docs(design): §09 still specifies 各阶段合同数漏斗 — the app stopped shipping a funnel in #48 and narrowed what it plots in #59

[claude]

Fixes #63

DESIGN.md §09 still specified the Legal Workbench stage widget as 各阶段合同数漏斗 — a funnel over every lifecycle stage. Two merged cards moved the app off that in two independent ways, and the design authority never caught up. This is a documentation-catches-up card: no app file changes.

What §09 said on 9b7538a, verbatim

Re-derived on the branch base rather than inherited: §09 spans lines 293–309 on 9b7538a (heading ## 09 分析 at 293, ## 10 种子数据 at 310). The target string was present verbatim on line 302, at the end of the 法务工作台 bullet:

- **法务工作台** — 待受理数 · 审查中 · 审查超 30 天（固定阈值，写进砖的标题）· 谈判停滞 · 审批吞吐（本月 vs 上月，用平台 `compareTo`）· 各阶段合同数漏斗

PR #40 (296fab8) had already rewritten the rest of this bullet — 超 SLA → 审查超 30 天（固定阈值…）, 平均周转 → 审批吞吐（…用平台 compareTo）— and appended the 各段时长 ruling paragraph. 漏斗 was the one stale item left in the bullet. This edit sits alongside that rewrite: the 各段时长 paragraph and every other row are byte-identical.

The change — two hunks, both inside §09

Hunk 1, line 302 — the bullet now names the mark and the coverage:

· 各阶段合同数（横向条形图，按合同数从多到少排；只画在办的六个阶段）

Hunk 2, new line 310 — a paragraph carrying the reasoning and, critically, the non-promise about stage order.

⚠️ Flagging hunk 2 explicitly so the seat can split it if it prefers. It is this widget's specification and nothing else, but it is an addition rather than a correction, so it is called out rather than slipped in. The case for keeping it: the card's items 2 and 3 cannot be discharged by hunk 1 alone. Item 2 (say what it plots, and on what criterion) does not fit in a list bullet; item 3 (⛔ do not re-promise an authored stage order) needs the negative recorded, or the next reader re-adds stageOrder — which is exactly the defect class PR #40 removed from this section by recording, not by deleting silently.

Every claim traced to src/dashboards/legal.dashboard.ts as merged

Claim in the new §09 text	Reading in the merged widget
横向条形图	type: 'horizontal-bar' (L347)
「各阶段合同数」 is this widget	id: 'stage_funnel' (L344); zh-CN bundle title 各阶段合同数 (src/translations/zh-CN/app.ts:202); en Pipeline by Stage
不是漏斗	type is not funnel; comment §「Why not a funnel (issue #48) — the mark was making a false claim」 (L278–302); description "not a conversion sequence" (L346)
漏斗断言逐级单调递减	"A funnel encodes MONOTONIC DECLINE: each stage narrower than the one before" (L280–281)
这本合同簿不递减，§10 分布 draft 10 · submitted 6 · in_review 12 · in_approval 8 · approved 4 · signing 6 · active 60	counted out of the running DB in the comment (L247–251); pinned by assertSpread in src/data/plan-contracts.ts; DESIGN.md §10 line 320 states the same spread
画成漏斗是在断言数据没说的事	"the mark ASSERTED something the numbers do not say — that contracts fall out at each stage" (L290–292)
收录判据 = 法务在这个阶段是否还有事要做	"The criterion that does discriminate … WHETHER LEGAL HAS WORK TO DO ON THE STAGE" (L224–225)
六个在办阶段 draft / submitted / in_review / in_approval / approved / signing 入图	filter: { status: { $in: ['draft','submitted','in_review','in_approval','approved','signing'] } } (L351)
active 与 rejected / expired / terminated / cancelled 四种终态都不入图	absent from that $in list (L351); comment L214–233; description "Contracts already in force are not plotted." (L346)
生效合同簿的规模归管理层看板（生效合同额 · 90 天内到期）	"§09 assigns the in-force sense to 管理层 and it is there — active_contract_value by currency, plus expiring_90_days" (L271–273); §09's own 管理层 bullet (line 303, unchanged)
排序按合同数从多到少 (sortBy: 'contract_count' / sortOrder: 'desc')	options: { sortBy: 'contract_count', sortOrder: 'desc' } (L352–355); "it lowers to order: { contract_count: 'desc' } on the dataset query" (L327–328)
⛔ 不承诺可编排的阶段顺序	options carries no stageOrder (L352–355)
console 17.4.0 只有漏斗分支读 stageOrder	"categoryOrder is consumed inside the chartType === 'funnel' branch and nowhere else" (L308–311)
条形图上静默丢弃	"this widget as a bar with stageOrder still authored came out Active · Approved · Draft · … i.e. alphabetically" (L311–313)
非 en 语言下连漏斗分支也因标签不匹配而丢	"Measured in a zh-CN console … the funnel lost stage order entirely … because the rows carry the API's English labels while the client's order map carries the bundle's Chinese ones and nothing matches" (L318–326)
objectstack-ai/objectstack#17344	reported upstream rather than patched (L316), per AGENTS.md 平台缺口

Gates — all four 0, on 8ecfb1c

Each redirected to its own file with $? read on the next statement; no pipe between the command and the exit code.

on 8ecfb1c — validate=0 lint=0 typecheck=0 lint:i18n-gate=0

✓ i18n gate
  LOCALES  : "en", "zh-CN" checked (required: en, zh-CN)
  REACH    : 12 surface classes walked — object=34 · field=273 · option=188 · section=35 ·
             view=37 · action=55 · app=2 · navigation=28 · dashboard=6 · widget=38 ·
             dataset=70 · page=2
  COVERAGE : 0 missing keys across 2 locale(s)

lint reports 21 warning(s), 5 suggestion(s) and validate 27 ⚠ — all pre-existing and none of them reachable from this diff, which changes no metadata (see the disjointness measurement below).

