docs(blog): map long-horizon work by goal and process clarity

[huangruiteng]

Goal And Delivered Outcome

• Outcome basis: accepted request to expand the bilingual application-scenarios article with a goal/process quadrant diagram.
• Gap: the three overview cards present solving modes and continuing responsibility as peers, without explaining their relationship.
• Before → after: readers can distinguish goal clarity from process clarity, place complex tasks and open exploration, and understand that digital-worker responsibility spans the quadrants. A concrete performance-maintenance journey shows how work moves between them.
• Intended base: main.

Scope And Continuation

Complete within this documentation scope: two article pages and two standalone SVG diagrams. Both editions add bounded exploration acceptance, version-bound validation, wake/quiet/escalation rules, and guidance for when continuity justifies a control plane.

The original title, lead, byline and benchmark observations/data/citations are preserved. Benchmark evidence moves after the scenario explanations with one conceptual bridge. No runtime, capability, benchmark scoring or evidence-policy change; no experiments are launched.

Future-facing simplification: reuse the existing bounded diagram viewport and presentation script, remove obsolete overview-card CSS, and avoid introducing another shared component or interaction owner.

Validation

• Tested revision: f6c9b240262105b38917d22f23404695efae77f0 (source content unchanged from the validated pre-commit scope).
• Run state: finished.
• Input classes: public_fixture, synthetic.

Check kind	Result	Public-safe evidence / limitation
static	passed	node examples/blog-bilingual-index-smoke.mjs: four paired articles
static	passed	Public site npm run build: TypeScript, client build and SSR
static	passed	Public site npm run smoke:seo: three prerendered routes and fourteen canonical sitemap targets
static	passed	Baseline comparison: original hero and benchmark section preserved; ten TOC anchors match section order per locale
static	passed	Built article links/assets resolve with the existing exporter's research-route copy
static	passed	Public/private scan of all four changed files; git diff --check
manual	passed	Both locales at desktop 1280×900 and mobile 390×844: readable diagram pan, no page-width overflow, full-size diagram navigation
manual	passed	Existing synthetic CI states and presentation-mode toggle/Escape retain their behavior
unit / real_backend	not_applicable	Static article and SVG changes; no runtime or backend contract changes

Coverage: publication build, bilingual navigation, metadata, local assets and responsive reading are checked. Full frontstage exporter and production deployment were not run; deployment remains after maintainer review/merge. No required scope check failed. Exact final change-quality receipt recorded and verified; no safe-fix pass was needed.

Frontend / Visual Evidence

• UI impact: changed (article content, not the site chrome).
• Before: published Chinese article, baseline English source.
• After: bilingual vector diagrams below; article viewport preview approved by the owner before commit/push.
• States/viewports checked: both locales, desktop and narrow mobile; unchanged synthetic CI demonstration states.
• Source data: synthetic conceptual map, not an experimental scatterplot.
• Attention review: one diagram replaces three isolated cards. The continuous-responsibility band prevents a false third-quadrant classification. Narrow screens pan within the figure instead of shrinking labels or overflowing the page.

English diagram

Type / Area

• Documentation update
• Public docs or presentation surface
• Shared-authority RFC fixture impact: N/A; no control-plane contract changes.

Boundary Checklist

• No private state, credentials, raw traces, internal links or local paths in the diff or public PR artifacts.
• No duplicated benchmark work or new experimental results.
• Only the accepted bilingual article/diagram scope is staged.
• Visual presentation reviewed before publication; no claim of deployed readback.
• Every commit carries DCO sign-off.

Maintainer merge requested after review; not self-merged.

[cocolord]

评审提交：f6c9b240262105b38917d22f23404695efae77f0。这是 policy-11 whole-PR、exact-head 评审。我检查了全部 4 个文件，并按公开文章入口走完双语结构、SVG、窄屏/键盘边界、构建、SEO、原有 benchmark 内容和 CI 归因。结论是 APPROVE：这是一个收益说清楚、范围有界、没有把概念图升级成运行时权威的正向改动。

动机

原页面把“复杂任务、开放探索、持续交付”做成三个并列卡片，但读者仍要自己推断它们的关系。这个 PR 给出了具体 before/after：用“目标明确度 × 过程明确度”解释 01/02 如何求解，再把 03 明确为横跨象限的持续职责；随后用性能维护的完整旅程和 When does a long-horizon control plane help? 回答什么时候值得承担额外状态与维护成本。它改善的是每个读者都会经过的文章开头，不是为了增加一张装饰图。

收益也有明确边界：文章没有宣称所有复杂任务都需要 LoopX，而是要求比较可验证产出、回退/重复工作、人的管理注意力以及新增执行/维护成本。这个判断门槛让概念图能影响真实选择；与原来的并列卡片相比，我认为是明显正向的内容增量。

改动思路

实现没有引入新的前端框架或交互 owner。两份 HTML 继续使用现有 blog.css 的 architecture-viewport、现有 presentation.js 和原有 canonical route；两份独立 SVG 只承载本地化标签。英文、中文均按 map → complex → exploration → employee → movement → layers → fit → benchmark → evolution → sources 展开，先建立模型，再给场景、迁移和采用条件。

