪnjuəs dɪˈlɪvəri/ n. 持续交付 Continuous delivery keeps software ready to release. 持续交付让软件随时可发布。 ★★
environment /ɪnˈvaɪrənmənt/ n. 环境 The code runs in a production environment. 代码运行在生产环境中。 ★★★
staging /ˈsteɪdʒɪŋ/ n. 预发布环境 Test the change in staging before going live. 上线前先在预发布环境测试。 ★★★
production /prəˈdʌkʃn/ n. 生产环境 Never make changes directly in production. 切勿直接在生产环境做修改。 ★★★
artifact /ˈɑːtɪfækt/ n. 构建产物 The build produces a deployable artifact. 构建产生可部署的产物。 ★★
container /kənˈteɪnə(r)/ n. 容器 The app runs inside a Docker container. 应用运行在 Docker 容器中。 ★★
rollback /ˈrəʊlbæk/ n. 回滚 We did a rollback to fix the broken release. 我们回滚以修复出问题的版本。 ★★★
automation /ˌɔːtəˈmeɪʃn/ n. 自动化 Automation reduces human error in delivery. 自动化减少交付中的人为错误。 ★★

四、运维 Operations(15 词)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
monitor /ˈmɒnɪtə(r)/ v./n. 监控 We monitor the server every minute. 我们每分钟监控服务器。 ★★★
metric /ˈmetrɪk/ n. 指标 Response time is a key metric. 响应时间是关键指标。 ★★★
latency /ˈleɪtənsi/ n. 延迟 Low latency is critical for real-time applications. 低延迟对实时应用至关重要。 ★★
log /lɒɡ/ n./v. 日志 Check the log to find the error. 查看日志找出错误。 ★★★
alert /əˈlɜːt/ n./v. 告警 The system sends an alert when the disk is full. 磁盘满时系统会发出告警。 ★★★
incident /ˈɪnsɪdənt/ n. 事件、故障 The team resolved the incident in 30 minutes. 团队在 30 分钟内解决了事件。 ★★★
outage /ˈaʊtɪdʒ/ n. 宕机、中断 A power failure caused a two-hour outage. 停电导致两小时宕机。 ★★
downtime /ˈdaʊntaɪm/ n. 停机时间 The website aims for zero downtime. 网站力求零停机。 ★★★
uptime /ˈʌptaɪm/ n. 正常运行时间 We promise 99.9 percent uptime to our customers. 我们向客户承诺 99.9% 的正常运行时间。 ★★★
on-call /ˌɒn ˈkɔːl/ adj. 值班的 I am on-call this weekend. 这周末我值班。 ★★
troubleshooting /ˈtrʌblˌʃuːtɪŋ/ n. 故障排查 Troubleshooting requires patience and logic. 故障排查需要耐心和逻辑。 ★★
root cause /ruːt kɔːz/ n. 根本原因 We found the root cause of the crash. 我们找到了崩溃的根本原因。 ★★
reliability /rɪˌlaɪəˈbɪləti/ n. 可靠性 Reliability matters more than new features. 可靠性比新功能更重要。 ★★
availability /əˌveɪləˈbɪləti/ n. 可用性 High availability keeps the service online almost all the time. 高可用性让服务几乎始终在线。 ★
scalability /ˌskeɪləˈbɪləti/ n. 可扩展性 Scalability allows the system to grow with user demand. 可扩展性让系统能随用户需求增长。 ★

五、过程与协作 Process(2 词)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
feedback loop /ˈfiːdbæk luːp/ n. 反馈回路 A short feedback loop helps teams improve quickly. 短的反馈回路帮助团队快速改进。 ★★
continuous improvement /kənˈtɪnjuəs ɪmˈpruːvmənt/ n. 持续改进 DevOps encourages continuous improvement of processes. DevOps 鼓励流程的持续改进。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
software development lifecycle /ˈsɒftweə(r) dɪˈveləpmənt ˈlaɪfsaɪkl/ 软件开发生命周期 The software development lifecycle has many phases. 软件开发生命周期包含多个阶段。 ★★★
go live /ɡəʊ laɪv/ 上线、投入使用 The system will go live next Monday. 系统下周一上线。 ★★★
version control system /ˈvɜːʃn kənˈtrəʊl ˈsɪstəm/ 版本控制系统 Git is a distributed version control system. Git 是分布式版本控制系统。 ★★
code review /kəʊd rɪˈvjuː/ 代码评审 Please submit your changes for code review. 请提交你的改动进行代码评审。 ★★★
pull request /pʊl rɪˈkwest/ 合并请求/拉取请求 She opened a pull request for the new feature. 她为新功能提交了一个合并请求。 ★★★
quality assurance /ˈkwɒləti əˈʃʊərəns/ 质量保证 Quality assurance is part of every release. 质量保证是每次发布的一部分。 ★★
test coverage /test ˈkʌvərɪdʒ/ 测试覆盖率 We track test coverage in every sprint. 我们在每个迭代中都追踪测试覆盖率。 ★★
continuous integration /kənˈtɪnjuəs ˌɪntɪˈɡreɪʃn/ 持续集成 Continuous integration builds the code on every push. 持续集成在每次推送时构建代码。 ★★★
continuous delivery /kənˈtɪnjuəs dɪˈlɪvəri/ 持续交付 Continuous delivery automates the release process. 持续交付使发布流程自动化。 ★★★
root cause analysis /ruːt kɔːz əˈnæləsɪs/ 根本原因分析 We did a root cause analysis after the outage. 宕机后我们做了根本原因分析。 ★★
on-call rotation /ˌɒn ˈkɔːl rəʊˈteɪʃn/ 值班轮换 The on-call rotation includes all senior engineers. 值班轮换包括所有资深工程师。 ★★
feedback loop /ˈfiːdbæk luːp/ 反馈回路 Short feedback loops keep teams aligned. 短的反馈回路让团队保持一致。 ★★
around the clock /əˈraʊnd ðə klɒk/ 全天候、24 小时不间断 The team monitors the service around the clock. 团队全天候监控服务。 ★★
best practice /best ˈpræktɪs/ 最佳实践 Following best practices reduces risk. 遵循最佳实践能降低风险。 ★★

第 2 课 需求分析与产品规划

Lesson 2: Requirements & Product Planning

学习目标:学会用英文描述"需求从哪里来、如何拆解、如何排优先级"。

📖 英文阅读(Reading)

原文:

Every software project starts with a problem to solve. Product teams translate that problem into clear requirements, then break them down into small, testable pieces called user stories. A typical user story sounds like: "As a customer, I want to reset my password, so that I can access my account again."

All stories are kept in a backlog — a prioritized list of everything the team might build. The team plans its work in short sprints, usually two weeks long, and pulls the most important stories into each sprint. During estimation, developers discuss how much effort each story needs, and the product owner confirms the acceptance criteria — the conditions a story must meet before it is considered done.

Priorities change all the time. A new request from an important client can jump to the top of the backlog, while low-value items sink to the bottom or get dropped. Good product planning is the art of saying "no" to most ideas, so the team can focus on what really matters. Every few weeks, the team reviews progress against milestones and adjusts the plan.

中文翻译:

每个软件项目都始于一个待解决的问题。产品团队把问题转化为清晰的需求,再拆解成小而可测试的单元——用户故事(user story)。一个典型的用户故事长这样:"作为客户,我想重置我的密码,以便我能重新访问我的账户。"

所有故事都保存在待办清单(backlog)里——这是团队可能构建的所有事项的优先级列表。团队以短冲刺(sprint)(通常两周)为单位规划工作,把最重要的故事拉进每个冲刺。在估算环节,开发者讨论每个故事需要多少工作量,产品负责人则确认验收标准(acceptance criteria)——即一个故事被视为"完成"前必须满足的条件。

优先级随时会变。一个重要客户的新请求可以跳到待办清单顶部,低价值的项则沉到底部或被丢弃。好的产品规划是一门"对大多数想法说不"的艺术,这样团队才能聚焦真正重要的事。每隔几周,团队对照里程碑检查进度并调整计划。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
user story /ˈjuːzə(r) ˈstɔːri/ n. 用户故事 Each user story should be small and testable. 每个用户故事都应当小而可测试。 ★★
backlog /ˈbæklɒɡ/ n. 待办清单、需求池 The backlog currently has 40 items. 待办清单目前有 40 项。 ★★★
sprint /sprɪnt/ n. 冲刺、迭代周期 Our team delivers a demo every sprint. 我们团队每个冲刺都做一次演示。 ★★★
estimation /ˌestɪˈmeɪʃn/ n. 估算 Estimation helps us plan the sprint. 估算帮助我们规划冲刺。 ★★
acceptance criteria /əkˈseptəns kraɪˈtɪəriə/ n. 验收标准 The story is done only when it meets the acceptance criteria. 故事只有满足验收标准才算完成。 ★★
priority /praɪˈɒrəti/ n. 优先级 Security fixes always have the highest priority. 安全修复总是拥有最高优先级。 ★★★
milestone /ˈmaɪlstəʊn/ n. 里程碑 We reached the first milestone ahead of schedule. 我们提前达成了第一个里程碑。 ★★
roadmap /ˈrəʊdmæp/ n. 路线图 The roadmap shows what we plan to ship this year. 路线图展示了我们今年计划发布的内容。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
break down /breɪk daʊn/ 拆解 We break down big features into small stories. 我们把大功能拆解成小故事。 ★★★
product owner /ˈprɒdʌkt ˈəʊnə(r)/ 产品负责人 The product owner decides what goes into the sprint. 产品负责人决定冲刺里放什么。 ★★
scope creep /skəʊp kriːp/ 范围蔓延(需求失控) Scope creep delayed the release by a month. 范围蔓延让发布推迟了一个月。 ★★
on track /ɒn træk/ 按计划进行 The project is on track for a June launch. 项目按计划将在六月上线。 ★★

第 3 课 敏捷开发与 Scrum

Lesson 3: Agile Development & Scrum

学习目标:掌握敏捷会议、流程与团队节奏的英文表达。

📖 英文阅读(Reading)

原文:

Agile is a philosophy of building software in small, continuous steps and responding to change quickly. Scrum is the most popular agile framework. It organizes work into sprints and relies on a few simple ceremonies: sprint planning, the daily standup, the sprint review, and the retrospective.

Each day, the team meets for a fifteen-minute standup meeting. Everyone answers three questions: What did I do yesterday? What will I do today? Are there any blockers? The standup keeps everyone aligned without long meetings.

At the end of a sprint, the team shows completed work in the sprint review, then holds a retrospective to discuss what went well and what should improve. Teams also track their velocity — the amount of work they complete per sprint — to predict future capacity, and they watch the burndown chart to see if the team is on pace to finish the sprint's committed work. Epics are large stories that span several sprints, while small tasks are often broken into fine-grained tickets. The definition of done states clearly when a task is truly finished, including tests, documentation, and deployment.

中文翻译:

敏捷(Agile)是一种以小步快跑、快速响应变化的方式构建软件的理念。Scrum 是最流行的敏捷框架:它把工作组织成冲刺,并依靠几个简单的仪式——冲刺规划(sprint planning)、每日站会(daily standup)、冲刺评审(sprint review)和回顾会(retrospective)。

团队每天开 15 分钟的站会。每个人回答三个问题:我昨天做了什么?我今天要做什么?有没有阻塞项(blocker)?站会让团队无需冗长会议就能保持一致。

冲刺结束时,团队在冲刺评审中展示已完成的工作,然后开回顾会(retrospective)讨论哪些做得好、哪些需要改进。团队还会追踪速率(velocity)——每个冲刺完成的工作量——用来预测未来产能;同时观察燃尽图(burndown chart),判断团队能否按计划完成冲刺承诺的工作。史诗(epic)是跨越多个冲刺的大故事,小任务则常被拆成细粒度的工单(ticket)。完成定义(definition of done)明确了一个任务真正完成的标准,包括测试、文档和部署。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
agile /ˈædʒaɪl/ adj. 敏捷的 Agile teams ship small changes frequently. 敏捷团队频繁交付小变更。 ★★★
Scrum /skrʌm/ n. Scrum(敏捷框架) Scrum is built on sprints and ceremonies. Scrum 建立在冲刺和仪式之上。 ★★★
standup /ˈstændʌp/ n. 每日站会 Our standup starts at ten sharp. 我们的站会十点整开始。 ★★★
retrospective /ˌretrəˈspektɪv/ n. 回顾会 We changed our process after the retrospective. 回顾会后我们调整了流程。 ★★
velocity /vəˈlɒsəti/ n. 团队速率 Our velocity dropped because of the refactor. 由于重构,我们的速率下降了。 ★★
blocker /ˈblɒkə(r)/ n. 阻塞项 The missing API key is a blocker for me. 缺少 API 密钥是我的阻塞项。 ★★
epic /ˈepɪk/ n. 史诗级需求 The billing overhaul is a three-sprint epic. 计费系统重构是一个跨三个冲刺的史诗需求。 ★★
definition of done /ˌdefɪˈnɪʃn əv dʌn/ n. 完成定义 Tests and docs are part of our definition of done. 测试和文档是我们完成定义的一部分。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
daily standup /ˈdeɪli ˈstændʌp/ 每日站会 The daily standup is kept under 15 minutes. 每日站会控制在 15 分钟以内。 ★★★
sprint planning /sprɪnt ˈplænɪŋ/ 冲刺规划会 Sprint planning sets the goal for the next two weeks. 冲刺规划会设定接下来两周的目标。 ★★
burndown chart /ˈbɜːndaʊn tʃɑːt/ 燃尽图 The burndown chart shows we are ahead of schedule. 燃尽图显示我们进度超前。 ★★
lessons learned /ˈlesnz ˈlɜːnd/ 经验教训 We documented the lessons learned from the failed release. 我们记录了失败发布中的经验教训。 ★★

第 4 课 版本控制与 Git 实操

Lesson 4: Version Control with Git

学习目标:掌握 Git 日常操作的英文动词与名词,能看懂英文 Git 文档和报错。

📖 英文阅读(Reading)

原文:

Git is a distributed version control tool that every developer uses daily. A project lives in a repository, and each developer works with a local clone of it. To share work, you push your commits to the remote, and to get others' changes you pull or fetch them. A tag marks an important point in history, such as a release version.

Conflicts are a normal part of collaboration. When two people change the same file, Git cannot merge them automatically, and you must resolve the merge conflict by hand. The rebase command rewrites history by replaying your commits on top of another branch, which keeps the history linear and clean.

Many teams adopt trunk-based development or a feature-branch workflow: create a branch, make changes, push, and open a pull request. Before merging, the branch must pass automated checks and code review. If you need to put work aside temporarily, you can stash it. You can also fork a repository to make your own copy, then submit changes back through a pull request. A clean, well-organized commit history makes debugging and code review much easier.