⛔ None of the four gates reads DESIGN.md — measured three ways, not asserted

1. Grep the four gate logs. No gate log mentions DESIGN.md, or any .md file at all:

grep -ni "DESIGN\.md|\.md\b" gate-{validate,lint,typecheck,i18n}.log  →  exit 1, 0 matching lines

2. tsc --noEmit --listFiles — what typecheck actually opens.

tsc exit=0
total files compiled: 305
of those, .md files:  0
lines mentioning DESIGN: 0

3. Ablation — corrupt the file and watch nothing notice. The decisive one. With the fix committed first (so the restore leg points at HEAD), DESIGN.md was overwritten with two lines of garbage, the mutation proven on disk, and all four gates re-run:

proof the mutation reached disk:
  lines now: 2   (HEAD version had: 303)
  marker 'DELIBERATELY CORRUPTED' present: 1
  on-disk hash c17cd06b…  ≠  HEAD blob 4bf1529b…

ABLATION exits: validate=0 lint=0 typecheck=0 i18n-gate=0

A file that has been reduced to two lines of non-markdown passes every gate unchanged. The gates verify that the metadata parses; they say nothing whatsoever about this diff. The only checks on this change are the reading table above and human review.

Restore verified by state, not by an exit code — git hash-object DESIGN.md = 4bf1529b73a88b366b8707f9a6357216d55887b8 = git rev-parse HEAD:DESIGN.md, git diff HEAD empty, git status clean, corruption marker absent, 303 lines back.

4. (bonus) Diff/walk disjointness. git diff --name-only 9b7538a..HEAD = DESIGN.md, and its intersection with the gates' walk root (objectstack.config.ts + src/**, per tsconfig.json include) is 0 lines. Every byte the gates read is identical to 9b7538a.

Edit verified on disk by before/after counts — never by an editor's exit code

spelling	before	after
各阶段合同数漏斗	1	0
横向条形图	0	2
在办	0	2
sortBy	0	1
stageOrder	0	1
17344	0	1

漏斗 still matches 1 line — deliberately. That line is the new paragraph, which uses the word only to deny it (不是漏斗 / 漏斗断言逐级单调递减 / 画成漏斗是在断言数据没说的事 / 只有漏斗分支读 stageOrder). It is no longer the widget's name anywhere.

Both edit anchors were asserted to match exactly once before replacement, so an exact-match edit that silently matched nothing could not have reported success.

Scope

• Touched: DESIGN.md §09 only. §09 now spans 293–311; the changed lines are 302 and 310–311, both inside it. §01–§04 (lines 24–205) are byte-identical — same sha256 before and after (9d1de298628d97616a705d6e4be904768a0c06b14591354e3b5939f8ed8abc67).
• No other §09 row touched, no renumbering, no changeset (this repo has no changeset gate).
• No browser run owed — the diff changes no runnable surface. No app file is touched, and the ablation above shows the running app cannot observe this file at all.

Acceptance notes — the grep for other descriptions of this widget

Grepped 漏斗 and funnel repo-wide and 各阶段 / stage_funnel / Pipeline by Stage across DESIGN.md.

§02 is clean for this widget. Its Insights row (line 80) reads 「阶段时间戳 → dataset → 看板：交付各阶段到达数与审批吞吐，各段时长等平台日期能力（§09）」 — that points at the contract_cycle_time dataset and the approval_throughput tile, not at the stage bar, and it says nothing about a funnel or about coverage. PR #40 already corrected that row. No stale pointer here, and §02 is inside the governed range regardless.

Both translation bundles are already correct — stage_funnel.title is Pipeline by Stage / 各阶段合同数, and the zh-CN description already says 「六个阶段…按数量从多到少排列，不是转化流程。已生效的合同不在此图内。」 The new §09 wording was aligned to these shipped strings rather than to the card's prose.

Two stale occurrences remain, both in app source, both left untouched on purpose. The card and the dispatch both say ⛔ do not touch the app; these are named here rather than silently left, so the seat can decide:

1. src/dashboards/legal.dashboard.ts:4-5 — the file header quotes §09's 法务工作台 list to identify which dashboard it implements, and that quote is stale twice over: it still says 平均周转（本月 vs 上月） (PR Record four maintainer rulings in DESIGN.md §03, §06 and §09 #40 changed §09 to 审批吞吐) and 各阶段合同数漏斗 (this PR). One line, mechanical.
2. src/datasets/cycle-time.dataset.ts:44-45 — "The status funnel on the legal workbench is the honest funnel; these are per-stage coverage counts." The status widget has not been a funnel since PR Draw 「Pipeline by Stage」 as a ranked bar, not a funnel #57, so the sentence's contrast no longer holds.

Neither is a runtime defect: both are comments, invisible to the gates and to the product. They are the same stale-pointer pattern the card warned about, pointing the other way — code quoting a document instead of a document describing code. Suggest one follow-up card covering both; not filed, so as not to pre-empt the seat's split.

🤖 Generated with Claude Code

https://claude.ai/code/session_01R3n3GGzobdegM4HUzah1iR

---

Generated by Claude Code