Back to Super Productivity

Architecture Decision Records

ARCHITECTURE-DECISIONS.md

18.21.138.8 KB
Original Source

Architecture Decision Records

This document tracks significant architectural decisions and patterns in the Super Productivity codebase. When making changes that affect these patterns, reference this document and update it if needed.

It is also the index of accepted decisions: a decision recorded somewhere else — because it is long enough to stand alone, or because it is enforced as a contributor rule — must still be listed under Decisions Recorded Elsewhere.

Active Patterns & Decisions

1. dueDay/dueWithTime Mutual Exclusivity Pattern

Status: ✅ Active (since commit 400ca8c1, 2026-01-29)

Decision: The task.dueDay and task.dueWithTime fields are mutually exclusive in new data. When setting dueWithTime, dueDay must be cleared (set to undefined). When reading, dueWithTime takes priority over dueDay.

Rationale:

  • Prevents state inconsistency bugs where both fields had conflicting values
  • Single source of truth for task scheduling
  • Simpler state management

Implementation:

  • Writing: Clear dueDay when setting dueWithTime (in meta-reducers)
  • Reading: Check dueWithTime first; only check dueDay if dueWithTime is not set (in selectors)
  • Legacy Data: Old data with both fields works via priority pattern (no migration needed)

Key Files:

When to Update This Pattern:

  • Adding new date/time scheduling fields
  • Modifying task scheduling logic
  • Working with task selectors that check due dates

2. TODAY_TAG Virtual Tag Pattern

Status: ✅ Active (established pattern)

Decision: TODAY_TAG (ID: 'TODAY') is a virtual tag whose membership is determined by task.dueWithTime or task.dueDay, not by task.tagIds. The tag's taskIds field stores only the ordering of tasks, not membership.

Key Invariant: TODAY_TAG.id must NEVER be added to task.tagIds

Rationale:

  • Uniform move operations across all tags (virtual and regular)
  • Single source of truth for "today" membership (date fields, not tagIds)
  • Self-healing ordering (stale entries automatically filtered)
  • Natural integration with planner (which uses date fields)

Related: Uses the dueDay/dueWithTime mutual exclusivity pattern (Decision #1)

Key Files:

When to Update This Pattern:

  • Adding new virtual tags
  • Modifying tag membership logic
  • Working with today's task list

3. Sync Package Boundary Direction

Status: ✅ Active (since May 2026)

Decision: Operation-log sync code is split by dependency direction: src/app composes host-specific wiring, @sp/sync-providers owns bundled provider implementations, and @sp/sync-core owns framework-agnostic reusable sync primitives.

Rationale:

  • Keeps reusable sync algorithms independent of Angular, NgRx, app models, and provider implementations
  • Prevents provider IDs, app action/entity enums, validation schemas, UI, OAuth, and platform bridges from leaking into the core engine package
  • Gives boundary lint a clear rule: packages never import app code, and providers consume only public sync-core exports

Implementation:

  • ESLint rejects Angular, NgRx, app, shared-schema, sync-core deep imports, and dynamic imports inside package sources
  • @sp/sync-core has no runtime dependencies and owns vector-clock algorithms used by client/server compatibility paths
  • packages/shared-schema compatibility-re-exports generic vector-clock algorithms from @sp/sync-core; @sp/sync-core must not import @sp/shared-schema
  • @sp/sync-providers depends on public @sp/sync-core plus provider runtime helpers, while app factories inject credentials, platform bridges, validators, OAuth routing, and config

Documentation: docs/sync-and-op-log/package-boundaries.md

Key Files:

When to Update This Pattern:

  • Moving sync code between app and packages
  • Adding a package export or dependency
  • Adding a provider implementation or plugin-facing provider contract
  • Changing vector-clock ownership or shared-schema compatibility

4. Upload Conflict Safety via the lastSeq Row Lock Under RepeatableRead

Status: ✅ Active (since May 2026; batch upload engine removed August 2026)

Decision: SuperSync uploads derive conflict-safety from the shared user_sync_state.lastSeq row write that reserves server sequence numbers, not from PostgreSQL RepeatableRead snapshot isolation alone.

