ljg-book

ljg-book: Read a book as x → f → f(x)

Reduce a book to a complete logical chain:

• x: What problem is the author addressing?
• f: What is the author's core answer? It could be a concept, distinction, framework, method, model, or narrative perspective.
• f(x): After applying that answer back to the problem, how should we re-understand, re-judge, or re-act?

The goal is not to compress the book into a platitude or to audit the author's framework. After reading the note, you should know both "what the book is about" and "how it changes my response to the problem."

Workflow Routing

Input	Required Reading	Output
Book title	After finding reliable material, read ReadingGuide.md	Save org note
PDF, full text, sample chapter, old notes	Read material first, then ReadingGuide.md	Save org note
User explicitly wants only oral explanation	ReadingGuide.md	No file written; explain using same path

When writing the org file, also read references/template.org. Default save location: ~/Documents/notes/.

Filename follows Denote: {YYYYMMDDTHHMMSS}--拆书-{书名}__book.org; timestamp generated with date +%Y%m%dT%H%M%S.

Completion Target

Start with three minimal lines:

- *x*: {What problem is the author discussing?}
- *f*: {What core answer does the author use to handle x?}
- *f(x)*: {After accepting f, how should we respond to x?}

The three lines must read as a connected chain: because x exists, the author proposes f; applying f to x yields f(x). If the three lines are just three related points, the logical chain is not yet connected.

Then expand the chain in the body. Typically 900–1600 Chinese characters, excluding frontmatter and source calibration; shorter books can be shorter; complex books exceed 1600 characters only when the logic truly requires it.

Understanding Bottleneck Scan

After completing f, identify the biggest understanding bottleneck in the book: which mechanism is hardest to form in the mind through text alone, but once grasped, all other concepts fall into place? Choose the lowest-cost representation for it, rather than defaulting to more prose.

Shape of Thought	Preferred Representation
Convex/concave, distribution, marginal change, spatial relationship	ASCII curve or shape diagram
Sequential steps, causal chain, state change	Flowchart or timeline
Loop, feedback, self-reinforcement	Loop diagram
Hierarchy, part-whole	Structure diagram
Two mechanisms compared repeatedly	Small table
One or two sentences suffice	Plain text

Diagrams are not mandatory; but when the thought itself has shape, spatial relationship, or dynamic change, do not merely paraphrase it in text. The diagram must serve x → f → f(x), so that a reader can tell from the diagram alone which judgment the author changed.

• ASCII diagrams go inside #+begin_example / #+end_example blocks.
• Keep the diagram undistorted, width ≤ 80 characters.
• One diagram is enough; if it does not reduce understanding cost, delete it.

Minimal form:

#+begin_example
x: original problem
        │
        ▼
f: author's concept / framework / method
        │ apply
        ▼
f(x): new understanding / judgment / action
#+end_example

Gotchas

• x is not a topic. "Technology and the future" is not x; "Does technological evolution have its own direction?" is.
• f is not a label. Writing only "systems thinking," "existentialism," or "antifragile" is not an answer; explain how it handles x.
• f is not a contribution type. Do not output source, weight, originality, contribution type, shift type, or other review judgments unless the user explicitly asks.
• Do not jump from x directly to advice. Without f, the note becomes only generic life lessons.
• Do not write f as a table of contents. There can be 2–4 necessary components, but they must together complete the same answer.
• f(x) is not a slogan. Specify exactly how understanding, judgment order, or action changes after accepting f.
• Literary works also have f. It may not be a methodology but an image, character relationship, or narrative movement that makes a problem visible again; do not force it into an efficiency tool.
• Keep research in the background. Author background, intellectual lineage, and external debates are only used to avoid misinterpretation; the body retains only what is truly needed to understand x, f, and f(x).
• Do not interpret "diagrams are not mandatory" as draw fewer diagrams. First find the understanding bottleneck; if the core idea depends on shape, space, or dynamic relationships, omitting the diagram means omitting part of f.
• Do not just line-wrap a summary into a diagram. The diagram must make the mechanism visible; if a reader cannot tell the judgment change from the diagram alone, it has no explanatory role.
• If material is thin, say less. When only a table of contents, interview, or book review is available, mark as "初拆" (initial breakdown); when evidence is even thinner, mark as "假设版" (hypothetical version).

Quick Reference

1. Narrow the topic to a specific x.
2. Identify the main f the author proposes for x, not a list of all concepts in the book.
3. Use 2–4 anchors from the book to explain how f works and why it holds.
4. Compute f(x): if f is truly accepted, what changes when facing x?
5. Scan the biggest understanding bottleneck; use the lowest-cost representation (diagram, table, or text) to make it visible.
6. Delete anything that does not connect to this logical chain.

Top-level sections are fixed:

1. * x：作者在讨论什么问题
2. * f：作者怎样回答
3. * f(x)：怎样回应这个世界
4. * 资料校准

Examples

Analytical book: Judgment in Managerial Decision Making

• x: Why does the same choice, when framed differently, lead people to opposite decisions?
• f: Prospect theory explains judgment bias through reference points, loss aversion, and nonlinear probability weighting.
• f(x): For important choices, first reframe the gains and losses; if the answer changes with the frame, postpone the decision.

Literary work: Siddhartha

• x: Can wisdom be transmitted from one person to another like knowledge?
• f: Teachings can point the way, but wisdom is generated through personal experience, listening, and integrating contradictory experiences; the river is the central image for this answer.
• f(x): When seeking answers from others, first distinguish whether you lack teachable knowledge or experience you have not yet personally digested.

Completion

After generating, read back to confirm:

• File is saved, frontmatter is complete.
• The opening has exactly three lines for x / f / f(x).
• x is a problem, f is an answer, f(x) is the result of applying it; they are not parallel summaries.
• All four body sections are present, and each contributes to the same logical chain.
• The body has at least 2 anchors from the book, not chapter summaries following the table of contents.
• f explains the internal mechanism, not degenerating into a term or concept list.
• f(x) specifies the change in understanding, judgment, or action.
• The body contains no review information such as source, weight, contribution type, or shift type.
• Understanding bottleneck scan is done; the hardest-to-form mechanism has an appropriate representation, or it is confirmed that a short sentence suffices.
• ASCII diagrams appear only when they truly improve understanding, with correct format and width.
• If the core idea depends on shape, space, or dynamic relationships, the diagram is not omitted; the diagram alone can convey the judgment the author changed.
• Source calibration includes only necessary sources, not expanded into a research review.
• Reading only the opening three lines answers: "What question does the book ask, how does it answer, and what should we do after the answer?"