Skip to content

Commit bc67fa5

Browse files
committed
docs(brainstorms): add post-checkpoint foundation direction
Record the 2026-04-21 source-of-truth checkpoint after sqlite plus local-ANN integration landed on main, and repoint current dashboard anchors to the superseding direction document.
1 parent 31845aa commit bc67fa5

4 files changed

Lines changed: 141 additions & 10 deletions

File tree

Lines changed: 128 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,128 @@
1+
---
2+
date: 2026-04-21
3+
topic: mainline-foundation-integration-and-post-checkpoint-direction
4+
---
5+
6+
# Mainline Foundation Integration and Post-Checkpoint Direction Requirements
7+
8+
## Problem Frame
9+
10+
`main` no longer needs another "did backend independence land?" checkpoint.
11+
That question is already answered by code and verification.
12+
13+
Current drift is elsewhere:
14+
15+
- the 2026-04-20 checkpoint document still describes foundation promotion as `in_progress` / `no-go`,
16+
- top-level EN/ZH dashboard sections were corrected, but the checkpoint source-of-truth chain still needs a superseding record,
17+
- the next lane question is no longer "can we land bounded graph/vector independence?" but "is the current local sqlite + ANN baseline enough, or is a heavier backend justified?"
18+
19+
This document supersedes the 2026-04-20 checkpoint as the current source-of-truth item for post-M8 direction.
20+
21+
## Verified Mainline Reality (2026-04-21)
22+
23+
- `main` already carries the full bounded notification-threshold governance closure through `M7.29`.
24+
- `main` already carries the bounded foundation integration closure that earlier checkpoint documents treated as future work:
25+
- embedded sqlite graph persistence,
26+
- dedicated query backend boundary,
27+
- deterministic local embedding generation,
28+
- ANN-style bucket pruning,
29+
- readiness route/report/runtime parity,
30+
- promotion criteria matrix and derived ids.
31+
- current foundation status on `main` is:
32+
- `status = integrated`
33+
- `decision = go`
34+
- `promotionCriteriaPassed = 7/7`
35+
- `promotionBlockers = []`
36+
- `graphBackendIndependent = true`
37+
- `vectorAdapterIndependent = true`
38+
- current bounded backend semantics on `main` are:
39+
- `storeType = sqlite`
40+
- `graphBackendSignalKind = embedded_graphdb`
41+
- `queryBackendScoreSignals = [keyword_matches, title_match_bonus, vector_ann_similarity_bonus, relation_bonus]`
42+
- `vectorAdapterSignalKind = embedding_ann`
43+
- query/runtime traces surface `mode = local_ann`, `status = independent`
44+
- local verification already confirms that this is not doc-only drift:
45+
- `npm run verify:foundation:readiness`
46+
- `npm run test:migration`
47+
- `npm run test:agent-workspace:contracts`
48+
- `npm run docs:diataxis:check`
49+
- `npm run docs:site:build`
50+
- remote CI for the latest `main` push should still be watched normally, but failures must be interpreted against this integrated baseline rather than against older `no-go` assumptions.
51+
52+
## Deep Comparison: Prior Direction vs Current Code
53+
54+
| Prior source | Earlier direction | Current code reality | Judgment | Consequence |
55+
|---|---|---|---|---|
56+
| `docs/brainstorms/2026-04-20-mainline-checkpoint-ci-hardening-and-post-m8-direction-requirements.md` | checkpoint closure while foundation still remains `in_progress` / `no-go` | sqlite graph persistence and bounded local ANN retrieval are already integrated on `main` | now stale as current source | preserve as historical checkpoint, but stop using it as current truth |
57+
| `docs/diataxis/en/explanation/development-progress-dashboard.md` + ZH counterpart | top-level summary corrected to integrated/go | current-facing summary now matches code reality | correct | treat dashboard top sections as current product-facing truth |
58+
| `docs/diataxis/*/explanation/foundation-reentry-readiness-checklist.md` | anti-overclaim contract for pre-integration state | checklist now correctly records integrated/go bounded readiness | correct | keep using it as the operator contract page |
59+
| earlier post-M8 direction framing | backend independence should be the next execution target by default | bounded backend independence already landed locally; the remaining question is whether external backends are warranted | premise changed | next plan must evaluate sufficiency/ROI, not replay already-landed work |
60+
61+
## Critical Findings
62+
63+
1. Treating backend independence as still "not landed" is now factually wrong for the bounded local-first contract on `main`.
64+
65+
2. The architecture question changed shape.
66+
The relevant decision is no longer whether `main` can escape file snapshots and lexical prefiltering.
67+
It already did.
68+
The relevant decision is whether local sqlite + deterministic ANN is sufficient for current product goals.
69+
70+
3. External graph/vector backends are no longer default next steps.
71+
They now require explicit justification such as:
72+
- corpus size beyond comfortable local indexing bounds,
73+
- multi-process or multi-user write coordination,
74+
- remote/shared retrieval requirements,
75+
- latency/SLA targets local baseline cannot hit,
76+
- observability or recovery needs local embedded storage cannot satisfy.
77+
78+
4. Markdown-reader and other product-surface work should not be blocked by a fictional "backend independence still missing" narrative.
79+
They should only remain secondary if higher-value user/runtime work is chosen deliberately.
80+
81+
## Requirements
82+
83+
**Truth Refresh**
84+
- R1. Add a fresh 2026-04-21 direction record instead of silently rewriting the 2026-04-20 checkpoint document.
85+
- R2. EN/ZH dashboard execution anchors must point to this 2026-04-21 document first.
86+
- R3. EN/ZH top-level delivery-focus text must describe post-checkpoint direction and CI interpretation, not pre-integration backend uncertainty.
87+
- R4. No current-facing summary section may regress to `file-backed`, `prefilter_only`, or `in_progress` / `no-go` wording for the present repository baseline.
88+
89+
**Next-Lane Governance**
90+
- R5. The next planning round must answer one concrete question: is the current local backend baseline enough for the next product phase?
91+
- R6. If the answer is "no", the plan must state which deficiency justifies a heavier backend and why sqlite + local ANN is insufficient.
92+
- R7. If the answer is "yes", backend work should pause and the next user-facing lane should be chosen explicitly between markdown-reader and broader agent/product improvements.
93+
94+
**CI Interpretation**
95+
- R8. Remote CI failures, if any, must be triaged against the integrated baseline now on `main`, not against older readiness assumptions.
96+
- R9. Long-running or queued jobs must not be misreported as architecture regressions without concrete failing evidence.
97+
98+
## Success Criteria
99+
100+
- a new 2026-04-21 checkpoint/direction document exists and becomes the top current item in `docs/brainstorms/index.md`,
101+
- EN/ZH progress dashboards point to it from their execution anchors and current-priority framing,
102+
- docs verification passes locally,
103+
- the work is pushed to `main`,
104+
- the worktree remains clean after push.
105+
106+
## Scope Boundaries
107+
108+
- This document does not reopen notification-threshold expansion work.
109+
- This document does not claim an external graph or vector service is required.
110+
- This document does not start markdown-reader execution by itself.
111+
- This document does not rewrite historical archive sections that are clearly chronological.
112+
113+
## Key Decisions
114+
115+
- **Preserve history by superseding, not mutating**: the 2026-04-20 checkpoint stays as an auditable snapshot of an earlier assumption set.
116+
- **Treat local sqlite + ANN as a real baseline, not a placeholder**: code and tests now support that claim directly.
117+
- **Raise the bar for heavier backends**: future graph/vector escalation must be justified by measurable need, not by roadmap inertia.
118+
- **Keep next-lane selection honest**: user-facing work should compete against backend escalation on product value, not lose to outdated architecture assumptions.
119+
120+
## Outstanding Questions
121+
122+
### Deferred to Planning
123+
- Which measurable thresholds would justify moving beyond local sqlite + deterministic ANN?
124+
- If the local backend baseline is sufficient, which next lane has higher product value: markdown-reader improvements or broader agent/product-surface work?
125+
126+
## Next Steps
127+
128+
-> /ce:plan for `local backend baseline sufficiency vs external backend escalation`, then choose the next user-facing lane explicitly.