中文翻译:

Git 是每个开发者每天都会用到的分布式版本控制工具。项目存放在仓库(repository)中,每个开发者操作它的本地克隆(clone)。要分享工作成果,就把提交推送(push)到远程;要获取别人的改动,就拉取(pull)或获取(fetch)。标签(tag)用来标记历史上的重要节点,比如发布版本。

冲突是协作中的正常现象。当两个人修改了同一个文件,Git 无法自动合并,你必须手动解决合并冲突(merge conflict)。变基(rebase)命令通过把你的提交"重放"到另一个分支之上来重写历史,让历史保持线性整洁。

许多团队采用主干开发(trunk-based development)或功能分支工作流(feature-branch workflow):创建分支、修改、推送、提交合并请求。合并前,分支必须通过自动化检查和代码评审。如果想把工作暂时放到一边,可以暂存(stash);也可以分叉(fork)一个仓库做成自己的副本,再通过合并请求把改动提交回去。干净、有组织的提交历史会让调试和代码评审轻松得多。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
repository /rɪˈpɒzətri/ n. 仓库(代码库) Clone the repository to your laptop. 把仓库克隆到你的笔记本上。 ★★★
clone /kləʊn/ v./n. 克隆 I cloned the repo and ran it locally. 我克隆了仓库并在本地运行。 ★★★
push /pʊʃ/ v. 推送 Don't push directly to the main branch. 不要直接推送到主分支。 ★★★
pull /pʊl/ v. 拉取 Pull the latest changes before you start. 开始前先拉取最新改动。 ★★★
fetch /fetʃ/ v. 获取(远程更新) Fetch the remote updates first. 先获取远程更新。 ★★
tag /tæɡ/ n./v. 标签 We tag each release with a version number. 我们给每个发布打上版本号标签。 ★★
merge conflict /mɜːdʒ ˈkɒnflɪkt/ n. 合并冲突 I resolved the merge conflict in the config file. 我解决了配置文件里的合并冲突。 ★★★
rebase /ˌriːˈbeɪs/ v. 变基 Rebase your branch onto the latest main. 把你的分支变基到最新的主干上。 ★★
stash /stæʃ/ v. 暂存(未提交改动) Stash your changes before switching branches. 切换分支前先暂存你的改动。 ★★
fork /fɔːk/ v./n. 分叉 Fork the project and send us a pull request. 分叉这个项目并向我们提交合并请求。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
feature branch /ˈfiːtʃə(r) brɑːntʃ/ 功能分支 Create a feature branch for every task. 为每个任务创建功能分支。 ★★★
main branch /meɪn brɑːntʃ/ 主分支 The main branch must always be deployable. 主分支必须始终可部署。 ★★★
trunk-based development /trʌŋk beɪst dɪˈveləpmənt/ 主干开发 Trunk-based development avoids long-lived branches. 主干开发避免长期存在的分支。 ★★
track changes /træk ˈtʃeɪndʒɪz/ 追踪变更 Version control lets us track every change. 版本控制让我们能追踪每一次变更。 ★★

第 5 课 代码评审与团队协作

Lesson 5: Code Review & Collaboration

学习目标:学会用英文提意见、批准、驳回,掌握协作场景的礼貌表达。

📖 英文阅读(Reading)

原文:

Code review is one of the most effective ways to keep code quality high. When you open a pull request, teammates read your diff, add comments on specific lines, and finally approve or reject the change. A reviewer might ask, "Could you extract this logic into a helper function?" or "This edge case is not covered — please add a test."

Good reviews are specific and kind. Instead of saying "this is bad", reviewers explain the impact of a problem and suggest a concrete alternative. Authors should respond to every comment, either fixing the issue or explaining why the current approach is fine. Disagreements are normal; the team resolves them with clear reasoning, not ego.

Beyond review, teams use pair programming — two developers working on one screen — to share knowledge and catch mistakes instantly. A strong culture of knowledge sharing means code is understood by more than one person, and no single engineer becomes an single point of failure (a person whose absence blocks everything). Written documentation for unusual decisions helps new members ramp up quickly.

中文翻译:

代码评审是保持代码质量最高效的方式之一。当你提交合并请求(pull request)时,队友阅读你的改动、在具体代码行上加评论,最终批准或驳回这个变更。评审者可能会问:"能不能把这个逻辑提取成一个辅助函数?"或"这个边界情况没覆盖到——请加个测试。"

好的评审具体而友善。评审者不会说"这写得很糟",而是解释问题的影响并给出具体的替代方案。作者应当回复每一条评论——要么修复问题,要么解释为什么当前做法没问题。意见分歧很正常,团队用清晰的论证而不是意气来解决。评审之外,团队还使用结对编程(pair programming)——两个开发者共用一块屏幕——来分享知识、即时发现错误。强大的知识共享文化意味着代码不止一个人理解,没有任何工程师成为单点故障(一个缺席就会阻塞一切的人)。为非典型决策写文档能帮助新成员快速上手。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
comment /ˈkɒment/ n./v. 评论 Please comment on the diff before merging. 合并前请对改动提出评论。 ★★★
approve /əˈpruːv/ v. 批准 I will approve the pull request once tests pass. 测试通过后我会批准这个合并请求。 ★★★
reject /rɪˈdʒekt/ v. 驳回、拒绝 The reviewer rejected the change due to missing tests. 评审者因缺少测试驳回了这个改动。 ★★
extract /ɪkˈstrækt/ v. 提取(代码) Extract the duplicate code into a shared module. 把重复代码提取成共享模块。 ★★
impact /ˈɪmpækt/ n. 影响 This change has a big impact on performance. 这个改动对性能影响很大。 ★★★
alternative /ɔːlˈtɜːnətɪv/ n. 替代方案 Can you suggest an alternative approach? 你能建议一个替代方案吗? ★★
pair programming /peə(r) ˈprəʊɡræmɪŋ/ n. 结对编程 Pair programming is great for onboarding. 结对编程非常适合新人上手。 ★★
single point of failure /ˈsɪŋɡl pɔɪnt əv ˈfeɪljə(r)/ n. 单点故障 Shared knowledge reduces single points of failure. 共享知识能减少单点故障。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
address a comment /əˈdres ə ˈkɒment/ 回应评论 Please address all review comments. 请回应所有评审评论。 ★★
knowledge sharing /ˈnɒlɪdʒ ˈʃeərɪŋ/ 知识共享 Knowledge sharing makes the team stronger. 知识共享让团队更强大。 ★★
ramp up /ræmp ʌp/ 快速上手 The docs help new hires ramp up fast. 文档帮助新员工快速上手。 ★★
work together /wɜːk təˈɡeðə(r)/ 协作 Developers and QA work together closely. 开发与测试紧密协作。 ★★

第 6 课 代码质量与重构

Lesson 6: Code Quality & Refactoring

学习目标:掌握描述"好代码/坏代码"的英文词汇,学会讨论技术债务。

📖 英文阅读(Reading)

原文:

Clean code is easy to read, easy to test, and easy to change. Readability matters because code is read far more often than it is written. Maintainability is the long-term measure of how cheaply a team can modify the code without breaking things. Complexity is the enemy: deeply nested functions and huge classes are hard to follow and even harder to test. Duplication — the same logic copied in many places — makes every future change risky, because you must remember to update every copy.

When teams move fast, they often accumulate technical debt: quick solutions that work today but will cost more tomorrow. Debt is not always bad — sometimes it is the right trade-off — but it must be tracked and paid down. Common code smells that signal trouble include long parameter lists, god classes, and dead code that nobody uses. Spaghetti code refers to tangled, unstructured logic that is almost impossible to maintain.

The remedy is refactoring: changing the internal structure without changing behavior, one small step at a time. A good style guide and consistent naming keep the codebase uniform, and testability — designing code so it can be tested easily — makes refactoring safe.

中文翻译:

整洁的代码易读、易测、易改。可读性(readability)很重要,因为代码被阅读的频率远高于被编写的频率。可维护性(maintainability)衡量的是团队长期以多低的成本修改代码而不破坏其他部分。复杂度(complexity)是敌人:深层嵌套的函数和巨型类难以理解,更难以测试。重复(duplication)——同样的逻辑在多处复制——让未来每一次修改都充满风险,因为你必须记得更新每一处副本。

团队赶进度时,常常积累技术债务(technical debt):今天能用、明天要付代价的快捷方案。债务不总是坏事——有时它是合理的取舍——但必须被记录并偿还。常见的代码坏味道(code smell)包括过长的参数列表、上帝类(god class)、没人用的死代码(dead code)。意大利面条式代码(spaghetti code)指的是纠缠不清、几乎无法维护的无结构逻辑。

解药是重构(refactoring):在行为不变的前提下改变内部结构,一次一小步。良好的风格指南(style guide)和一致的命名让代码库统一,可测试性(testability)——让代码容易被测试的设计——让重构变得安全。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
readability /ˌriːdəˈbɪləti/ n. 可读性 Naming matters most for readability. 命名对可读性最重要。 ★★
maintainability /meɪnˌteɪnəˈbɪləti/ n. 可维护性 Maintainability beats clever tricks. 可维护性胜过花哨技巧。 ★★
complexity /kəmˈpleksəti/ n. 复杂度 We simplified the logic to reduce complexity. 我们简化逻辑以降低复杂度。 ★★
duplication /ˌdjuːplɪˈkeɪʃn/ n. 重复(代码) Duplication makes fixes twice as hard. 重复让修复加倍困难。 ★★
technical debt /ˈteknɪkl det/ n. 技术债务 We pay down technical debt every quarter. 我们每个季度偿还技术债务。 ★★★
code smell /kəʊd smel/ n. 代码坏味道 Long parameter lists are a code smell. 过长的参数列表是一种坏味道。 ★★
dead code /ded kəʊd/ n. 死代码 Remove dead code to avoid confusion. 删除死代码以免混淆。 ★★
testability /ˌtestəˈbɪləti/ n. 可测试性 Dependency injection improves testability. 依赖注入能提升可测试性。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
clean code /kliːn kəʊd/ 整洁代码 Clean code is simple and obvious. 整洁代码简单而清晰。 ★★★
spaghetti code /spəˈɡeti kəʊd/ 面条式代码(混乱代码) The old module is spaghetti code. 那个旧模块是面条式代码。 ★★
style guide /staɪl ɡaɪd/ 风格指南 Follow the style guide in every PR. 每个合并请求都要遵循风格指南。 ★★
pay down debt /peɪ daʊn det/ 偿还(技术)债务 We dedicate one day a month to pay down debt. 我们每月花一天偿还技术债务。 ★★

第 7 课 软件架构与设计模式

Lesson 7: Architecture & Design Patterns

学习目标:掌握描述系统结构和设计原则的英文词汇。

📖 英文阅读(Reading)

原文:

Architecture is the high-level structure of a system — how it is divided into modules and components, and how they communicate. A good architecture makes the system easy to understand, easy to change, and easy to scale. Layered architecture separates concerns: presentation, business logic, and data access each live in their own layer.

The goal of good design is decoupling: components should depend on interfaces and contracts, not on each other's internals, so that one part can change without breaking the rest. Abstraction hides implementation details behind a simple interface, letting developers reason about the system at a higher level. Extensibility means you can add new features without rewriting existing code.

Design patterns are proven, reusable solutions to common problems. The factory pattern creates objects without exposing the creation logic; the observer pattern lets one object notify others when its state changes; the singleton pattern ensures a class has only one instance. Patterns give teams a shared vocabulary — saying "use a factory here" communicates an entire design in four words. Robustness — the ability to keep working under unexpected conditions — is also a key architectural goal, supported by fault tolerance and graceful fallback strategies.

中文翻译:

架构是系统的高层结构——如何划分成模块和组件,以及它们如何通信。好的架构让系统易于理解、易于修改、易于扩展。分层架构(layered architecture)分离关注点:表现层、业务逻辑层、数据访问层各自独立。

优秀设计的目标是解耦(decoupling):组件应依赖接口和契约,而不是彼此的内部实现,这样一部分改变时不会破坏其他部分。抽象(abstraction)把实现细节隐藏在简单接口之后,让开发者能在更高层面思考系统。可扩展性(extensibility)意味着无需重写现有代码就能添加新功能。

设计模式(design pattern)是经过验证的、可复用的常见问题解决方案。工厂模式(factory pattern)在不暴露创建逻辑的情况下创建对象;观察者模式(observer pattern)让一个对象在状态变化时通知其他对象;单例模式(singleton pattern)确保一个类只有一个实例。模式给团队提供了共享词汇——说"这里用工厂"四个词就传达了一整套设计。健壮性(robustness)——在意外条件下继续工作的能力——也是关键的架构目标,由容错(fault tolerance)和优雅的降级(fallback)策略支撑。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
architecture /ˈɑːkɪtektʃə(r)/ n. 架构 The architecture separates concerns clearly. 该架构清晰分离了关注点。 ★★★
module /ˈmɒdjuːl/ n. 模块 Each module has a single responsibility. 每个模块只有单一职责。 ★★★
component /kəmˈpəʊnənt/ n. 组件 Components communicate through interfaces. 组件通过接口通信。 ★★★
interface /ˈɪntəfeɪs/ n. 接口 Program to an interface, not an implementation. 面向接口编程,而非面向实现。 ★★★
decouple /ˌdiːˈkʌpl/ v. 解耦 Decouple the services to reduce risk. 解耦各服务以降低风险。 ★★
abstraction /æbˈstrækʃn/ n. 抽象 A good abstraction hides complexity. 好的抽象隐藏复杂性。 ★★
extensibility /ɪkˌstensəˈbɪləti/ n. 可扩展性(架构层面) Extensibility lets us add plugins easily. 可扩展性让我们能轻松添加插件。 ★★
robustness /rəʊˈbʌstnəs/ n. 健壮性 Robustness testing feeds bad data to the system. 健壮性测试向系统输入异常数据。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
design pattern /dɪˈzaɪn ˈpætn/ 设计模式 The observer pattern fits this use case. 观察者模式适合这个场景。 ★★★
layered architecture /ˈleɪəd ˈɑːkɪtektʃə(r)/ 分层架构 A layered architecture keeps code organized. 分层架构让代码井井有条。 ★★
single responsibility /ˈsɪŋɡl rɪˌspɒnsəˈbɪləti/ 单一职责 Each class should have a single responsibility. 每个类应当只有单一职责。 ★★
fault tolerance /fɔːlt ˈtɒlərəns/ 容错 Fault tolerance keeps the service alive during failures. 容错让服务在故障期间保持存活。 ★★

第 8 课 单元测试与测试金字塔

Lesson 8: Unit Testing & Test Pyramid

学习目标:掌握编写、组织、评估单元测试的英文词汇。

📖 英文阅读(Reading)

原文:

Unit tests are the foundation of the test pyramid: they are fast, cheap, and numerous, while integration and end-to-end tests sit above them. A good unit test focuses on one behavior, uses clear assertions to check expected results, and runs in milliseconds. Tests that depend on external services are made reliable with mocks and stubs — fake objects that replace real dependencies.

To keep tests fast and independent, developers use test fixtures to set up data, and parameterized tests to run the same logic with many different inputs. Each test should verify a single test case, and the whole collection is called a test suite. A test runner executes the suite and reports which tests passed and which failed. Slow, unreliable tests are worse than no tests: flaky tests pass sometimes and fail randomly, destroying trust in the suite. Teams treat a failing test as an urgent problem — never ignore it or skip it casually, because tests are the safety net that makes refactoring and rapid change possible. A fast, trustworthy unit test suite is the single biggest productivity tool a developer can have.