仓库复用是合理的：architecture-viewport 已拥有横向滚动、block link、响应式图片与 focus-visible；本 PR 只给象限图设置 640px 的文章内最小宽度，并删除已失去调用点的 card CSS。新象限被明确写成连续的解释维度，而非持久 enum、调度 selector 或权限来源；“数字员工”也被限定为持续职责，不因此获得花费、写仓库或合并权限。

具体改动

关键内容讲解

• 英文 section#map 把三张 peer cards 替换为带 title/desc、alt、figcaption、键盘焦点和原图链接的二维图；正文逐象限解释，并明确 01/02 是求解方式、03 是责任时间轴。
• 新增 section#movement，用“澄清指标 → 探索瓶颈 → 实现验证 → 持续跟进”说明工作会双向跨象限，交接需要目标修订、反例、产物/验收版本和继续/停止条件。
• 新增 section#fit，把是否采用长程控制面落到跨轮次解释成本、失败经验丢失、人的调度负担以及新增执行/维护成本，避免把难度本身当成采用理由。
• 中文页面保持相同十个 section id、顺序、轴含义、权限边界和成本表述；两份 SVG 使用同一 1040×920 几何结构和 LoopX 的黑白/蓝色视觉语言。
• 原 benchmark 段落、数据、限制与引用整体移动但未改写；base/head 抽取比较只多出一段“benchmark 多数目标明确、路线未定”的概念桥。

没有发现需要行内修复的 blocker。4 个 changed files 共 +294/-122，其中 108 行是双语 SVG；HTML 的大量 diff 来自 section 移动，实际净增量约每个 locale 32 行。对一个需要同时解决关系图、本地化、移动案例和采用门槛的公开文章改动，这个体量是相称的。

对主干的风险

主要风险是首屏附近的大图在窄屏不可读、造成整页横向溢出，或中英文导航/资产漂移。独立验证结果：blog-bilingual-index-smoke 在 base/head 均通过；exact head 的 public-site client/SSR build、SEO smoke、git diff --check、两份 SVG XML 校验均通过；构建后的英文/中文页面有相同十个 section，TOC 完整对应，文章和 SVG 本地 URL 返回 200。额外 mutation check 删除中文 #movement 链接后会被 oracle 拒绝，原 exact head 通过。DOM/CSS 检查确认滚动被限制在 architecture-viewport，region 可聚焦，图片有尺寸/alt，原图链接存在。

本机当前没有可连接的浏览器实例，因此我没有重复作者记录的 1280×900、390×844 像素级手测；这是本次 residual risk。它没有上升为 blocker，因为 PR 已记录 owner 在 commit/push 前批准 article viewport，并明确列出双语 desktop/mobile、页面无横向溢出、原图导航和 presentation state 的手测；同时独立构建、DOM/CSS 和 SVG 渲染没有发现矛盾。

GitHub 的 test-shard (2/3/4)、聚合 pytest 与 merge-gate 仍为红色。shard 3/4 的 generated-twin 与两个 prompt-upgrade-hook assertion 在 immutable base 3ec049e 和 exact head 上以相同签名失败；shard 2 的 scheduler case 在 CI 记录了两个 1 秒 quota timeout，而同一节点在 base/head 本地均通过。这个 PR 只改两份静态 HTML 和两份 SVG，未改 tests/Python/scheduler/prompt/quota 路径，所以这些红项不归因于本 PR，但 merge readiness 仍应由仓库 owner 单独恢复或处置。

语义与 CI 对齐

新“目标/过程明确度”只扩展文章本地解释词汇，未注册为共享 schema 或机器状态；它与现有 Goal、Acceptance、Evidence、Kernel/Capability/Provider 边界一致，并明确 guidance 不等于 authority。CI 红项不能证明本文正确，但其独立归因也不应要求这个 docs PR 修改无关运行时代码。

我的整体评价

结论是 APPROVE。long-horizon 与 user experience 两个维度都改善：读者更早理解三类工作的关系，也得到一个可用于判断“是否值得加控制面”的成本/收益路径。改动复用现有页面 owner，双语和视觉资产范围可控，没有新增运行时状态、权限或默认行为，也没有重复扩写 benchmark 结论。

后续仍可在 CI 恢复后补一份公开的实际 mobile/desktop 页面截图，但现有 exact-head 证据已足以支持本次内容结论。本次 review 不修改 PR，也不授权 merge；红色 required checks 仍是独立的合并门槛。

English verdict: APPROVE exact head f6c9b240262105b38917d22f23404695efae77f0. The PR replaces three peer cards with a clear goal/process model, keeps continuing responsibility separate from machine authority, adds a concrete movement and adoption-cost test, preserves the benchmark body, and passes the focused bilingual/build/SEO/asset checks. The remaining red CI assertions are reproducibly pre-existing or timing-related outside this static four-file diff; they still hold merge readiness but do not justify changes to this PR.