Note — batch upload engine deleted (2026-08, #9508): this decision was originally written for the batch upload engine (processOperationBatch, prefetchLatestEntityOpsForBatch, the SUPERSYNC_BATCH_UPLOAD flag). That engine was never enabled in production and was deleted rather than rolled out; the serial per-op path (processOperation) is the only upload engine. The invariant below is engine-neutral and applies unchanged to the serial path. The deleted batch code last lived at commit 924ddd7019. Re-open condition: the batch engine processed a 25-op upload in ~10 SQL statements vs ~127 for serial — resurrect it (from that commit, re-reviewed) only if per-upload latency or transaction lock-hold time becomes a measured production problem.

Rationale:

  • PostgreSQL RepeatableRead does not provide full serializable snapshot isolation
  • Two concurrent upload transactions can both pass conflict checks when they read the same pre-insert snapshot
  • Reserving sequence numbers through one user_sync_state.lastSeq row forces accepted writers for the same user to serialize on that row lock
  • A causal REPAIR snapshot must prove that its state includes the current server prefix; the same row serializes that base-cursor check with later writes
  • If two uploads race, the later writer blocks on the row and the transaction retry path handles the serialization failure rather than silently accepting conflicting operations
  • The serial path's post-allocation conflict re-check ("FIX 1.5", removed 2026-08; last lived at commit 07511ab45c) was dead code: under RepeatableRead both conflict checks read one snapshot fixed at the transaction's first statement, and the lastSeq increment raises a serialization failure (40001) against any committed concurrent upload before a re-check could run. Lowering the isolation level below REPEATABLE READ would require reinstating a post-allocation re-check.

Implementation:

  • An upsert ensures the user_sync_state row exists (lastSeq: 0); each accepted operation then reserves its sequence number with an atomic update({ lastSeq: { increment: 1 } }) on that row (operation-upload.service.ts)
  • The operation insert uses createMany(..., skipDuplicates: true): a lost duplicate-ID race surfaces as count === 0 and is handled in-transaction (sequence rolled back, op classified as DUPLICATE_OPERATION) rather than aborting the whole upload with a unique-constraint error; only a non-ID unique conflict aborts the transaction
  • REPAIR uploads persist repairBaseServerSeq on the operation row. The HTTP handler rejects an obviously stale base before quota cleanup, and the upload transaction repeats the check under SELECT ... FOR UPDATE before insertion
  • Regular uploads carrying lastKnownServerSeq use the same per-user row lock to reject an upload behind the latest SYNC_IMPORT or BACKUP_IMPORT before insertion. The durable replacement marker is reconciled lazily from retained operations for rows created before the marker existed.
  • The marker carries import semantics — it forces a stale client to download a replacement whose concurrent ops the client then drops. Only an import advances it. The lazy reconcile falls back to the newest causal REPAIR purely as a stand-in for an import row that pruning already deleted; a repair is not a fence in its own right, and the client replays concurrent work on top of one rather than dropping it
  • Markerless legacy repairs are compatibility records, not causal boundaries: they cannot drive download fast-forward, snapshot trust, history pruning, or server-generated restore points; snapshot replay across one fails closed
  • Removing or sharding the lastSeq write requires replacing this safety mechanism with an equivalent per-user serialization primitive

Documentation: packages/super-sync-server/docs/architecture.md, docs/sync-and-op-log/sync-architecture.html#transport

Key Files:

When to Update This Pattern:

  • Changing upload conflict detection
  • Changing server sequence assignment
  • Changing transaction isolation for upload operations
  • Changing repair base-cursor validation or full-state history pruning
  • Changing the state-replacement upload fence
  • Introducing multi-writer or multi-region upload processing

5. Project Completion: Decoupled Resolution over Atomic Multi-Entity Op

Status: ✅ Active (since 2026-06-06, branch feat/completing-projects-48eeb4)

Decision: "Complete project" is a plain single-entity PROJECT flag flip (completeProject, OpType.Update, mirroring archiveProject → sets isDone/doneOn/isArchived). The accompanying resolution of unfinished tasks ("move to Inbox" / "mark done") runs first, as the normal per-task actions (moveToOtherProject / updateTask isDone) dispatched in a loop with the Rule #6 bulk-dispatch flush — not bundled into a single atomic multi-entity op.

Rationale: An earlier iteration made completion one atomic Batch op (completeProject) that marked/moved tasks inside the project-shared meta-reducer. Because that op deliberately routed around the normal per-task actions, every system that observes those actions had to be re-taught about completeProject separately:

  • Conflict detection needed a whole new affectedEntities multi-entity-ref feature threaded through sync-core, the sync server (+ a Prisma migration), shared-schema and the op-log — ~1,565 LOC, of which completeProject was the only producer.
  • Native-reminder cancellation, issue two-way-sync, time-block sync and repeat-cfg effects each needed a dedicated completeProject listener to re-derive the task changes the atomic op skipped.