中文翻译:

单元测试是测试金字塔(test pyramid)的基石:它们快速、便宜、数量众多,集成测试和端到端测试位于其上。好的单元测试聚焦一个行为,用清晰的断言(assertion)检查预期结果,毫秒级运行。依赖外部服务的测试通过模拟对象(mock)和桩(stub)——替代真实依赖的假对象——来保证可靠性。

为了保持测试快速、独立,开发者用测试夹具(fixture)准备数据,用参数化测试(parameterized test)用大量不同输入运行同一逻辑。每个测试应验证一个测试用例(test case),全部测试合称测试套件(test suite)。测试运行器(test runner)执行套件并报告哪些通过、哪些失败。慢而不稳定的测试比没有测试更糟:不稳定的(flaky)测试时好时坏,摧毁人们对套件的信任。团队把失败的测试当作紧急问题——绝不随意忽略或跳过,因为测试是让重构和快速变更成为可能的安全网。快速、可信的单元测试套件是开发者能拥有的最大生产力工具。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
assertion /əˈsɜːʃn/ n. 断言 The assertion checks the returned value. 断言检查返回值。 ★★★
mock /mɒk/ n./v. 模拟对象 Mock the database in unit tests. 在单元测试中模拟数据库。 ★★★
stub /stʌb/ n. 桩(测试替身) Use a stub for the payment API. 为支付接口使用一个桩。 ★★
fixture /ˈfɪkstʃə(r)/ n. 测试夹具、固定数据 The fixture loads sample users into the test. 夹具把示例用户加载进测试。 ★★
test case /test keɪs/ n. 测试用例 This test case covers the empty-input scenario. 这个测试用例覆盖空输入场景。 ★★★
test suite /test swiːt/ n. 测试套件 The full test suite runs in three minutes. 完整测试套件三分钟跑完。 ★★
test runner /test ˈrʌnə(r)/ n. 测试运行器 The test runner stops at the first failure. 测试运行器在第一个失败处停止。 ★★
flaky /ˈfleɪki/ adj. 不稳定的(测试) Flaky tests waste everyone's time. 不稳定的测试浪费所有人的时间。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
test pyramid /test ˈpɪrəmɪd/ 测试金字塔 The test pyramid guides our testing strategy. 测试金字塔指导我们的测试策略。 ★★
edge case /edʒ keɪs/ 边界情况 Handle the edge case of an empty list. 处理空列表的边界情况。 ★★★
pass / fail a test /pɑːs feɪl ə test/ 测试通过/失败 The new test fails until we fix the bug. 在我们修复 bug 之前新测试会失败。 ★★★
run the tests /rʌn ðə tests/ 运行测试 Run the tests before pushing. 推送前先运行测试。 ★★★

第 9 课 集成与端到端测试

Lesson 9: Integration & End-to-End Testing

学习目标:掌握"测试系统如何协同工作"的英文词汇,区分各测试层次。

📖 英文阅读(Reading)

原文:

While unit tests check pieces in isolation, integration tests verify that pieces work together. They connect real modules — a service talking to a database, an API calling another API — and confirm the contract between them is honored: the request format matches, the response is correct, and errors are handled. API testing sends real HTTP requests to endpoints and checks status codes, response bodies, and headers.

End-to-end (E2E) tests go further: they drive the whole application through the user interface, using browser automation to click buttons, fill forms, and verify pages. UI testing catches problems that unit tests cannot, such as a broken layout or a wrong link. Because E2E tests are slow and fragile, teams run a small, critical set — a smoke test — on every release, and keep the full suite for nightly runs. Tests often run headless (without a visible browser window) to save resources. The team also creates a clean test environment with test data, so results are repeatable and one test never interferes with another. A balanced strategy: many fast unit tests, a moderate layer of integration tests, and a few high-value E2E tests.

中文翻译:

单元测试孤立地检查零件,而集成测试验证零件能否协同工作。它连接真实模块——服务对接数据库、API 调用另一个 API——并确认它们之间的契约得到遵守:请求格式匹配、响应正确、错误被处理。接口测试(API testing)向端点发送真实 HTTP 请求,检查状态码、响应体和请求头。

端到端测试(E2E)更进一步:它通过用户界面驱动整个应用,用浏览器自动化点击按钮、填写表单、验证页面。界面测试(UI testing)能发现单元测试发现不了的问题,比如布局损坏或链接错误。由于 E2E 测试慢且脆弱,团队每次发布只跑一小部分关键用例——冒烟测试(smoke test)——完整套件留到夜间运行。测试常常以无头(headless)方式运行(不显示浏览器窗口)以节省资源。团队还会准备干净的测试环境和测试数据,保证结果可复现、测试之间互不干扰。一个均衡的策略是:大量快速的单元测试、中等数量的集成测试、少数高价值的 E2E 测试。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
contract /ˈkɒntrækt/ n. 契约、接口约定 The API contract defines the request format. API 契约定义了请求格式。 ★★
endpoint /ˈendpɔɪnt/ n. 端点(接口地址) Test the login endpoint with valid credentials. 用有效凭据测试登录端点。 ★★★
browser automation /ˈbraʊzə(r) ˌɔːtəˈmeɪʃn/ n. 浏览器自动化 Browser automation simulates user clicks. 浏览器自动化模拟用户点击。 ★★
smoke test /sməʊk test/ n. 冒烟测试 Run a smoke test after every deployment. 每次部署后运行冒烟测试。 ★★
headless /ˈhedləs/ adj. 无头的(无界面运行) Run the browser in headless mode on CI. 在 CI 上以无头模式运行浏览器。 ★★
UI /ˌjuː ˈaɪ/ n. 用户界面 The UI test verifies the checkout flow. 界面测试验证结账流程。 ★★★
test environment /test ɪnˈvaɪrənmənt/ n. 测试环境 Spin up a clean test environment for the suite. 为测试套件启动一个干净的测试环境。 ★★
repeatable /rɪˈpiːtəbl/ adj. 可重复的 Tests must be repeatable and deterministic. 测试必须可重复且结果确定。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
API testing /ˌeɪ piː ˈaɪ ˈtestɪŋ/ 接口测试 API testing checks status codes and payloads. 接口测试检查状态码和响应体。 ★★★
user interface /ˈjuːzə(r) ˈɪntəfeɪs/ 用户界面 The user interface must be accessible. 用户界面必须易于访问。 ★★★
status code /ˈsteɪtəs kəʊd/ 状态码 A 500 status code means a server error. 500 状态码表示服务器错误。 ★★★
end to end /end tə end/ 端到端 We test the payment flow end to end. 我们端到端地测试支付流程。 ★★★

第 10 课 自动化测试框架

Lesson 10: Automation Testing Frameworks

学习目标:掌握自动化测试基础设施的英文词汇(脚本、并行、门禁、CI 集成)。

📖 英文阅读(Reading)

原文:

Automated testing turns manual checks into code. A test framework provides the structure: how to write tests, how to group them, and how to report results. Popular frameworks include JUnit for Java, pytest for Python, and Jest for JavaScript. Teams write test scripts that are committed to the repository and run on every code change, which is the essence of continuous testing.

Modern test runs are parallelized across many machines to finish in minutes, and results are collected into clear reports with pass/fail summaries and screenshots on failure. A quality gate blocks merging when coverage drops below a threshold or when critical tests fail — nobody can merge code that breaks the build.

Frameworks also support retries for flaky tests, test harnesses that simulate the surrounding environment, and mocking libraries to isolate components. The nightly build runs the full, slower suite once a day, while the fast subset runs on every commit. The goal is a short feedback cycle: the moment you introduce a bug, a test fails, and you know exactly where the problem is. Automation is not just about speed — it is about consistency: the same checks run the same way, every time, for every developer.

中文翻译:

自动化测试把人工检查变成代码。测试框架(test framework)提供结构:如何写测试、如何分组、如何报告结果。流行的框架包括 Java 的 JUnit、Python 的 pytest、JavaScript 的 Jest。团队把测试脚本(test script)提交进仓库,在每次代码变更时运行——这就是持续测试(continuous testing)的本质。

现代测试在多台机器上并行(parallelize)运行,几分钟内跑完;结果汇总成清晰的报告(report),包含通过/失败汇总和失败时的截图(screenshot)。质量门禁(quality gate)在覆盖率低于阈值或关键测试失败时阻止合并——任何人都不能合并破坏构建的代码。

框架还支持对不稳定测试的重试(retry)、模拟周边环境的测试工具(test harness)以及隔离组件的模拟库(mocking library)。夜间构建(nightly build)每天运行一次完整但较慢的套件,快速子集则在每次提交时运行。目标是一个短反馈周期(short feedback cycle):你一旦引入 bug,测试立刻失败,你马上就知道问题在哪。自动化不只是为了速度——更是为了一致性(consistency):同样的检查以同样的方式、每次、为每个开发者运行。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
framework /ˈfreɪmwɜːk/ n. 框架 Choose a framework your team already knows. 选择团队已经熟悉的框架。 ★★★
script /skrɪpt/ n. 脚本 The test script runs in five minutes. 测试脚本五分钟跑完。 ★★★
parallelize /ˈpærəlelaɪz/ v. 并行化 Parallelize the suite across eight workers. 把测试套件并行到 8 个工作进程上。 ★★
report /rɪˈpɔːt/ n. 测试报告 The report shows 12 passed and 1 failed. 报告显示 12 个通过、1 个失败。 ★★★
quality gate /ˈkwɒləti ɡeɪt/ n. 质量门禁 The quality gate blocks below 80% coverage. 质量门禁拦截覆盖率低于 80% 的合并。 ★★
retry /ˌriːˈtraɪ/ v./n. 重试 The runner retries flaky tests twice. 运行器对不稳定测试重试两次。 ★★
harness /ˈhɑːnəs/ n. 测试工具/测试台 The harness provides a fake database. 测试工具提供假数据库。 ★
consistency /kənˈsɪstənsi/ n. 一致性 Automation gives us consistency. 自动化带来一致性。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
continuous testing /kənˈtɪnjuəs ˈtestɪŋ/ 持续测试 Continuous testing catches bugs early. 持续测试能及早发现 bug。 ★★
nightly build /ˈnaɪtli bɪld/ 夜间构建 The nightly build runs the full suite. 夜间构建运行完整测试套件。 ★★
short feedback cycle /ʃɔːt ˈfiːdbæk ˈsaɪkl/ 短反馈周期 A short feedback cycle speeds up development. 短反馈周期加速开发。 ★★
in parallel /ɪn ˈpærəlel/ 并行地 The tests run in parallel on CI. 测试在 CI 上并行运行。 ★★

第 11 课 持续集成(CI)

Lesson 11: Continuous Integration

学习目标:掌握 CI 流水线配置与"主干开发"文化的英文词汇。

📖 英文阅读(Reading)

原文:

Continuous integration (CI) is the practice of merging code into a shared mainline frequently — often many times a day — so the team always has a tested, working version. Every push triggers a build job on a build server, which runs a series of stages: install dependencies, compile, run unit tests, run static analysis, and package the result. Each stage is a step in the pipeline, and a failure in any step stops the flow.

The CI server shows the status of every commit with a green (passing) or red (failing) indicator — hence the phrase "keeping the build green". A broken build is treated as the team's top priority: someone fixes it immediately, or the change is reverted to keep the mainline healthy. Caching speeds up builds by reusing downloaded dependencies, and artifacts from each job are stored for later use. In trunk-based development, branches are short-lived and everyone integrates constantly, which minimizes merge conflicts and keeps integration pain small. Good CI turns integration from a painful event into an invisible, everyday habit.

中文翻译:

持续集成(CI)是频繁(常常每天多次)把代码合并进共享主干(mainline)的实践,这样团队始终拥有经过测试、可以工作的版本。每次推送都会触发(trigger)构建服务器上的构建任务(build job),运行一系列阶段(stage):安装依赖、编译、运行单元测试、运行静态分析、打包结果。每个阶段是流水线中的步骤(step),任何一步失败都会停止流程。

CI 服务器用绿色(通过)或红色(失败)指示灯显示每次提交的状态(status)——所以有"保持构建绿色"这个说法。构建失败(broken build)被当作团队最高优先级:有人立即修复,或者改动被回退(revert)以保持主干健康。缓存(caching)通过复用已下载的依赖加速构建,每个任务产生的产物(artifact)被保存供后续使用。在主干开发中,分支生命周期很短,所有人持续集成,从而把合并冲突和集成痛苦降到最低。好的 CI 把集成从痛苦的事件变成无形、日常的习惯。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
mainline /ˈmeɪnlaɪn/ n. 主干(主分支) Keep the mainline green at all times. 让主干始终保持绿色(可发布)。 ★★
trigger /ˈtrɪɡə(r)/ v./n. 触发 A push triggers the CI pipeline. 一次推送触发 CI 流水线。 ★★★
job /dʒɒb/ n. (CI)任务 The build job failed at the test stage. 构建任务在测试阶段失败了。 ★★★
stage /steɪdʒ/ n. 阶段 The pipeline has five stages. 流水线有五个阶段。 ★★★
step /step/ n. 步骤 Each step logs its own output. 每个步骤记录自己的输出。 ★★★
revert /rɪˈvɜːt/ v. 回退(代码) We reverted the commit that broke the build. 我们回退了破坏构建的提交。 ★★★
cache /kæʃ/ v./n. 缓存 Cache the dependencies to speed up builds. 缓存依赖以加速构建。 ★★
status /ˈsteɪtəs/ n. 状态 Check the build status before merging. 合并前检查构建状态。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
broken build /ˈbrəʊkən bɪld/ 构建失败/损坏的构建 A broken build blocks all merges. 构建失败会阻塞所有合并。 ★★★
build server /bɪld ˈsɜːvə(r)/ 构建服务器 The build server runs 200 jobs a day. 构建服务器每天运行 200 个任务。 ★★
keep the build green /kiːp ðə bɪld ɡriːn/ 保持构建通过 Everyone is responsible for keeping the build green. 每个人都有责任保持构建通过。 ★★
static analysis /ˈstætɪk əˈnæləsɪs/ 静态分析 Static analysis finds bugs before tests run. 静态分析在测试运行前发现 bug。 ★★

第 12 课 持续交付与部署(CD)

Lesson 12: Continuous Delivery & Deployment

学习目标:掌握衡量交付速度的指标词汇(部署频率、前置时间、MTTR)与发布技巧。

📖 英文阅读(Reading)

原文:

Continuous delivery extends CI: every change that passes the pipeline is ready to ship, but a human decides when to press the button. Continuous deployment removes even that decision — every green build goes to production automatically. Both practices rely on high deployment frequency and low-risk releases.

Teams measure their delivery with a few key metrics: lead time (the time from a code change to it running in production), deployment frequency, change failure rate (how often a release breaks something), and mean time to recovery (MTTR) (how fast the team recovers from a failure). Improving these four is the core of the DORA framework.

To release safely, teams use feature flags — switches that hide unfinished features in production until they are ready — and release trains that ship on a fixed schedule. A release candidate is a build that has passed all checks and is waiting for final approval. The goal is zero-downtime deployment: users never notice that a new version went live. If something does go wrong, the team can cut losses fast, disable the feature flag, or roll back to the previous version.

