Share E-Book
Scan to open this page

Scan with your phone to open this page

Author: Stefan Baumgartner

Rating 3.5 AI initial rating

AI writes a lot of code now, so an engineer's real output becomes the decisions made around the code, choices that will impact projects for years, admitting little room for failure. This book takes engineers on a deep dive beyond the hype to the core of web application development. With this book, you'll understand the consequences of architectural and technical decisions while examining web apps from every angle, enabling the creation of long-lasting solutions, faster bug detection, and the ability to deliver focused, high-impact code reviews. With this knowledge, you'll zero in on what truly matters for your projects.

AI Reading Assistant

Whole-book reading guide from stratified index samples; jump to passages in the text

AI guide
【One-Line Pitch】 A senior-engineer's field guide to web development that treats architecture, tooling, and debugging as a chain of human decisions rather than API trivia—written for developers who have shipped real software and now need to understand the whole stack well enough to make calls that outlive them. 【Book Arc】 - **Opening (~0%–10%)**: Frames the premise—AI absorbs the typing, so an engineer's real output becomes decisions—and defines who the book serves: developers on the way to senior/principal who keep hitting the edges of their knowledge. Introduces the "V-shaped" engineer and Conway's law as the book's organizing tension. - **Early (~10%–30%)**: Establishes the two bookends. Chapter 1 asks what a senior engineer is now that coding is partly "solved," arguing for decisions and reach over memorized APIs; Chapter 11 (Egoless Programming) returns to the people those decisions land on. Also covers the "layers of the web" model and why lower layers move slowly by design. - **Early–Middle (~30%–50%)**: The decisions made first and regretted longest—JavaScript tooling (npm, dependency graphs, keeping builds comprehensible a year later) and architecture (client/server split, the C4 model, why system boundaries reappear in the org chart). Tooling is reframed as a shared team vocabulary. - **Middle (~50%–70%)**: Follows a request from wire to screen: networking (what HTTP actually gives you, API design, designing for unreliable networks), rendering (client, server, and in-between), and routing/state management (keeping URLs and app state in sync). - **Late (~70%–90%)**: The daily material—robust CSS (specificity and the cascade behind most styling fights at scale), effective JavaScript (event loop, bundlers, error propagation), and frameworks as a bundle of all prior decisions that outlives the enthusiasm behind it. - **Ending (~90%–100%)**: Debugging as the payoff chapter—DevTools panels worth knowing, techniques that narrow problems down, and two real-world bugs solved by method rather than persistence. Closes the frame with the human/ego dimension of senior work. 【Key Takeaways】 - **A senior engineer's output is a series of decisions, not code volume** (Early): framework choice, system layout, backend contracts, dependency trust, and on-call ownership are the real deliverables—AI can draft the code but not own these calls. - **The V-shaped engineer replaces the T-shaped one** (Early): depth in a couple of areas with reach that keeps extending sideways, because no decision stays contained in the code. - **Architecture happens whether or not you design it** (Early–Middle): remove the architect and the architecture only becomes implicit; Conway's law will draw your boundaries for you if nobody does it deliberately. - **The web's layers move at different speeds** (Middle): protocols barely change, frameworks change fast—and the slow layers are slow by design, refusing to break existing code (the `includes` vs. `contains` naming story). - **Good ideas sink down the stack** (Middle): patterns proven in libraries (jQuery's selectors, for example) eventually become native platform features, which is a useful lens for judging what's worth adopting. - **Tooling is a shared vocabulary and an interface** (Middle): build tools, package managers, linters, and test frameworks are how a team agrees to work—and maintaining that interface is senior work. - **Debugging is a method, not persistence** (Late): understanding how the layers work converts debugging from guesswork into narrowing techniques and knowing which DevTools panels to actually use. - **The frame is technical on one side, human on the other** (Ending): decisions land on people, and ego is the main obstacle to working with them—the book's closing argument ties engineering back to collaboration. 【Reading Tips】 - Read Chapter 1 and Chapter 11 first as the frame; everything between fills it in, so knowing the endpoints makes the middle chapters cohere. - Deep-read the chapters on decisions you're currently facing (tooling, architecture, frameworks); skim the ones where you already have production scar tissue. - Treat the debugging chapter as the payoff—return to it after the middle chapters so the layer-by-layer understanding has somewhere to land. - Watch for the recurring Conway's law thread across architecture, tooling, and team structure; it's the book's spine, not a one-off aside. - Note the author's explicit scope disclaimer: accessibility is out of scope and deferred to a dedicated reference, so don't expect coverage there. 【Coverage Limits】 These excerpts cover the front matter, preface, and chapter map plus early framing chapters; the detailed content of the middle and late chapters (networking, rendering, CSS, JavaScript, debugging) is described by the author's overview rather than shown directly, so specifics beyond that outline are not covered here.
Excerpt 1
eilly logo is a registered trademark of O’Reilly Media, Inc. Fluent Web Development , the cover image, and related trade dress are trademarks of O’Reilly Med...
View in text
Excerpt 2
keeps trying to replace the language instead of learning it. JavaScript gets one because the distance between code that works and code that works well runs t...
View in text
Excerpt 3
al That reach matters because no decision stays in the code. Each one lands on a person: the backend engineer who has to build the API you specified, whoever...
View in text
Excerpt 4
the current year. The layers do not move at the same speed. The lower ones barely move at all. Today, TCP/IP works the way it did decades ago. HTTP went year...
View in text
Excerpt 5
ies as they arrive, without rewriting what you already have. Leave that code untouched for two years and it will still run. Bind yourself to a framework and...
View in text
Excerpt 6
iting JavaScript code that doesn’t need to run in a browser. With that, folks started writing JavaScript tools in the browser using Node.js. One of the first...
View in text
Excerpt 7
s using our tool as it was intended is fundamentally flawed. If vulnerabilities can be exploited, we can’t rely on the knowledge of our end users to work aro...
View in text
Excerpt 8
by adding module systems, code splitting, and tree shaking. Before that, tools like the closure compiler and Ant used Java-based software to minify and bundl...
View in text
Tags
AI categories
Web TechnologySoftwareProgramming
Publish Year: 2026
Language: English
Pages: 427
File Format: EPUB
File Size: 1.3 MB