The atomic op's headline benefit — reversing the whole thing as one unit — was never realized: reopenProject only clears the project flags; it does not un-move or un-complete the resolved tasks. So the bundle paid a large cross-cutting cost for an undo guarantee it didn't provide. Decoupling makes the existing effects and per-entity conflict detection fire naturally and deletes ~1,750 LOC total (revert + decouple). Trade-off accepted: completion now emits N+1 ops (one per resolved task + the flag flip) instead of one, and there is a brief intermediate state — both fine for a rare, user-initiated action whose resolution is not atomically reversible anyway. One behavioral nuance vs. the old atomic op: when unfinished work is moved to Inbox, a task that was being actively tracked stays the current task (it was carried forward, not finished — consistent with Inbox's carry-forward intent); the mark-done path stops tracking the current task via the existing autoSetNextTask$ effect. The atomic op cleared the current task in both cases; the decoupled design intentionally keeps it for the carry-forward case.

Implementation:

  • Action/reducer: completeProject({ id, doneOn }) in project.actions.ts; on(completeProject) flag flip in project.reducer.ts (guards INBOX_PROJECT). reopenProject clears the flags only.
  • Service: ProjectService.complete(id, doneOn) dispatches the flag flip; moveTasksToInbox() / markTasksDone() loop the normal per-task actions + setTimeout(0) flush.
  • Flow: work-context-menu resolves unfinished work before calling complete().
  • Do NOT reintroduce a multi-entity completeProject op or affectedEntities for it without re-justifying the full downstream cost above. Prior atomic implementation is preserved in history at commit 0893a86162.

Key Files:

When to Update This Decision:

  • Adding a true bulk meta-reducer action for general use (revisit whether completion should adopt it)
  • Reworking how completion resolves unfinished tasks
  • Any proposal to make completion a single synced op again

6. Passkeys Stay Pending Until Email Verification

Status: ✅ Active (since July 2026)

Decision: A passkey submitted during account registration is stored as a PendingPasskeyRegistration tied to its exact email-verification token. It is promoted to the user's active Passkey set only when that token is consumed.

Rationale:

  • A WebAuthn registration ceremony proves possession of a credential, not ownership of the email address entered alongside it.
  • Storing a submitted credential directly on an unverified user lets an attacker pre-register a victim's address, then have the victim's later magic-link verification activate the attacker's passkey.
  • Keeping separate pending attempts prevents concurrent registrations from replacing or activating one another. The email owner chooses the credential by consuming the link produced by that same registration attempt.
  • Failed email delivery leaves the bounded, expiring pending attempt in place. Deleting the shared unverified user can race a concurrent registration and invalidate a link that was successfully delivered.

Implementation:

  • Passkey registration stores no active credential and creates one pending row per verification token.
  • Email verification atomically claims the unverified user, replaces active passkeys with the credential bound to that token, and deletes the user's remaining pending attempts.
  • Passkey verification tokens live only on pending registrations; user-row verification tokens belong to magic-link registrations. Consuming a user-row token verifies the email but removes untrusted active and pending passkeys.
  • The migration moves the latest legacy credential for each unverified user to the pending table and removes all active credentials from unverified users.
  • The resend cap bounds pending rows per unverified account; rows also expire with their verification tokens.

Key Files:

When to Update This Pattern:

  • Changing passkey enrollment or email-verification flows
  • Adding another credential type to registration
  • Changing verification-token persistence or cleanup

7. Versioned Delete-Wins Semantics for Project Deletion

Status: ✅ Active (since July 2026)

Decision: Project deletions created with schema v4 or newer carry an explicit projectDeleteWins marker and beat concurrent project updates. Historical, unmarked deletions keep timestamp-based LWW semantics.

This is a deliberate semantic trade-off: a concurrent project rename or field edit that is vector-clock CONCURRENT with a marked delete loses, regardless of which has the newer wall-clock timestamp. Deleting an entity another device is editing wins over the edit — the alternative (timestamp LWW) resurrects an empty project shell and silently loses its task subtree. The lost edit is only recoverable via local undo, not via sync.