中文翻译:

持续交付(continuous delivery)扩展了 CI:每个通过流水线的变更都随时可以发布,但由人来决定何时按下按钮。持续部署(continuous deployment)连这个决定也去掉——每个绿色构建都自动进入生产环境。两种实践都依赖高部署频率和低风险发布。

团队用几个关键指标衡量交付:前置时间(lead time)(从代码变更到它在生产环境运行的时间)、部署频率(deployment frequency)、变更失败率(change failure rate)(发布破坏东西的频率)和平均恢复时间(MTTR)(团队从故障中恢复的速度)。改进这四项是 DORA 框架的核心。

为了安全发布,团队使用功能开关(feature flag)——在生产环境中隐藏未完成功能的开关,等就绪后再打开——以及按固定节奏发布的发布列车(release train)。候选版本(release candidate)是已通过所有检查、等待最终批准的构建。目标是零停机部署(zero-downtime deployment):用户完全察觉不到新版本已上线。如果真出了事,团队能迅速止损、关闭功能开关,或回滚到上一个版本。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
lead time /liːd taɪm/ n. 前置时间 Our lead time is down to two hours. 我们的前置时间已缩短到两小时。 ★★
frequency /ˈfriːkwənsi/ n. 频率 Increase deployment frequency gradually. 逐步提高部署频率。 ★★★
failure rate /ˈfeɪljə(r) reɪt/ n. 失败率 Change failure rate must stay under 15%. 变更失败率必须保持在 15% 以下。 ★★
recovery /rɪˈkʌvəri/ n. 恢复 Fast recovery matters more than avoiding failure. 快速恢复比避免失败更重要。 ★★★
feature flag /ˈfiːtʃə(r) flæɡ/ n. 功能开关 Roll out the feature behind a feature flag. 在功能开关后面灰度发布该功能。 ★★★
candidate /ˈkændɪdət/ n. 候选(版本) Promote the release candidate to production. 把候选版本提升到生产环境。 ★★
disable /dɪsˈeɪbl/ v. 关闭、禁用 Disable the flag if the error rate spikes. 如果错误率飙升就关闭开关。 ★★★
rollout /ˈrəʊlaʊt/ n. 发布、推广(渐进) The rollout reached 50% of users today. 今天发布已覆盖 50% 的用户。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
continuous deployment /kənˈtɪnjuəs dɪˈplɔɪmənt/ 持续部署 Continuous deployment ships every green build. 持续部署发布每一个通过的构建。 ★★★
release candidate /rɪˈliːs kænˈdɪdət/ 候选版本 The release candidate passed all checks. 候选版本通过了所有检查。 ★★
zero-downtime deployment /ˈzɪərəʊ ˈdaʊntaɪm dɪˈplɔɪmənt/ 零停机部署 Zero-downtime deployment keeps users happy. 零停机部署让用户无感。 ★★
press the button /pres ðə ˈbʌtn/ 按下发布按钮 A human presses the button for the release. 由人来按下发布按钮。 ★★

第 13 课 构建与产物管理

Lesson 13: Build & Artifact Management

学习目标:掌握"从源代码到可运行产物"环节的英文词汇。

📖 英文阅读(Reading)

原文:

A build transforms source code into something runnable: for compiled languages, it runs the compiler to produce binary executables; for interpreted languages, it bundles, minifies, and prepares the code for deployment. The build package gathers everything the app needs into one unit, such as a JAR, a wheel, or a zip file. Every build produces a build artifact — a versioned, immutable output that can be deployed or rolled back to.

Artifacts are stored in a registry (or artifact repository), where every version is tracked and checksums verify that files have not been tampered with. Versioning matters: each artifact carries a unique version number, so you always know exactly what is running where. Good builds are reproducible — the same source and the same environment produce the same output, which makes debugging old versions possible.

A build failure blocks the pipeline and should be fixed immediately. Teams publish release notes with every version, summarizing new features, fixes, and breaking changes (changes that require users to modify their code). Together, build and artifact management give the team traceability: any running system can be traced back to the exact source that produced it.

中文翻译:

构建(build)把源代码(source code)转换成可运行的东西:对编译型语言,它运行编译器(compiler)生成二进制(binary)可执行文件;对解释型语言,它做打包、压缩(minify)并为部署准备代码。构建会打包(package)把应用所需的一切收集成一个单元,比如 JAR、wheel 或 zip 文件。每次构建产生一个构建产物(build artifact)——一个带版本号、不可变的输出,可被部署,也可回滚到它。

产物存放在制品仓库(registry / artifact repository)中,每个版本都被跟踪,校验和(checksum)确保文件没有被篡改。版本管理(versioning)很重要:每个产物携带唯一版本号,让你始终确切知道什么版本运行在哪里。好的构建是可复现的(reproducible)——相同源码和相同环境产生相同输出,这让调试旧版本成为可能。

构建失败(build failure)会阻塞流水线,必须立即修复。团队每次发版都发布发布说明(release notes),总结新功能、修复和破坏性变更(breaking change)(要求用户修改自己代码的变更)。构建与产物管理一起给团队带来可追溯性(traceability):任何运行中的系统都能追溯到产生它的确切源码。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
build /bɪld/ n./v. 构建 The build takes four minutes. 构建需要四分钟。 ★★★
source code /sɔːs kəʊd/ n. 源代码 The source code lives in the repository. 源代码存放在仓库中。 ★★★
compiler /kəmˈpaɪlə(r)/ n. 编译器 The compiler reports type errors. 编译器报告类型错误。 ★★
binary /ˈbaɪnəri/ n./adj. 二进制(文件) Ship the binary with its checksum. 发布二进制文件及其校验和。 ★★
package /ˈpækɪdʒ/ n./v. 软件包、打包 Package the app into a single file. 把应用打包成单个文件。 ★★★
registry /ˈredʒɪstri/ n. 制品库/仓库 Push the image to the registry. 把镜像推送到制品库。 ★★
checksum /ˈtʃeksʌm/ n. 校验和 Verify the checksum before installing. 安装前校验校验和。 ★
reproducible /ˌriːprəˈdjuːsəbl/ adj. 可复现的 Builds must be reproducible. 构建必须可复现。 ★★
versioning /ˈvɜːʃnɪŋ/ n. 版本管理 Semantic versioning uses major.minor.patch. 语义化版本管理用 主.次.补丁 三段式。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
build artifact /bɪld ˈɑːtɪfækt/ 构建产物 Download the build artifact from the registry. 从制品库下载构建产物。 ★★★
release notes /rɪˈliːs nəʊts/ 发布说明 Read the release notes before upgrading. 升级前阅读发布说明。 ★★
breaking change /ˈbreɪkɪŋ tʃeɪndʒ/ 破坏性变更 This version has no breaking changes. 这个版本没有破坏性变更。 ★★
trace back /treɪs bæk/ 追溯到 Every build can be traced back to source. 每个构建都能追溯到源码。 ★★

第 14 课 环境与配置管理

Lesson 14: Environment & Configuration

学习目标:掌握环境隔离、配置、密钥管理的英文词汇。

📖 英文阅读(Reading)

原文:

Modern teams run the same software in several environments: development, testing, staging, and production. The environments must be as similar as possible, so code that works locally also works in production. Configuration — settings that change between environments — is kept separate from code: API keys, database addresses, and URLs live in config files or environment variables, not in the source.

Sensitive values such as passwords and tokens are called secrets. Secrets must never be committed to a repository; they are stored in a secret manager and injected at runtime. A leaked credential is a serious security incident. Teams also use feature toggles to switch behavior without redeploying.

Environment provisioning creates servers, databases, and networks automatically. A common problem is configuration drift: servers slowly become different from each other as people change them manually. The fix is immutable infrastructure — never modify a running server; instead, replace it with a fresh one built from a baseline template. Finally, every change to configuration should follow change management: documented, reviewed, and reversible.

中文翻译:

现代团队在多个环境中运行同一套软件:开发、测试、预发布和生产。环境之间必须尽可能相似,这样本地能跑的代码在生产中也能跑。配置(configuration)——随环境变化的设置——与代码分离:API 密钥、数据库地址和 URL 存放在配置文件(config file)或环境变量(environment variable)中,而不是写进源码。

密码、令牌等敏感值叫密钥/机密(secret)。密钥绝不能提交进仓库,应存放在密钥管理器(secret manager)中,运行时注入。凭据(credential)泄露是严重的安全事件。团队还用功能开关(feature toggle)在无需重新部署的情况下切换行为。

环境配置供应(provisioning)自动创建服务器、数据库和网络。一个常见问题是配置漂移(configuration drift):人们手动修改服务器,导致服务器之间逐渐变得不同。解决办法是不可变基础设施(immutable infrastructure)——从不修改运行中的服务器,而是用基线(baseline)模板构建的全新服务器替换它。最后,对配置的每一次变更都应遵循变更管理(change management):有文档、经评审、可回退。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
configuration /kənˌfɪɡəˈreɪʃn/ n. 配置 The configuration differs between environments. 配置在不同环境间不同。 ★★★
config file /kənˈfɪɡ faɪl/ n. 配置文件 Store timeouts in a config file. 把超时时间放在配置文件里。 ★★★
environment variable /ɪnˈvaɪrənmənt ˈveəriəbl/ n. 环境变量 Read the database URL from an environment variable. 从环境变量读取数据库地址。 ★★★
secret /ˈsiːkrət/ n. 密钥、机密 Never log secrets to the console. 绝不把密钥打印到控制台。 ★★★
credential /krəˈdenʃl/ n. 凭据(账号密码等) Rotate credentials every 90 days. 每 90 天轮换一次凭据。 ★★
provisioning /prəˈvɪʒənɪŋ/ n. 资源供应、配置环境 Provisioning a new environment takes ten minutes. 配置一个新环境需要十分钟。 ★★
drift /drɪft/ n. 漂移(配置不一致) Drift causes "works on my machine" problems. 配置漂移导致"在我机器上能跑"的问题。 ★★
immutable /ɪˈmjuːtəbl/ adj. 不可变的 Immutable infrastructure prevents drift. 不可变基础设施防止漂移。 ★★
baseline /ˈbeɪslaɪn/ n. 基线 Rebuild servers from the baseline template. 从基线模板重建服务器。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
secret manager /ˈsiːkrət ˈmænɪdʒə(r)/ 密钥管理器 The secret manager injects tokens at runtime. 密钥管理器在运行时注入令牌。 ★★
feature toggle /ˈfiːtʃə(r) ˈtɒɡl/ 功能开关 A feature toggle hides the new UI. 功能开关隐藏新界面。 ★★
configuration drift /kənˌfɪɡəˈreɪʃn drɪft/ 配置漂移 Drift detection keeps environments aligned. 漂移检测让环境保持一致。 ★★
change management /tʃeɪndʒ ˈmænɪdʒmənt/ 变更管理 All config changes go through change management. 所有配置变更都经过变更管理。 ★★

第 15 课 容器化与 Docker

Lesson 15: Containerization & Docker

学习目标:掌握容器概念与 Docker 操作的英文词汇。

📖 英文阅读(Reading)

原文:

A container packages an application with its runtime, libraries, and settings into a single, portable unit. Unlike a virtual machine, a container shares the host's kernel, so it is lightweight and starts in seconds. Containerization solves the classic "works on my machine" problem: the container runs identically on a developer's laptop, a test server, or a cloud instance.

Docker is the most popular container platform. Developers write a Dockerfile that describes how to build an image — a read-only template made of layers, where each instruction adds one layer. Images are stored in a container registry, such as Docker Hub, and pulled to run anywhere. Running an image creates a container instance, with its own runtime, filesystem, and network. Data that must survive container restarts is stored in a volume — a storage area outside the container's ephemeral filesystem.

Teams expose application ports so the outside world can reach the service, and manage container lifecycle with commands like start, stop, and restart. Isolation keeps containers from interfering with each other. Because images are versioned and immutable, rollback is trivial: just start the previous image.

中文翻译:

容器(container)把应用连同其运行时、库和设置打包成一个单一、可移植的单元。与虚拟机不同,容器共享宿主机的内核(kernel),因此轻量(lightweight),几秒内就能启动。容器化(containerization)解决了经典的"在我机器上能跑"问题:容器在开发者的笔记本、测试服务器或云实例上运行结果完全一致。

Docker 是最流行的容器平台。开发者编写Dockerfile,描述如何构建镜像(image)——一个由层(layer)组成的只读模板,每条指令添加一层。镜像存放在容器镜像仓库(container registry)(如 Docker Hub)中,可被拉取到任何地方运行。运行镜像会创建容器实例(container instance),拥有自己的运行时(runtime)、文件系统和网络。需要在容器重启后保留的数据存放在卷(volume)中——一个位于容器临时文件系统之外的存储区域。

团队暴露(expose)应用端口(port)让外界能访问服务,并用 start、stop、restart 等命令管理容器生命周期。隔离(isolation)让容器互不干扰。由于镜像是带版本且不可变的,回滚轻而易举:只需启动之前的镜像。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
image /ˈɪmɪdʒ/ n. (容器)镜像 Pull the latest image from the registry. 从镜像仓库拉取最新镜像。 ★★★
Dockerfile /ˈdɒkəfaɪl/ n. 容器构建文件 The Dockerfile installs Python dependencies. Dockerfile 安装 Python 依赖。 ★★★
layer /ˈleɪə(r)/ n. 镜像层 Rebuild only the layers that changed. 只重建变更过的层。 ★★
lightweight /ˈlaɪtweɪt/ adj. 轻量的 Containers are more lightweight than VMs. 容器比虚拟机更轻量。 ★★
runtime /ˈrʌntaɪm/ n. 运行时 The runtime version is pinned in the image. 运行时版本固定在镜像中。 ★★★
volume /ˈvɒljuːm/ n. 卷(持久化存储) Mount a volume for the database files. 为数据库文件挂载一个卷。 ★★
port /pɔːt/ n. 端口 Expose port 8080 to the host. 把 8080 端口暴露给宿主机。 ★★★
isolation /ˌaɪsəˈleɪʃn/ n. 隔离 Isolation prevents containers from interfering. 隔离防止容器互相干扰。 ★★
kernel /ˈkɜːnl/ n. 内核 Containers share the host kernel. 容器共享宿主内核。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
container registry /kənˈteɪnə(r) ˈredʒɪstri/ 容器镜像仓库 Push the image to the container registry. 把镜像推送到容器镜像仓库。 ★★★
run a container /rʌn ə kənˈteɪnə(r)/ 运行容器 Run a container from the image. 从镜像运行一个容器。 ★★★
start / stop / restart /stɑːt stɒp ˌriːˈstɑːt/ 启动/停止/重启 Restart the container after changing the config. 修改配置后重启容器。 ★★★
expose a port /ɪkˈspəʊz ə pɔːt/ 暴露端口 Expose port 3000 for the web app. 为网页应用暴露 3000 端口。 ★★

第 16 课 容器编排与 Kubernetes

Lesson 16: Orchestration & Kubernetes

学习目标:掌握 Kubernetes 核心概念(集群、Pod、服务、伸缩)的英文词汇。

📖 英文阅读(Reading)

原文:

Running hundreds of containers by hand is impossible. Orchestration tools manage containers at scale: they deploy, scale, network, and heal them automatically. Kubernetes (K8s) is the industry-standard orchestrator.

A Kubernetes cluster is a set of machines called nodes. Workloads run in pods — the smallest deployable units, each containing one or more containers. Pods are scheduled by the scheduler onto nodes with enough resources. A service provides a stable network endpoint and performs load balancing, distributing traffic among pod replicas so no single pod is overwhelmed. The controller continuously compares the desired state with reality: if a pod crashes, the replica set immediately creates a new one — this self-healing behavior is one of Kubernetes' biggest wins.

Teams use namespaces to divide the cluster between teams or environments, and an ingress to route external traffic to services. Autoscaling adjusts the number of replicas based on CPU usage or request volume, and rolling updates replace pods gradually with zero downtime. A health check ensures that only healthy pods receive traffic. Kubernetes turns infrastructure into a declarative system: you describe the desired state, and the cluster makes it real.

