返回 ppt-master
design_spec.md
1 ---
2 style_id: science-explainer
3 kind: style
4 summary: Public-explanation method that builds understanding from familiar ground through visual analogy, without trading accuracy for accessibility.
5 keywords: [explainer, science-communication, popular, analogy, public]
6 ---
7
8 # Science Explainer — Style Specification
9
10 > Method and design defaults only. No project communication contract, brand identity, page structure, or SVG prototypes.
11
12 ## I. Style Overview
13
14 | Property | Value |
15 |---|---|
16 | Style Name | Science Explainer |
17 | Best Fit | Public lectures, museum and outreach talks, popular-science and technology explainers, journalist and policy briefings, and internal cross-discipline explanations |
18 | Reusable Intent | Take an audience with no background from curiosity to genuine understanding of one idea, leaving them able to explain it themselves and knowing what remains uncertain |
19 | Sources | Authored in-repo as a bundled reference Style, 2026-08-07; distilled from science-communication practice, not a single external document |
20
21 ## II. Communication Method
22
23 - **Preferred Mode**: instructional
24 - **Argument Flow**: Start from something the audience already knows or has felt, introduce the puzzle that familiar ground cannot explain, build the new idea one necessary step at a time, then return to the opening ground and show it differently. Every step earns the next; nothing is introduced because the field's own curriculum introduces it there. Prefer one idea understood deeply over five mentioned.
25 - **Page Message Discipline**: Give each page one step in the build, and title it in plain language as the idea rather than its technical name. Introduce a term only after the audience has met the thing it names, and keep the term with its plain-language meaning on the page where it first appears. Keep a visual and the idea it carries together; a diagram that needs the previous page has broken the build.
26 - **Claim Discipline**: Keep established consensus, active debate, and speculation distinct, and say which is which in plain words rather than through hedging that a lay audience cannot decode. Mark every analogy as an analogy and say where it stops being true before the audience finds the edge themselves. Simplify by omitting detail, never by stating something false; when a simplification would mislead, say the real thing is more complicated and name the direction of the difference. Give scale and uncertainty in terms the audience can feel, and never overstate certainty to make a cleaner story.
27
28 ## III. Page Role Vocabulary
29
30 | Role | Communication Job | Evidence Obligation | Composition Tendency |
31 |---|---|---|---|
32 | Familiar ground | Start where the audience already stands | Use an experience the audience genuinely shares, not one assumed of them | One concrete image or situation; no terminology yet |
33 | The puzzle | Create the question the explanation answers | Show why the familiar understanding fails here; make the gap felt | Keep the tension dominant and unresolved on this page |
34 | Building block | Introduce one necessary piece | Give only what the next step requires; connect it to what is already held | One idea, one visual, minimal text |
35 | Visual analogy | Make an abstraction graspable | State what the analogy maps to and where it breaks down | Let the analogy image dominate with its limit noted plainly nearby |
36 | Mechanism | Show how it actually works | Follow one concrete path; keep the sequence honest even when simplified | Sequence drives the composition; each step's state stays visible |
37 | Scale and magnitude | Make a number mean something | Anchor to a comparison the audience can feel; keep the arithmetic honest | One comparison per page; resist stacking multiple scale tricks |
38 | Evidence | Show how we know rather than asking for trust | Give the observation, experiment, or measurement in plain terms, with its source | Keep the evidence and what it establishes side by side |
39 | Common misconception | Correct what the audience likely believes | State the misconception plainly, then what is actually true and why the error is natural | Place the two accounts in direct correspondence; never mock the misconception |
40 | Edge of knowledge | Show where certainty ends | Distinguish what is settled from what is actively debated and what is unknown | Give uncertainty its own space rather than a closing hedge |
41 | Why it matters | Connect the idea to the audience's world | Tie to a real consequence or decision, without inflating stakes | One consequence, concretely stated |
42 | Going further | Point the curious onward | Give real, reachable sources at the audience's level | Quiet and simple; the talk has already landed |
43
44 ## IV. Evidence & Data Expression
45
46 - **Argument Trace**: Every claim traces to evidence the audience could in principle check, described in plain terms. Where the explanation rests on expert consensus rather than something showable, say so directly instead of implying the audience has just seen it demonstrated.