docs/brainstorms/index.md

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -4,6 +4,7 @@ This page tracks active and archived brainstorming outputs used before implement
44

55
## Current
66

7+
- [2026-04-21 Mainline Foundation Integration and Post-Checkpoint Direction Requirements](2026-04-21-mainline-foundation-integration-and-post-checkpoint-direction-requirements.md)
78
- [2026-04-20 Mainline Checkpoint: CI Hardening and Post-M8 Direction Requirements](2026-04-20-mainline-checkpoint-ci-hardening-and-post-m8-direction-requirements.md)
89
- [2026-04-19 Mainline Progress Reconciliation and Post-M7 Direction Requirements](2026-04-19-mainline-progress-reconciliation-and-post-m7-direction-requirements.md)
910
- [2026-04-16 Mainline CI Stabilization and M7 Direction Requirements](2026-04-16-mainline-ci-stabilization-and-m7-direction-requirements.md)

docs/diataxis/en/explanation/development-progress-dashboard.md

Lines changed: 6 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -6,7 +6,7 @@ It explicitly separates branch-level implementation status from mainline integra
66
## Scope
77

88
- Focus area: local-first knowledge mastery platform (ingest, retrieval, learning path, tutor, memory, governance).
9-
- Time window: `v1.7.0` through the 2026-04-20 checkpoint and 2026-04-21 UTC CI review cycle.
9+
- Time window: `v1.7.0` through the 2026-04-21 foundation-integration checkpoint and 2026-04-21 UTC CI review cycle.
1010
- Evidence rule: every progress claim must map to:
1111
- contract layer (`src/learning/api.ts`, `src/learning/types.ts`),
1212
- route/runtime layer (`src/server.ts`),
@@ -15,11 +15,11 @@ It explicitly separates branch-level implementation status from mainline integra
1515
## Current Delivery Focus
1616