中文翻译:

手工管理几百个容器是不可能的。编排(orchestration)工具在规模上管理容器:自动部署、伸缩、组网和自愈。Kubernetes(K8s)是行业标准的编排器。

Kubernetes 集群(cluster)由一组叫节点(node)的机器组成。工作负载运行在 Pod 中——最小的可部署单元,每个 Pod 包含一个或多个容器。调度器(scheduler)把 Pod 调度到资源充足的节点上。服务(service)提供稳定的网络端点并执行负载均衡(load balancing),在 Pod 副本之间分发流量,避免单个 Pod 过载。控制器(controller)持续比较期望状态与现实:如果某个 Pod 崩溃,副本集(replica set)立即创建新 Pod——这种自愈(self-healing)行为是 Kubernetes 最大的优势之一。

团队用命名空间(namespace)在团队或环境之间划分集群,用入口(ingress)把外部流量路由到服务。自动伸缩(autoscaling)根据 CPU 使用率或请求量调整副本数量,滚动更新(rolling update)零停机地逐步替换 Pod。健康检查(health check)确保只有健康的 Pod 接收流量。Kubernetes 把基础设施变成声明式系统:你描述期望状态,集群让它成为现实。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
orchestration /ˌɔːkɪˈstreɪʃn/ n. 编排 Orchestration manages containers at scale. 编排在大规模上管理容器。 ★★★
cluster /ˈklʌstə(r)/ n. 集群 The cluster has 20 nodes. 集群有 20 个节点。 ★★★
node /nəʊd/ n. 节点(机器) Add another node before the holiday traffic. 节假日前再加一个节点。 ★★★
pod /pɒd/ n. Pod(容器组) Each pod runs one application instance. 每个 Pod 运行一个应用实例。 ★★★
scheduler /ˈʃedjuːlə(r)/ n. 调度器 The scheduler places pods on free nodes. 调度器把 Pod 放到空闲节点上。 ★★
replica /ˈreplɪkə/ n. 副本 We run three replicas of the API. 我们运行 API 的三个副本。 ★★
namespace /ˈneɪmspeɪs/ n. 命名空间 Each team owns a namespace. 每个团队拥有一个命名空间。 ★★
autoscaling /ˌɔːtəʊˈskeɪlɪŋ/ n. 自动伸缩 Autoscaling handles traffic spikes. 自动伸缩应对流量高峰。 ★★
ingress /ˈɪnɡres/ n. 入口(流量路由) The ingress routes traffic to services. 入口把流量路由到服务。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
load balancing /ləʊd ˈbælənsɪŋ/ 负载均衡 Load balancing spreads requests across pods. 负载均衡把请求分摊到多个 Pod。 ★★★
health check /helθ tʃek/ 健康检查 The health check failed on pod 3. 3 号 Pod 的健康检查失败了。 ★★★
rolling update /ˈrəʊlɪŋ ˈʌpdeɪt/ 滚动更新 Rolling updates avoid downtime. 滚动更新避免停机。 ★★
self-healing /ˌself ˈhiːlɪŋ/ 自愈 Self-healing restarts crashed pods. 自愈机制重启崩溃的 Pod。 ★★

第 17 课 微服务架构

Lesson 17: Microservices Architecture

学习目标:掌握微服务、分布式系统设计的关键英文词汇。

📖 英文阅读(Reading)

原文:

A monolith is an application built as one big unit. It is simple to start but hard to scale and deploy — one small change means rebuilding everything. Microservices split the system into small, independent services, each owning one business capability, such as billing, search, or notifications. Each service can be developed, deployed, and scaled independently by its own team.

Microservices communicate over the network using APIs, so decoupling between services is essential: teams aim for loose coupling (services know little about each other) and high cohesion (each service does one job well). Because services can fail, the system must be designed for fault tolerance: a circuit breaker stops calling a failing service and returns a fallback response instead of hanging. Service discovery lets services find each other's addresses automatically without hard-coded URLs.

An API gateway sits at the edge, routing requests, handling authentication, and aggregating responses. Distributed systems bring new challenges — network latency, partial failures, and data consistency — so teams invest heavily in monitoring and testing. The trade-off is real: microservices offer scalability and team autonomy, but they demand much more operational maturity than a monolith.

中文翻译:

单体应用(monolith)是作为一个大单元构建的应用。它起步简单,但难以扩展和部署——一次小改动意味着要重新构建一切。微服务(microservices)把系统拆成小而独立的服务,每个服务负责一项业务能力,比如计费、搜索或通知。每个服务可以由自己的团队独立开发、部署和扩展。

微服务通过网络用 API 通信,因此服务之间的解耦至关重要:团队追求松耦合(loose coupling)(服务之间相互了解很少)和高内聚(high cohesion)(每个服务把一件事做好)。由于服务可能失败,系统必须为容错(fault tolerance)而设计:熔断器(circuit breaker)停止调用失败的服务并返回降级响应,而不是一直挂起。服务发现(service discovery)让服务自动找到彼此的地址,无需硬编码 URL。

API 网关(API gateway)位于边缘,负责路由请求、处理认证、聚合响应。分布式系统(distributed system)带来新挑战——网络延迟、部分失败、数据一致性——因此团队在监控和测试上投入巨大。取舍是真实的:微服务带来可扩展性和团队自治,但对运维成熟度的要求远高于单体。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
monolith /ˈmɒnəlɪθ/ n. 单体应用 The monolith is hard to scale. 单体应用难以扩展。 ★★
microservice /ˈmaɪkrəʊˌsɜːvɪs/ n. 微服务 Each microservice owns one capability. 每个微服务负责一项能力。 ★★★
loose coupling /luːs ˈkʌplɪŋ/ n. 松耦合 Loose coupling lets teams move fast. 松耦合让团队快速前进。 ★★
cohesion /kəʊˈhiːʒn/ n. 内聚 High cohesion keeps services focused. 高内聚让服务保持聚焦。 ★★
circuit breaker /ˈsɜːkɪt ˈbreɪkə(r)/ n. 熔断器 The circuit breaker tripped after five failures. 熔断器在五次失败后跳闸。 ★★
gateway /ˈɡeɪtweɪ/ n. 网关 The gateway handles authentication. 网关处理认证。 ★★★
discovery /dɪˈskʌvəri/ n. 服务发现 Discovery finds healthy service instances. 服务发现找到健康的服务实例。 ★★
distributed /dɪˈstrɪbjuːtɪd/ adj. 分布式的 Distributed systems fail in partial ways. 分布式系统以部分方式失败。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
API gateway /ˌeɪ piː ˈaɪ ˈɡeɪtweɪ/ API 网关 All traffic passes through the API gateway. 所有流量都经过 API 网关。 ★★★
service discovery /ˈsɜːvɪs dɪˈskʌvəri/ 服务发现 Service discovery replaces hard-coded URLs. 服务发现取代硬编码地址。 ★★
fallback response /ˈfɔːlbæk rɪˈspɒns/ 降级响应 Return a fallback response when the service is down. 服务宕机时返回降级响应。 ★★
split into /splɪt ˈɪntə/ 拆分成 We split the monolith into ten services. 我们把单体拆成十个服务。 ★★★

第 18 课 部署策略

Lesson 18: Deployment Strategies

学习目标:掌握蓝绿、金丝雀、滚动等部署策略的英文词汇。

📖 英文阅读(Reading)

原文:

How you deploy matters as much as what you deploy. A simple in-place deployment stops the old version, installs the new one, and restarts — fast, but it causes downtime, and if the new version is broken, you are stuck. Modern teams prefer strategies that keep the service available and make rollback easy.

Blue-green deployment runs two identical environments: blue (the current version) and green (the new version). You deploy to green, run tests against it, then switch traffic over with a router change. If problems appear, you switch back instantly. Canary release is more gradual: route a small percentage of traffic — say 5% — to the new version, watch the error rate, and slowly increase the percentage until 100%. If the canary misbehaves, cut it off before most users are affected.

Rolling deployment replaces instances one by one, keeping the rest serving traffic. Feature flags give even finer control: you can ship code that is dark (inactive) and flip it on later. Whatever the strategy, a clear rollback plan is mandatory, and automated verification after each step tells the team whether to continue or abort the release.

中文翻译:

怎么部署和部署什么同样重要。简单的原地部署(in-place deployment)停掉旧版本、安装新版本、再重启——速度快,但会造成停机,而且如果新版本坏了,你就卡住了。现代团队更偏好能保持服务可用、且回滚容易的策略。

蓝绿部署(blue-green deployment)运行两个相同的环境:蓝色(当前版本)和绿色(新版本)。你部署到绿色,对它跑测试,然后通过路由切换把流量(traffic)切过去。如果出问题,瞬间切回。金丝雀发布(canary release)更渐进:把一小部分流量(比如 5%)路由到新版本,观察错误率,再慢慢提升百分比直到 100%。如果金丝雀表现异常,在影响大多数用户之前就切断它。

滚动部署(rolling deployment)逐个替换实例,其余实例继续服务流量。功能开关提供更细的控制:你可以发布未激活(dark)的代码,之后再打开。无论哪种策略,清晰的回滚预案(rollback plan)都是必需的,每一步之后的自动化验证告诉团队是继续还是中止(abort)发布。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
blue-green /ˌbluː ˈɡriːn/ adj. 蓝绿(双环境部署) Blue-green deployment enables instant rollback. 蓝绿部署支持瞬间回滚。 ★★
canary /kəˈneəri/ n. 金丝雀(灰度发布) The canary received 5% of traffic. 金丝雀接收了 5% 的流量。 ★★
traffic /ˈtræfɪk/ n. 流量 Shift traffic gradually to the new version. 逐步把流量切到新版本。 ★★★
switch /swɪtʃ/ v./n. 切换 Switch the router to the green environment. 把路由器切到绿色环境。 ★★★
error rate /ˈerə(r) reɪt/ n. 错误率 The error rate spiked after the canary. 金丝雀发布后错误率飙升。 ★★★
rolling /ˈrəʊlɪŋ/ adj. 滚动的、分批的 Rolling deployment updates one pod at a time. 滚动部署一次更新一个 Pod。 ★★
dark /dɑːk/ adj. 未激活的(暗中发布) Ship the feature dark, then enable it later. 先暗中发布功能,之后再启用。 ★★
abort /əˈbɔːt/ v. 中止 Abort the release if the check fails. 如果检查失败就中止发布。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
blue-green deployment /ˌbluː ˈɡriːn dɪˈplɔɪmənt/ 蓝绿部署 Blue-green deployment halves release risk. 蓝绿部署把发布风险减半。 ★★★
canary release /kəˈneəri rɪˈliːs/ 金丝雀发布/灰度发布 We did a canary release to test the new API. 我们用金丝雀发布测试新接口。 ★★★
rollback plan /ˈrəʊlbæk plæn/ 回滚预案 Every release needs a rollback plan. 每次发布都需要回滚预案。 ★★★
in-place deployment /ˌɪn ˈpleɪs dɪˈplɔɪmənt/ 原地部署 In-place deployment causes downtime. 原地部署会造成停机。 ★★

第 19 课 基础设施即代码(IaC)

Lesson 19: Infrastructure as Code

学习目标:掌握用代码管理基础设施的英文词汇(声明式、幂等、状态)。

📖 英文阅读(Reading)

原文:

Infrastructure as code (IaC) means managing servers, networks, and databases with code and version control, instead of clicking through consoles. Tools like Terraform and Ansible treat infrastructure as declarative configuration: you write the desired state — "I want three web servers and one database" — and the tool makes it real. This brings the benefits of software engineering to infrastructure: code review, testing, and versioning of everything.

A key property of good IaC is idempotency: running the same script twice produces the same result, with no harmful side effects. The tool keeps track of state — a record of what has been created — so it knows exactly what to add, change, or remove. Provisioning creates new resources; apply pushes your configuration; destroy tears everything down. Teams store their infrastructure definitions in repositories and review changes like code, so every server is reproducible and nothing is configured by hand.

IaC also enables immutable infrastructure and fast environment provisioning: spin up an identical staging environment in minutes, test a change, and tear it down. The result is fewer mistakes, faster delivery, and environments that never drift from the definition.

中文翻译:

基础设施即代码(IaC)意味着用代码和版本控制来管理服务器、网络和数据库,而不是在控制台里点点点。Terraform、Ansible 等工具把基础设施当作声明式(declarative)配置:你写出期望状态(desired state)——"我要三台 Web 服务器和一个数据库"——工具让它成为现实。这把软件工程的好处带给基础设施:一切都可以代码评审、测试和版本管理。

好的 IaC 的一个关键属性是幂等性(idempotency):同样的脚本跑两次产生同样的结果,没有有害副作用。工具会跟踪状态(state)——一份关于已创建内容的记录——因此它确切知道该添加、修改或删除什么。配置供应(provisioning)创建新资源;apply(应用)推送你的配置;destroy(销毁)拆除一切。团队把基础设施定义存进仓库,像代码一样评审变更,因此每台服务器都可复现,没有任何东西是手工配置的。

IaC 还支持不可变基础设施和快速环境配置:几分钟内启动一个一模一样的预发布环境、测试变更、然后拆除。结果是更少的错误、更快的交付,以及永远不会偏离定义的环境。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
infrastructure /ˈɪnfrəstrʌktʃə(r)/ n. 基础设施 The infrastructure is managed as code. 基础设施以代码方式管理。 ★★★
declarative /dɪˈklærətɪv/ adj. 声明式的 Declarative config describes the desired state. 声明式配置描述期望状态。 ★★
desired state /dɪˈzaɪəd steɪt/ n. 期望状态 The tool converges to the desired state. 工具收敛到期望状态。 ★★
idempotent /ˌaɪdemˈpəʊtənt/ adj. 幂等的 Terraform applies are idempotent. Terraform 的 apply 是幂等的。 ★★
state /steɪt/ n. 状态(记录) The state file tracks created resources. 状态文件记录已创建的资源。 ★★★
apply /əˈplaɪ/ v. 应用(配置) Run apply to create the resources. 运行 apply 创建资源。 ★★★
destroy /dɪˈstrɔɪ/ v. 销毁 Destroy the test environment after the demo. 演示结束后销毁测试环境。 ★★
provision /prəˈvɪʒn/ v. 配置(资源) Provision a database for the new service. 为新服务配置一个数据库。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
infrastructure as code /ˈɪnfrəstrʌktʃə(r) æz kəʊd/ 基础设施即代码 Infrastructure as code ends manual setup. 基础设施即代码终结手工搭建。 ★★★
spin up /spɪn ʌp/ 快速启动(资源) Spin up a staging environment in minutes. 几分钟内启动预发布环境。 ★★
tear down /teə(r) daʊn/ 拆除 Tear down unused environments to save cost. 拆除不用的环境以节省成本。 ★★
versioned infrastructure /ˈvɜːʃnd ˈɪnfrəstrʌktʃə(r)/ 版本化管理的基础设施 Versioned infrastructure makes change auditable. 版本化管理的基础设施让变更可审计。 ★★

第 20 课 监控与指标

Lesson 20: Monitoring & Metrics

学习目标:掌握监控体系(指标、仪表盘、阈值、黄金信号)的英文词汇。

📖 英文阅读(Reading)

原文:

Monitoring answers one question: is the system healthy? Teams collect metrics — numeric measurements like CPU usage, memory, request counts, and error rates — and display them on a dashboard. A good dashboard shows the golden signals of service health: latency (how fast requests complete), traffic (how much demand there is), errors (how often requests fail), and saturation (how close to capacity the system is).

Metrics are only useful with a baseline for comparison and thresholds for action. When a metric crosses a threshold — say, latency above 500 ms for five minutes — the system raises an alert. Dashboards help humans see trends over time and spot anomalies that automated rules miss. Teams often look at percentiles (p95, p99) rather than averages, because averages hide the slowest users' pain.

Observability goes beyond predefined metrics: it lets you ask new questions about any request by combining metrics, logs, and traces. Visualization tools turn raw numbers into charts that make problems obvious. Good monitoring is proactive: you detect a slow degradation days before users complain, not after an outage.