47 - **Charts**: Build a chart in stages rather than revealing it complete, explain the axes before the data, and label directly on the mark so no legend lookup is required. Use one relationship per chart and annotate what to read from it in plain words. Keep axes honest and unbroken, avoid log scales unless they are explained, and never assume familiarity with a conventional plot type the field takes for granted.
48 - **Tables**: Tables rarely help a lay audience. Use one only for a small, concrete comparison of a few items on a few axes, and keep every heading in plain language rather than field notation.
49 - **Sources**: Cite in a form the audience can actually follow up — the study, the mission, the institution — with the date, at the point of the claim. Name where a figure is an estimate or a range, and say plainly when sources disagree.
50 - **Native Editability**: Prefer editable native charts for the few figures the talk carries, so labels can be rewritten in plain language and built up in stages. Keep photographic evidence and instrument imagery at original fidelity; where a research figure is the evidence, present it with its labels translated alongside rather than silently redrawn. Producing publication figures is outside this style's scope.
51
52 ## V. Visual System Defaults
53
54 - **Preferred Visual Style**: paper-cut
55 - **Composition**: Build each page around one visual carrying one idea, with generous space and a clear entry point. Keep a consistent spatial arrangement for anything that recurs across the build so the audience recognizes rather than re-reads it. Let elements accumulate across pages in a stable position so the growing picture stays coherent.
56 - **Density**: Very light. A page should be understood in seconds by someone hearing this for the first time. Never place two new ideas on one page, and never fall back to bullet text where a visual is doing the explaining. Depth belongs in what the speaker says, not in what the page holds.
57 - **Decoration**: Warm and inviting, but every element carries meaning: layers separate concepts, arrows show direction, size shows magnitude. Avoid decorative scientific ornament — floating molecules, glowing brains, circuit-board backgrounds, generic laboratory imagery — that signals a topic without explaining anything.
58 - **Color Behavior**: Keep a light, approachable field and assign color a consistent explanatory job — this substance, this force, this stage — fixed for the whole talk so a returning element is recognized instantly. Keep categories distinguishable for color-vision differences and never encode a required distinction in color alone. Where a field has an established convention, follow it rather than inventing a new mapping. Any confirmed Brand or Deck identity replaces these tendencies.
59 - **Typography Character**: Use a warm, highly legible sans-serif at generous size, set in plain-language sentences rather than compressed notation. Keep technical terms visually marked when first introduced and consistently styled afterward, and keep mathematical or chemical notation properly rendered and large enough to read from the back of a hall. Exact families remain current-project or resolved identity decisions.
60
61 ## VI. Image & Icon Direction
62
63 - **Preferred Image Rendering**: vector-illustration
64 - **Image Usage**: Imagery carries the explanation here. Use authored illustration and diagram where a mechanism must be understood, and real photography where seeing the actual thing — the organism, the instrument, the site, the phenomenon — is what makes it real. Never use decorative science stock imagery in place of an explanatory visual.
65 - **Image Treatment**: Keep illustration clean, layered, and labeled in plain language, with a scale reference wherever size matters and any deliberate simplification noted. Present real evidence images honestly, with any false color or processing stated, and caption what the audience is looking at. Avoid full-bleed atmospheric treatment that crowds out labels, and synthetic text inside generated images.
66 - **Icon Treatment**: Use one coherent icon family at consistent weight to mark recurring entities or stages, with meaning fixed for the whole talk and always paired with a word on first use. Avoid icon grids substituting for explanation and mixed visual languages within one build.
67
68 ## VII. Review Focus
69 <!-- visual-review-trigger: explicit-user-only -->
70 > Apply this section only after the user explicitly activates visual review. It never triggers that stage.
71
72 - Each page carries one idea, understandable by someone meeting the topic for the first time.
73 - Every technical term appears with its plain-language meaning where it is first introduced.
74 - Each analogy is marked as an analogy and its breaking point is stated on the page.
75 - Visuals explain rather than decorate; no page relies on a diagram carried over from the previous one.
76 - Chart axes are explained before the data, labeled directly, and neither broken nor unexplained in scale.
77 - The color-to-concept mapping stays fixed across the talk and survives color-vision differences.
78 - Consensus, active debate, and speculation remain distinguishable in plain words.
79 - Text and labels are legible from the back of a hall, not only on a laptop.
80
80 lines MARKDOWN