1717
Current top priority is not feature-count expansion.
18-
Current top priority is **checkpoint closure**:
18+
Current top priority is **post-checkpoint direction closure**:
1919

20-
- truth-refresh the EN/ZH progress narrative against the real `main` state after `M8.18`,
21-
- harden the Windows restore-cache migration path so `main` CI does not depend on a timing-sensitive duplicate-request assumption,
22-
- keep the next lane choice disciplined instead of drifting into more inspectability-only micro-slices.
20+
- keep EN/ZH current-summary narrative aligned with the integrated sqlite + local-ANN baseline on `main`,
21+
- interpret remote CI against current code truth instead of older `no-go` assumptions,
22+
- force the next lane choice to answer whether the current local backend baseline is sufficient before escalating to heavier backend work.
2323

2424
## Current Mainline Snapshot (2026-04-21 M8.18)
2525

@@ -37,6 +37,7 @@ Current top priority is **checkpoint closure**:
3737

3838
Execution anchor:
3939

40+
- [Mainline Foundation Integration and Post-Checkpoint Direction Requirements (2026-04-21)](../../../brainstorms/2026-04-21-mainline-foundation-integration-and-post-checkpoint-direction-requirements.md)
4041
- [Mainline Checkpoint: CI Hardening and Post-M8 Direction Requirements (2026-04-20)](../../../brainstorms/2026-04-20-mainline-checkpoint-ci-hardening-and-post-m8-direction-requirements.md)
4142
- [Mainline Progress Reconciliation and Post-M7 Direction Requirements (2026-04-19)](../../../brainstorms/2026-04-19-mainline-progress-reconciliation-and-post-m7-direction-requirements.md)
4243
- [Mainline CI Stabilization and M7 Direction Requirements (2026-04-16)](../../../brainstorms/2026-04-16-mainline-ci-stabilization-and-m7-direction-requirements.md)

docs/diataxis/zh/explanation/development-progress-dashboard.md

Lines changed: 6 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -6,7 +6,7 @@
66
## 范围
77

88
- 聚焦对象:本地优先学习平台(摄入、检索、学习路径、导师、记忆、治理)。
9-
- 时间窗口:`v1.7.0` 至 2026-04-20 checkpoint 与 2026-04-21 UTC CI 复盘周期。
9+
- 时间窗口:`v1.7.0` 至 2026-04-21 foundation-integration checkpoint 与 2026-04-21 UTC CI 复盘周期。
1010
- 证据规则:每条进展结论都需映射到:
1111
- 契约层(`src/learning/api.ts``src/learning/types.ts`),
1212
- 路由/运行时层(`src/server.ts`),
@@ -15,11 +15,11 @@
1515
## 当前交付重心
1616

1717
当前不是继续堆功能数量。
18-
当前重心是 **checkpoint 收口**
18+
当前重心是 **post-checkpoint 方向收口**
1919

20-
- `M8.18` 后的真实主线状态重写 EN/ZH 进度叙事
21-
- 加固 Windows 下的 restore-cache migration 路径,避免 `main` CI 继续依赖时间敏感的重复请求假设
22-
- 在进入下一条主线前,先停止 inspectability-only 微增量惯性
20+
- 保持 EN/ZH 当前摘要叙事与 `main` 上已集成的 sqlite + 本地 ANN 基线一致
21+
- 远端 CI 解释必须基于当前代码事实,而不是基于旧的 `no-go` 假设
22+
- 在继续升级更重 backend 之前,先明确判断“当前本地后端基线是否已足够”
2323

2424
## 当前主线快照(2026-04-21 M8.18)
2525

@@ -37,6 +37,7 @@
3737

3838
执行锚点:
3939

40+
- [2026-04-21 主线 foundation integration 与 post-checkpoint 方向要求](../../../brainstorms/2026-04-21-mainline-foundation-integration-and-post-checkpoint-direction-requirements.md)
4041
- [2026-04-20 主线 checkpoint:CI 加固与 Post-M8 后续方向要求](../../../brainstorms/2026-04-20-mainline-checkpoint-ci-hardening-and-post-m8-direction-requirements.md)
4142
- [2026-04-19 主线进度重校准与 Post-M7 后续方向要求](../../../brainstorms/2026-04-19-mainline-progress-reconciliation-and-post-m7-direction-requirements.md)
4243
- [2026-04-16 主线 CI 稳定化与 M7 方向要求](../../../brainstorms/2026-04-16-mainline-ci-stabilization-and-m7-direction-requirements.md)

0 commit comments

Comments
 (0)