中文翻译:

监控回答一个问题:系统健康吗?团队收集指标(metric)——CPU 使用率、内存、请求数、错误率等数值——并显示在仪表盘(dashboard)上。好的仪表盘展示服务健康的黄金信号(golden signals):延迟(latency)(请求完成有多快)、流量(traffic)(需求有多大)、错误(error)(请求失败的频率)和饱和度(saturation)(系统离容量上限有多近)。

指标只有配上比较用的基线(baseline)和行动用的阈值(threshold)才有用。当指标越过阈值——比如延迟连续五分钟超过 500 毫秒——系统就发出告警(alert)。仪表盘帮助人类观察随时间变化的趋势(trend),发现自动化规则漏掉的异常(anomaly)。团队常看百分位(percentile)(p95、p99)而不是平均值,因为平均值掩盖了最慢用户的痛苦。

可观测性(observability)超越预定义指标:它让你通过结合指标、日志和链路追踪,对任意请求提出新问题。可视化(visualization)工具把原始数字变成让问题一目了然的图表。好的监控是主动的:你在用户抱怨前几天就发现缓慢的劣化,而不是在宕机之后。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
dashboard /ˈdæʃbɔːd/ n. 仪表盘 The dashboard shows all key metrics. 仪表盘显示所有关键指标。 ★★★
threshold /ˈθreʃhəʊld/ n. 阈值 Set the alert threshold at 90% CPU. 把告警阈值设为 CPU 90%。 ★★★
trend /trend/ n. 趋势 The chart shows an upward error trend. 图表显示错误率上升趋势。 ★★★
anomaly /əˈnɒməli/ n. 异常 The anomaly detector found unusual traffic. 异常检测器发现不寻常的流量。 ★★
percentile /pəˈsentaɪl/ n. 百分位 P99 latency is under 300 ms. P99 延迟低于 300 毫秒。 ★★
saturation /ˌsætʃəˈreɪʃn/ n. 饱和度 Saturation near 100% means the system is full. 饱和度接近 100% 意味着系统满载。 ★★
observability /əbˌzɜːvəˈbɪləti/ n. 可观测性 Observability answers unexpected questions. 可观测性回答意料之外的问题。 ★★★
visualization /ˌvɪʒuəlaɪˈzeɪʃn/ n. 可视化 Good visualization makes problems obvious. 好的可视化让问题一目了然。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
golden signals /ˈɡəʊldən ˈsɪɡnlz/ 黄金信号 Latency and errors are golden signals. 延迟和错误是黄金信号。 ★★
real-time monitoring /ˌrɪəl ˈtaɪm ˈmɒnɪtərɪŋ/ 实时监控 Real-time monitoring catches issues early. 实时监控及早发现问题。 ★★★
system health /ˈsɪstəm helθ/ 系统健康 The health dashboard is green. 健康仪表盘是绿色的。 ★★
keep an eye on /kiːp ən aɪ ɒn/ 密切关注 Keep an eye on memory usage after deploy. 部署后密切关注内存使用。 ★★

第 21 课 日志与可观测性

Lesson 21: Logging & Observability

学习目标:掌握日志、链路追踪、遥测等可观测性英文词汇。

📖 英文阅读(Reading)

原文:

When something goes wrong, the first question is: what happened? The answer lives in logs — time-stamped records of events written by the application. Teams practice structured logging: instead of free-form text, each log entry is a machine-readable record with fields like timestamp, level, service, and message. Log levels (debug, info, warn, error, fatal) let you filter noise. Logs are aggregated into a central system for searching, with retention policies that define how long they are kept and rotation that prevents disks from filling up.

Logs alone tell you what happened on one service. To see the whole story of a request across many services, you need distributed tracing: a trace is a request's journey, broken into spans — individual operations with their durations. A correlation ID links logs and traces for the same request. Together, logs, metrics, and traces form the three pillars of observability — the ability to understand a system's internal state from its external outputs. Telemetry is the umbrella term for all this data collected from systems. Good telemetry turns debugging from guesswork into science: you can replay any incident and see exactly what happened, step by step.

中文翻译:

出问题时,第一个问题是:发生了什么?答案在日志(log)里——应用记录的事件时间戳条目。团队实践结构化日志(structured logging):每条日志不是自由文本,而是带时间戳、级别、服务、消息等字段的机器可读记录。日志级别(log level)(debug、info、warn、error、fatal)让你过滤噪音。日志被汇总到中央系统便于搜索,用保留策略(retention policy)定义保存时长,用轮转(rotation)防止磁盘写满。

日志单独只能告诉你单个服务发生了什么。要看一个请求跨多个服务的完整故事,需要分布式追踪(distributed tracing):链路(trace)是一个请求的旅程,被拆成跨度(span)——带各自时长的独立操作。关联 ID(correlation ID)把同一请求的日志和链路关联起来。日志、指标和链路三者共同构成可观测性(observability)的三大支柱——即从系统外部输出理解其内部状态的能力。遥测(telemetry)是所有这些从系统收集的数据的总称。好的遥测把调试从猜测变成科学:你可以重放任何事件,一步步看清到底发生了什么。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
trace /treɪs/ n. 链路(追踪) The trace shows a 2-second wait at the database. 链路显示在数据库处等待了 2 秒。 ★★★
span /spæn/ n. 跨度(追踪单元) Each span records one operation's duration. 每个跨度记录一个操作的时长。 ★★
telemetry /təˈlemətri/ n. 遥测数据 Telemetry flows into the monitoring platform. 遥测数据流入监控平台。 ★★
retention /rɪˈtenʃn/ n. 保留(期限) Log retention is 30 days for compliance. 出于合规要求日志保留 30 天。 ★★
rotation /rəʊˈteɪʃn/ n. 轮转(日志) Rotation keeps log files from growing forever. 轮转防止日志文件无限增长。 ★★
aggregation /ˌæɡrɪˈɡeɪʃn/ n. 聚合 Log aggregation enables full-text search. 日志聚合支持全文搜索。 ★★
correlation /ˌkɒrəˈleɪʃn/ n. 关联 Use the correlation ID to join logs and traces. 用关联 ID 把日志和链路连接起来。 ★★
pillar /ˈpɪlə(r)/ n. 支柱 Metrics, logs, and traces are the three pillars. 指标、日志、链路是三大支柱。 ★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
structured logging /ˈstrʌktʃəd ˈlɒɡɪŋ/ 结构化日志 Structured logging makes logs searchable. 结构化日志让日志可搜索。 ★★
distributed tracing /dɪˈstrɪbjuːtɪd ˈtreɪsɪŋ/ 分布式追踪 Distributed tracing follows a request across services. 分布式追踪跟踪请求跨服务的过程。 ★★★
log level /lɒɡ ˈlevl/ 日志级别 Set the log level to debug for this module. 把这个模块的日志级别设为 debug。 ★★★
log aggregation /lɒɡ ˌæɡrɪˈɡeɪʃn/ 日志聚合 Log aggregation collects logs from all servers. 日志聚合收集所有服务器的日志。 ★★

第 22 课 告警与值班

Lesson 22: Alerting & On-call

学习目标:掌握告警流程、值班制度、升级机制的英文词汇。

📖 英文阅读(Reading)

原文:

Monitoring produces data; alerting turns data into action. An alert fires when a metric crosses a threshold, and the on-call engineer receives a page on their phone. Alerts carry a severity level: a critical alert means users are affected now; a warning means something needs attention soon. Every alert should have a runbook — a written procedure explaining what to check and how to fix the problem — so a sleepy engineer at 3 a.m. knows exactly what to do.

The on-call team works on a rotation: engineers take turns carrying the pager, usually for a week at a time. When an alert fires, the engineer must acknowledge it quickly, or the alert escalates to the next person and eventually to a manager. Too many irrelevant alerts cause alert fatigue: people stop noticing alarms, and real problems get missed. Good teams tune their alerts relentlessly, favoring a few high-signal rules over hundreds of noisy ones.

Burnout is the biggest risk in on-call work; teams protect engineers with follow-the-sun rotations, fair schedules, and a blameless culture where pages are treated as system problems, not personal failures. The goal is not zero alerts, but alerts that are actionable — each one tells you a real problem and how to respond.

中文翻译:

监控产生数据,告警(alerting)把数据变成行动。当指标越过阈值时触发一条告警,值班工程师(on-call engineer)的手机收到寻呼。告警带有严重级别(severity):critical(严重)意味着用户此刻正受影响;warning(警告)意味着有事情需要尽快关注。每条告警都应配有操作手册(runbook)——一份书面流程,说明该检查什么、如何修复——这样凌晨三点昏昏欲睡的工程师也能确切知道该做什么。

值班团队按轮换(rotation)工作:工程师轮流携带寻呼机,通常一次一周。告警触发时,工程师必须快速确认(acknowledge),否则告警会升级(escalate)到下一个人,最终到经理那里。太多无关告警会造成告警疲劳(alert fatigue):人们不再注意警报,真正的问题被漏掉。好的团队不懈地调优告警,宁要少数高信号规则,不要数百条嘈杂的规则。

倦怠(burnout)是值班工作最大的风险;团队用随日轮换(follow-the-sun)、公平排班和无责文化(blameless culture)保护工程师——寻呼被当作系统问题,而不是个人失败。目标不是零告警,而是可行动的(actionable)告警——每一条都告诉你真实的问题以及如何应对。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
severity /sɪˈverəti/ n. 严重级别 Severity one means total outage. 一级严重度意味着完全宕机。 ★★★
critical /ˈkrɪtɪkl/ adj. 严重的、关键的 The database failure is critical. 数据库故障是严重的。 ★★★
runbook /ˈrʌnbʊk/ n. 操作手册 Check the runbook for the restart procedure. 查看操作手册中的重启流程。 ★★
acknowledge /əkˈnɒlɪdʒ/ v. 确认(收到告警) Acknowledge the alert within five minutes. 五分钟内确认告警。 ★★★
escalate /ˈeskəleɪt/ v. 升级 The alert escalates if nobody responds. 无人响应时告警升级。 ★★★
fatigue /fəˈtiːɡ/ n. 疲劳 Alert fatigue causes missed incidents. 告警疲劳导致漏掉事件。 ★★
burnout /ˈbɜːnaʊt/ n. 倦怠 Good scheduling prevents on-call burnout. 好的排班防止值班倦怠。 ★★
actionable /ˈækʃənəbl/ adj. 可行动的 Every alert must be actionable. 每条告警都必须可行动。 ★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
on-call engineer /ˌɒn ˈkɔːl ˌendʒɪˈnɪə(r)/ 值班工程师 The on-call engineer was paged at 3 a.m. 值班工程师凌晨三点被寻呼。 ★★★
alert fatigue /əˈlɜːt fəˈtiːɡ/ 告警疲劳 Reduce noise to avoid alert fatigue. 减少噪音以避免告警疲劳。 ★★
blameless culture /ˈbleɪmləs ˈkʌltʃə(r)/ 无责文化 A blameless culture encourages honest reporting. 无责文化鼓励诚实上报。 ★★
follow-the-sun /ˈfɒləʊ ðə sʌn/ 随日轮换(跨时区值班) Follow-the-sun rotations share night duty globally. 随日轮换让全球分担夜间值班。 ★

第 23 课 事件管理与故障处理

Lesson 23: Incident Management & Troubleshooting

学习目标:掌握事件处理全流程(发现→响应→缓解→复盘)的英文词汇。

📖 英文阅读(Reading)

原文:

An incident is an unplanned event that disrupts service. The moment an incident is declared, the team shifts into response mode. An incident commander leads the response, while others do triage — quickly classifying the problem by severity and impact: how many users are affected, how badly, and for how long.

The immediate goal is containment and mitigation, not deep debugging. Teams apply a workaround — a temporary fix that restores service while the real fix is developed. It could be restarting a service, scaling up, or switching traffic to a healthy region. Restoring service is called recovery, and the time until then is measured as time to restore. Communication matters: the team updates a status page so users and stakeholders know what is happening.

Afterward comes the postmortem — a written review of the incident. It documents the timeline, the impact, the root cause, and the action items that will prevent recurrence. Great teams run post-incident reviews in a blameless way: the goal is improving the system, not blaming people. Every major incident is an opportunity to make the whole platform more robust.

中文翻译:

事件(incident)是打断服务的计划外状况。事件一经宣布,团队就进入响应模式。一名事件指挥官(incident commander)领导响应,其他人做分诊(triage)——按严重级别和影响(impact)快速分类问题:多少用户受影响、影响多严重、持续多久。

当下的目标是遏制(containment)和缓解(mitigation),而不是深度调试。团队采用临时解决方案(workaround)——在开发真正修复方案的同时恢复服务的临时手段。可能是重启服务、扩容(scale up),或把流量切换到健康区域。恢复服务叫恢复(recovery),在此之前的时间用恢复时间(time to restore)衡量。沟通很重要:团队更新状态页(status page),让用户和利益相关者了解进展。

之后是事后复盘(postmortem)——对事件的书面回顾。它记录时间线、影响、根本原因,以及防止复发要做的行动项(action item)。优秀的团队以无责方式开展事件后评审(post-incident review):目标是改进系统,而不是指责人。每一次重大事件都是让整个平台更健壮的机会。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
commander /kəˈmɑːndə(r)/ n. 指挥官 The commander coordinates all responders. 指挥官协调所有响应者。 ★★
triage /ˈtriːɑːʒ/ n./v. 分诊、分类 Triage decides which issue to fix first. 分诊决定先修哪个问题。 ★★
impact /ˈɪmpækt/ n. 影响 Assess the impact before acting. 行动前先评估影响。 ★★★
containment /kənˈteɪnmənt/ n. 遏制 Containment stops the blast radius from growing. 遏制阻止影响范围扩大。 ★★
mitigation /ˌmɪtɪˈɡeɪʃn/ n. 缓解 Mitigation restores service temporarily. 缓解临时恢复服务。 ★★
workaround /ˈwɜːkəraʊnd/ n. 临时解决方案 Use the workaround until the patch ships. 在补丁发布前使用临时方案。 ★★★
recovery /rɪˈkʌvəri/ n. 恢复 Recovery took 40 minutes. 恢复花了 40 分钟。 ★★★
postmortem /ˌpəʊstˈmɔːtəm/ n. 事后复盘 The postmortem lists five action items. 事后复盘列出了五个行动项。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
incident commander /ˈɪnsɪdənt kəˈmɑːndə(r)/ 事件指挥官 The incident commander assigns roles. 事件指挥官分配角色。 ★★
status page /ˈsteɪtəs peɪdʒ/ 状态页 Update the status page every 30 minutes. 每 30 分钟更新一次状态页。 ★★
action item /ˈækʃn ˌaɪtəm/ 行动项 Each postmortem produces action items. 每次复盘产生行动项。 ★★★
scale up / down /skeɪl ʌp daʊn/ 扩容/缩容 Scale up the API fleet during the sale. 促销期间给 API 服务扩容。 ★★★

第 24 课 站点可靠性工程(SRE)

Lesson 24: Site Reliability Engineering

学习目标:掌握 SLO、SLA、错误预算、混沌工程等可靠性英文词汇。

📖 英文阅读(Reading)

原文:

Site reliability engineering (SRE) applies software engineering to operations, with one core idea: reliability is a product decision, not an accident. SRE teams define a service level objective (SLO) — a target like "99.9% of requests succeed this month" — and customers sign a service level agreement (SLA) that promises a certain level of availability, often with penalties if missed.

