diff --git a/02-DECISIONS/0290-a-decision-record-lives-whole-in-the-knowledge-base-made-of-clauses.md b/02-DECISIONS/0290-a-decision-record-lives-whole-in-the-knowledge-base-made-of-clauses.md index 5a82cf12..11393a8b 100644 --- a/02-DECISIONS/0290-a-decision-record-lives-whole-in-the-knowledge-base-made-of-clauses.md +++ b/02-DECISIONS/0290-a-decision-record-lives-whole-in-the-knowledge-base-made-of-clauses.md @@ -15,6 +15,10 @@ extends: 02-DECISIONS/0153-the-record-is-read-by-a-module-and-the-console-lists- > and asked for this record to be updated in place rather than superseded. Nothing had been built on the earlier > version. It is in this repository's history at commit `08b1077b`. +> **Updated — 2026-10-11, on the operator's word.** Rule 9 left open whether this repository keeps a generated +> public copy of accepted decision records and research efforts (issue 435). The operator answered: no copy is needed. Rule 9 and the consequences +> say so. + ## Context [Research 045](../01-RESEARCH/045-a-knowledge-base-the-mesh-writes/00-overview.md) measured, on 2026-10-11, how an @@ -139,9 +143,9 @@ rule 8 already decided that the issue tracker hands out decision and research nu in force. 9. **This repository stops holding decision records and research efforts** once the import is checked against it, record by record: one pull request removes them, rewrites the mentions of their paths to "ADR NNNN" and "research - NNN", and changes playbooks 01 and 02, tracked by issue 426. Whether this repository then carries a generated, - read-only public copy is not decided here, tracked by issue 435. The records module stays the literal reader of - what this repository still holds. + NNN", and changes playbooks 01 and 02, tracked by issue 426. This repository keeps no copy of them, generated or + otherwise; the knowledge base is the only place a decision record or research effort is read. The records module + stays the literal reader of what this repository still holds. 10. **It is built in slices, each tracked by an issue.** The first is the module for decisions and research: import, clauses, answers, and proposing and reviewing a record, tracked by issue 425. The records leave this repository under issue 426. Designs, playbooks and glossary entries follow the same way, tracked by issue 428. Meeting notes, @@ -160,8 +164,8 @@ rule 8 already decided that the issue tracker hands out decision and research nu - Until the first slice takes writes, records are still written here as pull requests, with numbered clauses, and the import takes them in with the rest. - What got harder: one more module the mesh's own process depends on. Reading or writing a decision needs the mesh to - be reachable; a node that cannot reach the control-node says so rather than guessing. The public repository stops - showing decisions as they are made, until issue 435 decides on a public copy. + be reachable; a node that cannot reach the control-node says so rather than guessing. Once the records leave it + (issue 426), the public repository no longer shows the mesh's decisions or research. ## How it is checked @@ -180,4 +184,4 @@ rule 8 already decided that the issue tracker hands out decision and research nu [findings](../01-RESEARCH/045-a-knowledge-base-the-mesh-writes/01-findings.md). - [To-be 55 — The knowledge base](../03-DESIGN/01-to-be/55-the-knowledge-base.md). - hq pull request 305: every partial supersession declared in both records, and the `partial-supersession` check. -- Issues 424, 425, 426, 428, 429 and 435. +- Issues 424, 425, 426, 428, 429 and 435 (answered: no copy in this repository). diff --git a/02-DECISIONS/0291-an-issue-lives-whole-in-the-issue-tracker.md b/02-DECISIONS/0291-an-issue-lives-whole-in-the-issue-tracker.md index 9977da3e..6df05e9b 100644 --- a/02-DECISIONS/0291-an-issue-lives-whole-in-the-issue-tracker.md +++ b/02-DECISIONS/0291-an-issue-lives-whole-in-the-issue-tracker.md @@ -12,6 +12,11 @@ supersedes-in-part: # 291. An issue lives whole in the issue tracker +> **Updated — 2026-10-11, on the operator's word.** The context and consequences said decision records stay readable +> in this public repository, then that this was open (issue 435). The operator answered that no copy of the decision +> records is kept here once they leave it ([ADR 0290](0290-a-decision-record-lives-whole-in-the-knowledge-base-made-of-clauses.md)), +> and the sentences that said otherwise now say so. + ## Context [ADR 0275](0275-the-mesh-numbers-and-tracks-its-issues-in-a-module-of-its-own-and-hq-keeps-only-the-narrative-of-a-design-or-governance-issue.md) @@ -36,15 +41,14 @@ differently, and which one has a folder is an author's judgement at the moment i **What the record keeps.** A design or governance question that matters ends in a decision record, which the knowledge base holds whole, with its reasoning ([ADR 0290](0290-a-decision-record-lives-whole-in-the-knowledge-base-made-of-clauses.md)). What an issue's -content adds is the story of how a question was found, not the answer. Whether accepted decision records stay -readable in this public repository is issue 435. +content adds is the story of how a question was found, not the answer. Neither stays in this public repository. ### Against GENESIS - **The mesh owns the truth** ([mission](../00-META/mission.md)): an issue is in one place, written through one module. - **The mesh remembers** ([effect](../00-META/effect.md)): an issue's report, diagnosis and resolution stay, in the - tracker, kept as written with their history; the decision it led to stays here. + tracker, kept as written with their history; the decision it led to stays in the knowledge base. ## Considered Options @@ -81,8 +85,8 @@ readable in this public repository is issue 435. ## Consequences - One issue, one place: an agent reads an issue whole with `issue-tracker.get`. -- The public repository no longer shows how a design question was found. Whether it still shows the answer, the - decision record, is issue 435. +- The public repository no longer shows how a design question was found, nor, once ADR 0290's records leave it, the + decision record that answered it. - The tracker's values grow from a few hundred bytes to some kilobytes each; the largest measured fits 10 times over. - Reviewing an issue's diagnosis moves from a pull request to the tracker's history of that document. An issue's content was never a decision; what is decided still goes through a pull request. diff --git a/03-DESIGN/01-to-be/55-the-knowledge-base.md b/03-DESIGN/01-to-be/55-the-knowledge-base.md index 907485c1..85520f5f 100644 --- a/03-DESIGN/01-to-be/55-the-knowledge-base.md +++ b/03-DESIGN/01-to-be/55-the-knowledge-base.md @@ -160,8 +160,8 @@ field by field and section by section, and the result is written on issue 426. | 3 | designs, playbooks and glossary entries as kinds | `holds` answers for a design's decisions and a glossary word | issue 428 | | 4 | meeting notes, drawings and blog posts | not decided; needs its own decision | issue 429 | -Numbers from the issue tracker, which `reserve` needs, are the tracker's own slice, tracked by issue 424. Whether this -repository keeps a generated public copy of accepted records is issue 435. +Numbers from the issue tracker, which `reserve` needs, are the tracker's own slice, tracked by issue 424. This +repository keeps no copy of the records, generated or otherwise (issue 435, answered by the operator). ## 10. How it is checked