The most powerful SRE concept is the error budget: if the SLO allows 0.1% failure, that 0.1% is a budget the team can "spend" on risky changes. As long as errors stay within budget, teams can deploy freely; once the budget is exhausted, releases slow down until reliability recovers. This turns reliability into a trade-off the whole company can discuss rationally.

SREs work on capacity planning (will we run out of resources at holiday traffic?), redundancy (are there backups for every component?), and failover (can traffic switch to a backup automatically?). They also practice chaos engineering — deliberately breaking parts of the system to test resilience. Finally, SREs fight toil: repetitive manual work that can and should be automated. The less toil, the more time for engineering that makes the system genuinely better.

中文翻译:

站点可靠性工程(SRE)把软件工程应用于运维,核心思想是:可靠性是产品决策,不是偶然。SRE 团队定义服务级别目标(SLO)——比如"本月 99.9% 的请求成功"——客户则签署服务级别协议(SLA),承诺一定水平的可用性,未达标往往伴随赔偿。

SRE 最强大的概念是错误预算(error budget):如果 SLO 允许 0.1% 的失败,那 0.1% 就是团队可以"花费"在有风险变更上的预算。只要错误在预算内,团队就能自由发布;一旦预算耗尽,发布就要放慢,直到可靠性恢复。这把可靠性变成全公司可以理性讨论的取舍(trade-off)。

SRE 的工作还包括容量规划(capacity planning)(节假日流量下资源会耗尽吗?)、冗余(redundancy)(每个组件都有备份吗?)和故障转移(failover)(流量能否自动切到备份?)。他们还实践混沌工程(chaos engineering)——故意破坏系统的一部分来测试韧性。最后,SRE 与琐事(toil)作斗争:那些可以也应该被自动化的重复性手工工作。琐事越少,花在真正改进系统的工程上的时间就越多。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
objective /əbˈdʒektɪv/ n. 目标 The SLO is our reliability target. SLO 是我们的可靠性目标。 ★★★
agreement /əˈɡriːmənt/ n. 协议 The SLA promises 99.9% availability. SLA 承诺 99.9% 的可用性。 ★★★
budget /ˈbʌdʒɪt/ n. 预算 We spent half the error budget this week. 这周我们花掉了一半错误预算。 ★★
capacity /kəˈpæsəti/ n. 容量 Capacity planning prevents holiday outages. 容量规划防止节假日宕机。 ★★★
redundancy /rɪˈdʌndənsi/ n. 冗余 Redundancy keeps data safe. 冗余保障数据安全。 ★★
failover /ˈfeɪləʊvə(r)/ n. 故障转移 Failover switches traffic to the backup. 故障转移把流量切到备份。 ★★
chaos engineering /ˈkeɪɒs ˌendʒɪˈnɪərɪŋ/ n. 混沌工程 Chaos engineering tests our resilience. 混沌工程测试我们的韧性。 ★★
toil /tɔɪl/ n. 琐事(重复性工作) Automate toil wherever possible. 尽可能把琐事自动化。 ★★
resilience /rɪˈzɪliəns/ n. 韧性、弹性 Resilience means failing without falling. 韧性意味着失败但不倒下。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
service level objective /ˈsɜːvɪs ˈlevl əbˈdʒektɪv/ 服务级别目标(SLO) Set an SLO you can actually measure. 设定一个可实际测量的 SLO。 ★★★
service level agreement /ˈsɜːvɪs ˈlevl əˈɡriːmənt/ 服务级别协议(SLA) The SLA includes a refund clause. SLA 包含赔偿条款。 ★★★
error budget /ˈerə(r) ˈbʌdʒɪt/ 错误预算 The error budget allows 43 minutes of downtime monthly. 错误预算允许每月 43 分钟停机。 ★★
capacity planning /kəˈpæsəti ˈplænɪŋ/ 容量规划 Capacity planning uses traffic forecasts. 容量规划使用流量预测。 ★★

第 25 课 性能优化与延迟

Lesson 25: Performance & Latency

学习目标:掌握性能问题的定位与优化英文词汇。

📖 英文阅读(Reading)

原文:

Performance is how fast and efficiently a system responds under load. The key measurements are response time (how long a request takes) and throughput (how many requests the system handles per second). When performance degrades, the first step is finding the bottleneck — the single slowest component that limits everything else. It might be a slow database query, a blocked thread pool, or an inefficient algorithm.

Engineers use profilers to see where CPU time and memory go, and run benchmarks to compare versions and measure improvement. Load testing pushes the system with simulated traffic to find its breaking point and observe how it degrades under stress: does it slow down gracefully or crash? A sudden spike in traffic — like a holiday sale — is the classic test of performance readiness.

The most powerful performance tool is the cache: storing frequent results in fast memory so repeated requests never touch the slow backend. Other techniques include batching database operations, adding indexes, and optimizing hot paths. Performance work is a constant loop: measure, find the bottleneck, fix it, measure again. Never guess — always measure.

中文翻译:

性能衡量的是系统在负载下响应有多快、效率有多高。关键度量是响应时间(response time)(一个请求耗时多少)和吞吐量(throughput)(系统每秒处理多少请求)。当性能下降时,第一步是找到瓶颈(bottleneck)——限制一切的单个最慢组件。它可能是慢的数据库查询、被阻塞的线程池,或低效的算法。

工程师用分析器(profiler)观察 CPU 时间和内存花在哪里,用基准测试(benchmark)比较版本、衡量改进(improvement)。压测(load testing)用模拟流量冲击系统,找到它的临界点,观察它在压力下如何劣化(degrade):是优雅变慢还是直接崩溃?流量峰值(spike)——比如节假日大促——是性能就绪度的经典考验。

最强大的性能工具是缓存(cache):把高频结果存进快速内存,让重复请求不再触碰慢速后端。其他技术包括批量处理(batching)数据库操作、加索引、优化(optimize)热点路径。性能工作是永恒的循环:测量、找瓶颈、修复、再测量。永远不要猜——永远要测量。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
bottleneck /ˈbɒtlnek/ n. 瓶颈 The database is the main bottleneck. 数据库是主要瓶颈。 ★★★
throughput /ˈθruːpʊt/ n. 吞吐量 Throughput dropped to 200 requests per second. 吞吐量降到每秒 200 个请求。 ★★
response time /rɪˈspɒns taɪm/ n. 响应时间 Average response time is 80 ms. 平均响应时间是 80 毫秒。 ★★★
benchmark /ˈbenʃmɑːk/ n./v. 基准测试 Benchmark the new version against the old. 用旧版本对比基准测试新版本。 ★★
profiler /ˈprəʊfaɪlə(r)/ n. 性能分析器 The profiler found a memory leak. 分析器发现了一个内存泄漏。 ★★
degrade /dɪˈɡreɪd/ v. 劣化、降级 The service degrades gracefully under load. 服务在负载下优雅降级。 ★★
spike /spaɪk/ n. 峰值、激增 A traffic spike hit the API at noon. 中午 API 遭遇流量峰值。 ★★
optimize /ˈɒptɪmaɪz/ v. 优化 Optimize the hot path first. 先优化热点路径。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
load testing /ləʊd ˈtestɪŋ/ 压力测试 Load testing revealed the breaking point. 压测暴露了崩溃临界点。 ★★★
performance tuning /pəˈfɔːməns ˈtjuːnɪŋ/ 性能调优 Performance tuning is an iterative process. 性能调优是迭代过程。 ★★
memory leak /ˈmeməri liːk/ 内存泄漏 A memory leak crashes the server eventually. 内存泄漏最终会让服务器崩溃。 ★★
caching strategy /ˈkæʃɪŋ ˈstrætədʒi/ 缓存策略 A good caching strategy cuts latency. 好的缓存策略大幅降低延迟。 ★★

第 26 课 安全与 DevSecOps

Lesson 26: Security & DevSecOps

学习目标:掌握应用安全领域的英文词汇(漏洞、加密、认证、审计)。

📖 英文阅读(Reading)

原文:

Security is not a phase at the end of development — it is a practice woven into every step. DevSecOps integrates security into the development lifecycle: vulnerability scanning runs automatically on code and dependencies, and threat modeling happens during design, asking "how could an attacker abuse this?"

Applications protect data with encryption — encoding information so only authorized parties can read it, both in transit and at rest. Access is controlled by authentication (proving who you are) and authorization (deciding what you may do). A user logs in with credentials; once authenticated, a role-based system decides which actions they are allowed to perform. Weaknesses here are the top target of attackers.

Teams verify their defenses with penetration testing — authorized attempts to break into the system — and security audits that review code, configs, and processes. Compliance with standards like GDPR or PCI-DSS may be legally required. When a vulnerability is discovered, a patch must be tested and deployed quickly, sometimes within hours. Security breaches are costly and damaging, so prevention, detection, and a fast response plan are non-negotiable.

中文翻译:

安全不是开发末尾的一个阶段——它是编织进每一步的实践。DevSecOps 把安全整合进开发生命周期:漏洞扫描(vulnerability scanning)在代码和依赖上自动运行,威胁建模(threat modeling)在设计阶段进行,问"攻击者会如何滥用这个?"

应用用加密(encryption)保护数据——把信息编码,只有授权方才能读取,传输中和静态存储时都如此。访问由认证(authentication)(证明你是谁)和授权(authorization)(决定你能做什么)控制。用户用凭据登录;认证之后,基于角色的系统决定允许他们执行哪些操作。这里的弱点正是攻击者的头号目标。

团队用渗透测试(penetration testing)——获得授权的入侵尝试——和安全审计(security audit)(审查代码、配置和流程)来验证防御。合规(compliance)——比如满足 GDPR 或 PCI-DSS 标准——可能是法律要求。发现漏洞(vulnerability)时,补丁(patch)必须在几小时内完成测试和部署。安全入侵(security breach)代价高昂且破坏巨大,因此预防、检测和快速的响应预案是底线。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
vulnerability /ˌvʌlnərəˈbɪləti/ n. 漏洞 The scanner found a high-risk vulnerability. 扫描器发现了一个高危漏洞。 ★★★
encryption /ɪnˈkrɪpʃn/ n. 加密 Use encryption for data at rest. 对静态数据使用加密。 ★★★
authentication /ɔːˌθentɪˈkeɪʃn/ n. 认证 Authentication verifies who you are. 认证验证你是谁。 ★★★
authorization /ˌɔːθəraɪˈzeɪʃn/ n. 授权 Authorization controls what you can do. 授权控制你能做什么。 ★★★
penetration /ˌpenɪˈtreɪʃn/ n. 渗透 The penetration test found three issues. 渗透测试发现了三个问题。 ★★
compliance /kəmˈplaɪəns/ n. 合规 The system must meet compliance requirements. 系统必须满足合规要求。 ★★★
audit /ˈɔːdɪt/ n./v. 审计 The security audit runs every quarter. 安全审计每季度执行一次。 ★★
patch /pætʃ/ n./v. 补丁 Patch the vulnerability before it is exploited. 在漏洞被利用前打上补丁。 ★★★
breach /briːtʃ/ n. 入侵、数据泄露 The breach exposed 2 million records. 入侵泄露了 200 万条记录。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
vulnerability scanning /ˌvʌlnərəˈbɪləti ˈskænɪŋ/ 漏洞扫描 Vulnerability scanning runs on every build. 每次构建都运行漏洞扫描。 ★★★
access control /ˈækses kənˈtrəʊl/ 访问控制 Access control follows the least-privilege rule. 访问控制遵循最小权限原则。 ★★★
zero trust /ˈzɪərəʊ trʌst/ 零信任 Zero trust verifies every request. 零信任验证每一个请求。 ★★
security breach /sɪˈkjʊərəti briːtʃ/ 安全入侵 A security breach is a company-wide crisis. 安全入侵是全公司范围的危机。 ★★★

第 27 课 云原生与云计算

Lesson 27: Cloud Native & Cloud Computing

学习目标:掌握云计算资源、弹性、区域部署的英文词汇。

📖 英文阅读(Reading)

原文:

Cloud computing delivers computing resources over the internet, on demand, at scale. Instead of buying physical servers, teams rent virtual machines — servers emulated in software — or use serverless platforms, where you upload code and the cloud runs it without you managing any servers at all. Each running machine is an instance, and you pay only for what you use, in a pay-as-you-go model.

The cloud's superpower is elasticity: resources scale up automatically when demand rises and scale down when it falls, so you never overpay and never run out. Resources live in regions (geographic areas) and availability zones (isolated data centers within a region); running in multiple zones gives high availability against local failures. Teams choose a cloud provider (AWS, Azure, or Google Cloud) and often combine on-premises infrastructure with cloud — a hybrid cloud — for legacy systems or regulatory reasons.

Cloud native means designing applications specifically for the cloud: containers, microservices, automation, and managed services that offload undifferentiated heavy lifting. The benefits are speed, scale, and cost efficiency — but the cloud bill can surprise you, so teams monitor usage carefully and shut down idle resources.

中文翻译:

云计算(cloud computing)通过互联网按需、规模化地交付计算资源。团队不再购买物理服务器,而是租用虚拟机(virtual machine)——用软件模拟的服务器——或使用无服务器(serverless)平台:上传代码,云自动运行,完全不用你管理服务器。每台运行的机器是一个实例(instance),按按量付费模式只为实际使用的部分付钱。

云的超级能力是弹性(elasticity):需求上升时资源自动扩容,需求下降时自动缩容,既不浪费钱也不至于不够用。资源分布在区域(region)(地理区域)和可用区(availability zone)(区域内隔离的数据中心)中;跨多个可用区运行能针对局部故障提供高可用。团队选择一家云服务商(cloud provider)(AWS、Azure 或 Google Cloud),并常把本地(on-premises)基础设施与云结合——即混合云(hybrid cloud)——用于遗留系统或合规原因。

云原生(cloud native)意味着专门为云设计应用:容器、微服务、自动化和把无差别繁重工作外包出去的托管服务(managed service)。好处是速度、规模与成本效益(cost efficiency)——但云账单可能会吓你一跳,所以团队密切监控用量、及时关闭闲置资源。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
cloud /klaʊd/ n. 云(云计算) The app runs entirely in the cloud. 应用完全运行在云端。 ★★★
virtual machine /ˈvɜːtʃuəl məˈʃiːn/ n. 虚拟机 Spin up a virtual machine for testing. 为测试启动一台虚拟机。 ★★★
instance /ˈɪnstəns/ n. 实例 We run four instances of the service. 我们运行服务的四个实例。 ★★★
serverless /ˈsɜːvləs/ adj. 无服务器的 Serverless functions scale automatically. 无服务器函数自动伸缩。 ★★
elasticity /ˌiːlæˈstɪsəti/ n. 弹性 Elasticity matches cost to demand. 弹性让成本与需求匹配。 ★★
region /ˈriːdʒn/ n. 区域 Deploy to the region closest to users. 部署到离用户最近的区域。 ★★★
availability zone /əˌveɪləˈbɪləti zəʊn/ n. 可用区 Run across two availability zones. 跨两个可用区运行。 ★★
on-demand /ˌɒn dɪˈmɑːnd/ adj. 按需的 Cloud resources are on-demand. 云资源是按需获取的。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
cloud provider /klaʊd prəˈvaɪdə(r)/ 云服务商 Compare pricing across cloud providers. 比较不同云服务商的价格。 ★★★
pay-as-you-go /ˌpeɪ æz juː ˈɡəʊ/ 按量付费 Pay-as-you-go pricing suits variable traffic. 按量付费适合流量波动。 ★★
hybrid cloud /ˈhaɪbrɪd klaʊd/ 混合云 The bank uses a hybrid cloud model. 这家银行采用混合云模式。 ★★
on-premises /ˌɒn ˈpremɪsɪz/ 本地的(自建机房) Legacy systems stay on-premises. 遗留系统留在本地机房。 ★★

第 28 课 数据库与数据管理

Lesson 28: Databases & Data Management

学习目标:掌握数据库设计、迁移、备份的英文词汇。

📖 英文阅读(Reading)

原文:

Almost every application stores data in a database. A relational database organizes data into tables with a schema that defines their structure, and applications read and write it with queries. Indexes speed up queries by organizing data for fast lookup — but they cost storage and slow down writes, so they must be chosen carefully. A transaction groups several operations into one atomic unit: either all succeed or none do, keeping data consistent.

As systems grow, databases need careful management. Migrations change the schema safely without downtime. Backups copy data so it can be restored after a disaster — tested regularly, because an untested backup is no backup at all. Replication keeps copies of the data on multiple servers, usually with a master handling writes and replicas handling reads, which improves both reliability and read performance. For very large datasets, sharding splits data across many machines, each holding a slice.

A database outage can bring the whole company down, so teams invest in monitoring, connection pooling, and failover configurations. Data integrity and consistency are the ground rules everything else builds on.

中文翻译:

几乎每个应用都把数据存放在数据库(database)中。关系型数据库(relational database)把数据组织成带模式(schema)(定义结构)的表,应用通过查询(query)读写数据。索引(index)通过组织数据实现快速查找来加速查询——但它占用存储、拖慢写入,所以必须谨慎选择。事务(transaction)把多个操作组合成一个原子单元:要么全部成功,要么全部失败,从而保持数据一致。

随着系统增长,数据库需要精心管理。迁移(migration)在不停机的情况下安全地改变模式。备份(backup)复制数据,让灾难后能恢复(restore)——而且要定期演练,因为没演练过的备份等于没有备份。复制(replication)在多台服务器上保留数据副本,通常主库(master)处理写入、副本(replica)处理读取,同时提升可靠性和读性能。对超大数据集,分片(sharding)把数据分散到多台机器,每台持有其中一片。

数据库宕机可能让整个公司瘫痪,所以团队在监控、连接池(connection pooling)和故障转移配置上投入大量精力。数据完整性(data integrity)和一致性(consistency)是其他一切赖以建立的基本规则。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
database /ˈdeɪtəbeɪs/ n. 数据库 The database stores user accounts. 数据库存储用户账户。 ★★★
schema /ˈskiːmə/ n. 模式(表结构) Changing the schema requires a migration. 修改模式需要一次迁移。 ★★
query /ˈkwɪəri/ n./v. 查询 The query returns the last ten orders. 该查询返回最近十个订单。 ★★★
migration /maɪˈɡreɪʃn/ n. 迁移 Run the migration before deploying. 部署前先执行迁移。 ★★★
backup /ˈbækʌp/ n./v. 备份 Back up the database every night. 每晚备份数据库。 ★★★
restore /rɪˈstɔː(r)/ v. 恢复(数据) We restored the data from the backup. 我们从备份中恢复了数据。 ★★★
replication /ˌreplɪˈkeɪʃn/ n. 复制 Replication provides read scalability. 复制提供读扩展能力。 ★★
sharding /ˈʃɑːdɪŋ/ n. 分片 Sharding splits data across servers. 分片把数据分散到多台服务器。 ★★
index /ˈɪndeks/ n. 索引 Add an index on the user_id column. 在 user_id 列上加索引。 ★★★
transaction /trænˈzækʃn/ n. 事务 The transaction rolls back on error. 出错时事务回滚。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
database migration /ˈdeɪtəbeɪs maɪˈɡreɪʃn/ 数据库迁移 Database migration must be reversible. 数据库迁移必须可回退。 ★★★
data backup /ˈdeɪtə ˈbækʌp/ 数据备份 Test your data backup restore monthly. 每月演练一次数据备份恢复。 ★★★
data integrity /ˈdeɪtə ɪnˈteɡrəti/ 数据完整性 Transactions protect data integrity. 事务保护数据完整性。 ★★
connection pooling /kəˈnekʃn ˈpuːlɪŋ/ 连接池 Connection pooling reuses database connections. 连接池复用数据库连接。 ★★

第 29 课 团队协作与沟通

Lesson 29: Team Collaboration & Communication

学习目标:掌握软件团队日常沟通、会议、文档的英文词汇。

📖 英文阅读(Reading)

原文:

Software is built by people, not just code. Strong communication keeps a team aligned: everyone knows the goals, the current state, and who owns what. Teams hold sync meetings to share progress, raise blockers, and make decisions together. Alignment means everyone is moving in the same direction; misalignment produces duplicated work and missed deadlines.

As work moves between people and teams, handoffs must be clean: a clear documentation trail — design docs, READMEs, and decision records — lets the next person pick up where you left off. A good engineer writes for the future reader: "why" matters as much as "what". Team culture values transparency (share problems early), accountability (own your commitments), and consensus (everyone understands the decision, even if they disagree). Follow-ups turn meetings into action: each discussion ends with owners and deadlines.

In a global team, async communication — written updates, reviews, and wikis — matters as much as meetings, because colleagues live in different time zones. Great teams communicate in writing by default, so knowledge survives after any meeting ends.

中文翻译:

软件由人构建,而不只是代码。强大的沟通(communication)让团队保持一致:每个人都知道目标、当前状态以及谁负责什么。团队开同步(sync)会议分享进度、提出阻塞项、共同做决策(decision)。对齐(alignment)意味着所有人朝同一方向前进;不对齐会产生重复劳动和错过的截止日期。

工作在人、团队之间流转时,交接(handoff)必须干净利落:清晰的文档(documentation)痕迹——设计文档、README、决策记录——让下一个人能接着你的进度继续。优秀的工程师为未来的读者而写:"为什么"和"是什么"同样重要。团队文化重视透明度(transparency)(尽早暴露问题)、责任制(accountability)(兑现你的承诺)和共识(consensus)(即使有人不同意,也理解这个决定)。跟进(follow-up)把会议变成行动:每场讨论都以责任人和截止日期收尾。

在全球团队中,异步沟通(async communication)——书面更新、评审、wiki——与会议同样重要,因为同事身处不同时区。优秀的团队默认用书面沟通,这样知识在会议结束后依然留存。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
communication /kəˌmjuːnɪˈkeɪʃn/ n. 沟通 Clear communication prevents misunderstandings. 清晰的沟通避免误解。 ★★★
sync /sɪŋk/ n./v. 同步(对齐) Let's sync on the API design today. 我们今天同步一下 API 设计。 ★★★
alignment /əˈlaɪnmənt/ n. 对齐、一致 Alignment on goals saves wasted work. 目标对齐避免浪费工作。 ★★★
handoff /ˈhændɒf/ n. 交接 A clean handoff includes docs and context. 干净的交接包括文档和上下文。 ★★
documentation /ˌdɒkjumenˈteɪʃn/ n. 文档 Good documentation reduces questions. 好的文档减少提问。 ★★★
transparency /trænsˈpærənsi/ n. 透明度 Transparency builds team trust. 透明度建立团队信任。 ★★
accountability /əˌkaʊntəˈbɪləti/ n. 责任制 Accountability means owning your work. 责任制意味着为自己的工作负责。 ★★
consensus /kənˈsensəs/ n. 共识 We reached consensus on the plan. 我们对计划达成了共识。 ★★
async /ˌeɪˈsɪŋk/ adj. 异步的 Async updates help remote colleagues. 异步更新帮助远程同事。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
status update /ˈsteɪtəs ˈʌpdeɪt/ 进度更新 Give a status update in the channel. 在群里发一条进度更新。 ★★★
follow up /ˈfɒləʊ ʌp/ 跟进 I will follow up on the action items. 我会跟进这些行动项。 ★★★
get aligned /ɡet əˈlaɪnd/ 达成一致 Let's get aligned before the sprint starts. 冲刺开始前我们先对齐。 ★★
keep in the loop /kiːp ɪn ðə luːp/ 保持知情 Keep the QA team in the loop on changes. 让测试团队对变更保持知情。 ★★

第 30 课 技术面试与职业英语

Lesson 30: Technical Interview & Career English

学习目标:掌握求职、面试、晋升场景的高频英文表达。

📖 英文阅读(Reading)

原文:

Your technical skills get you the interview; your English gets you through it. A strong resume lists achievements with numbers — "reduced latency by 40%" — rather than duties. A portfolio with real projects and a GitHub profile shows more than words. In behavioral interviews, interviewers ask about past experiences: "Tell me about a time you disagreed with a teammate." Use the STAR method: Situation, Task, Action, Result.

In technical interviews, you may walk through your projects, solve problems on a whiteboard, or explain design trade-offs. Speak in structured terms: architecture, testing, deployment, and monitoring. When you do not know an answer, say so honestly and explain how you would find out — interviewers value honesty and problem-solving over memorized answers.

Career growth needs more than code. Negotiation matters when you receive an offer: research salaries, know your expectations, and discuss compensation calmly. After joining, focus on onboarding — learning the codebase and culture quickly — and seek mentorship from senior engineers. Promotion comes from impact, visibility, and helping others grow. Finally, plan your career path: decide whether you want to go deep as an individual contributor or wide as an engineering manager.

中文翻译:

你的技术能力让你拿到面试,你的英语让你通过面试。一份优秀的简历(resume)用数字列出成就——"延迟降低 40%"——而不是罗列职责。有真实项目和 GitHub 主页的作品集(portfolio)比空话更有说服力。在行为面试(behavioral interview)中,面试官会问过去的经历:"讲一次你和队友意见不合的事。"用 STAR 法则:情境(Situation)、任务(Task)、行动(Action)、结果(Result)。

技术面试中,你可能要讲解(walk through)自己的项目、在白板上解题,或解释设计取舍。用结构化的语言表达:架构、测试、部署、监控。遇到不会的问题,坦诚说不知道,并解释你会如何查找答案——面试官看重诚实和解决问题(problem-solving)的能力,胜过背下来的答案。

职业成长不止需要代码。收到录用通知(offer)时要谈薪资(negotiation):调研薪酬水平,清楚自己的预期(expectation),冷静讨论薪酬待遇(compensation)。入职后,重点做好入职培训(onboarding)——快速熟悉代码库和文化——并向资深工程师寻求指导(mentorship)。晋升(promotion)来自影响力、可见度和帮助他人成长。最后,规划你的职业路径(career path):决定是走深度(个人贡献者)还是走广度(工程管理者)。

📚 重点词汇(Key Vocabulary)

单词 音标 词性 中文释义 英文例句 中文翻译 频度
resume /ˈrezjumeɪ/ n. 简历 Tailor your resume to the job. 针对职位定制你的简历。 ★★★
portfolio /pɔːtˈfəʊliəʊ/ n. 作品集 A portfolio beats a long resume. 作品集胜过长长的简历。 ★★
behavioral /bɪˈheɪvjərəl/ adj. 行为的 Behavioral questions test soft skills. 行为问题考察软技能。 ★★
negotiation /nɪˌɡəʊʃiˈeɪʃn/ n. 谈判、议价 Negotiation starts after the offer. 议价从收到录用通知后开始。 ★★
onboarding /ˈɒnbɔːdɪŋ/ n. 入职培训 Good onboarding shortens time to productivity. 好的入职培训缩短上手时间。 ★★★
mentorship /ˈmentəʃɪp/ n. 导师指导 Mentorship accelerates career growth. 导师指导加速职业成长。 ★★
promotion /prəˈməʊʃn/ n. 晋升 Promotion rewards consistent impact. 晋升奖励持续的影响力。 ★★★
expectation /ˌekspekˈteɪʃn/ n. 预期 Know your salary expectations. 清楚你的薪资预期。 ★★★

🔗 重点短语(Key Phrases)

短语 音标 中文释义 英文例句 中文翻译 频度
walk through /wɔːk θruː/ 讲解、走一遍 Walk me through your last project. 给我讲讲你上一个项目。 ★★★
salary expectation /ˈsæləri ˌekspekˈteɪʃn/ 薪资预期 What are your salary expectations? 你的薪资预期是多少? ★★★
career path /kəˈrɪə(r) pɑːθ/ 职业路径 Plan your career path with your manager. 和你的经理一起规划职业路径。 ★★★
soft skills /sɒft skɪlz/ 软技能 Communication is a key soft skill. 沟通是一项关键软技能。 ★★

附录 A 词汇量统计

统计项 数量
总课时 30 课
重点词汇(Key Vocabulary) 约 300 个
重点短语(Key Phrases) 约 130 个
英文阅读材料 30 篇(约 5,000 词)
★★★ 极高频词汇 约 130 个(优先记忆)

说明:词汇与短语在全教材中刻意不重复——第 1 课 55 个基础词,第 2~30 课每课引入 8~10 个新词和 4 个新短语,学完即覆盖软件开发全领域核心词汇。


附录 B 30 天学习计划

阶段 天数 学习内容 复习安排
第一阶段:入门 第 1~6 天 第 1~2 课 + 第 3~4 课 + 第 5~6 课(每天 1 课) 第 2 天复习第 1 课
第二阶段:测试 第 7~12 天 第 7~8 课 + 第 9~10 课 + 第 11~12 课 第 8 天复习第 1~3 课
第三阶段:部署 第 13~18 天 第 13~14 课 + 第 15~16 课 + 第 17~18 课 第 15 天复习第 4~8 课
第四阶段:运维 第 19~24 天 第 19~20 课 + 第 21~22 课 + 第 23~24 课 第 21 天复习第 9~15 课
第五阶段:进阶 第 25~28 天 第 25~26 课 + 第 27~28 课 第 26 天复习第 16~22 课
第六阶段:收尾 第 29~30 天 第 29~30 课 第 30 天总复习 + 自测造句

每天建议用时 30~45 分钟:精读原文 → 对照翻译 → 朗读词汇例句 → 用 ★★★ 词造 2 个句子。


附录 C 高频缩写速查表

缩写 全称 中文含义 所在课时
CI Continuous Integration 持续集成 第 1、11 课
CD Continuous Delivery / Deployment 持续交付 / 持续部署 第 1、12 课
QA Quality Assurance 质量保证 第 1 课
PR Pull Request 合并请求 第 4、5 课
E2E End-to-End 端到端(测试) 第 9 课
UAT User Acceptance Testing 用户验收测试 第 2 课(验收标准)
API Application Programming Interface 应用程序接口 第 9、17 课
IaC Infrastructure as Code 基础设施即代码 第 19 课
K8s Kubernetes 容器编排平台 第 16 课
VM Virtual Machine 虚拟机 第 27 课
SLO Service Level Objective 服务级别目标 第 24 课
SLA Service Level Agreement 服务级别协议 第 24 课
SRE Site Reliability Engineering 站点可靠性工程 第 24 课
MTTR Mean Time To Recovery 平均恢复时间 第 12 课
DORA DevOps Research and Assessment DevOps 研究评估框架(四个核心指标) 第 12 课
STAR Situation, Task, Action, Result 行为面试回答法则 第 30 课
DDoS Distributed Denial of Service 分布式拒绝服务(攻击) 第 26 课(安全延伸)

完成祝贺:学完 30 课时,你已掌握软件开发、测试、部署、运维全流程的约 290 个核心词汇和约 130 个常用短语。建议每周重读一篇原文保持语感,并在工作中主动使用这些词汇——英语和工程能力一样,都是用出来的。

posted @ 2026-09-11 13:38  渔樵江渚  阅读(17)  评论(0)    收藏  举报