Rationale:

  • deleteProject is one user intent whose reducer cascade removes the project, active tasks, notes, sections, repeat configuration, and related archive data. Reversing only the project entity after that operation loses data and violates replay determinism.
  • Capturing every cascaded entity in the delete payload or emitting restoration sidecars makes payload size scale with project size and still cannot restore every side effect safely.
  • Deletion is the only complete, deterministic result already represented by the operation. A concurrent rename or project-field edit must not partially undo it.
  • The schema-v4 barrier makes clients that do not understand this conflict policy stop before applying the operation (they block on the newer-schema gate rather than mis-resolving). The absence of the payload marker on historical deletions — never added by the no-op v3→v4 migration — is what preserves their timestamp-LWW semantics; the marker, not the version number, is the real discriminator. The classifier additionally requires the marked delete's plaintext entityId to match its authenticated payload projectId, so a tampered/replayed delete retargeted onto a live entity cannot win.

Implementation:

  • New deleteProject actions include projectDeleteWins: true; replacement delete operations preserve that payload.
  • The shared LWW planner accepts a host-supplied delete-wins classifier. A remote marked delete is applied regardless of timestamps. A local marked delete is replaced with one operation whose vector clock dominates both conflict sides.
  • SuperSync keeps its generic conflict protocol: if the first delete upload is rejected, the existing retry path uploads the causally dominant replacement. File-based providers use the same client planner and marker.
  • Do not add per-task/note restoration operations or project-sized snapshots to compensate a losing marked project delete.

Key Files:

When to Update This Pattern:

  • Changing the cascade performed by deleteProject
  • Adding another operation with delete-wins conflict semantics
  • Changing schema compatibility or LWW replacement behavior

8. Additive Data-Model Evolution over Schema Bumps

Status: ✅ Active (since August 2026)

Decision: Persisted and synced data evolves additively. Pick the change channel by what actually changed (table below). Do not raise CURRENT_SCHEMA_VERSION unless a change is both inexpressible as an additive or derived field and would be misapplied — not merely ignored — by older clients. This is the constructive counterpart to the bump policy (sync rule 10), which says when not to bump but not what to do instead.

What changedChannelPrecedent
Local storage layout (stores, indexes, derived meta)DB_VERSION ladder — local only, never transmitteddb-upgrade.ts v7 seeds the full-state-ops meta store
Shape of stored state (new field, legacy key, changed default)Read-time normalization in the loadAllData reducermigrateFocusModeConfig, migrateKeyboardConfig
Representation of an existing synced fieldDual field — new field wins, legacy re-derived from it on every writenormalizeStartOfNextDayConfig
Semantics of an operationPayload marker / envelope, inert on older clientsLwwUpdatePayload; the v4 projectDeleteWins marker

Rationale:

  • A bump fences only receivers that ship after it. Released v17.0.0–v18.14.0 clients apply ops up to schema 5 unmigrated and, at ≥ 6, block them while still advancing the cursor — permanently skipping them. A bump therefore never buys safety against the clients actually writing today's data.
  • It cannot be reverted once any op carries the new version, and it hard-blocks every lagging post-v18.14.0 client on a frozen cursor.
  • The legacy fleet does not age out on its own: there is no desktop auto-updater (the block in electron/start-app.ts is commented out) and the update banner's dismissal is persisted. Any policy gated on "wait for the old fleet to shrink" is a permanent no in disguise.
  • Additive fields are safe by construction here: typia uses createValidate (excess properties are neither rejected nor stripped) and LWW patch application goes through updateOne, a shallow merge that retains unknown keys. Renames and removals are the dangerous shape — an old client that wins a conflict re-emits the entity without the field, destroying it fleet-wide — and no bump prevents that, because old clients keep writing regardless.

Evaluation record (2026-08): raising CURRENT_SCHEMA_VERSION to 5 was considered and declined. Neither candidate motivation survived: the accumulated optional-field/runtime-default debt needs no migration (that pattern is the answer, per sync rule 11), and the typed RRULE recurrence model can ship as an additive field while the flat fields stay canonical and re-derived — see #9664, which also corrects that plan's inverted cross-version gate. A migration with no payload is pure cost.

Implementation: no new machinery — each channel above already exists and has a shipped precedent.

Documentation: Bump Policy §A.7.11, persisted-model-fields.md, AGENTS.md sync rules 10 and 11

Key Files:

When to Update This Pattern:

  • A change genuinely requires removing or renaming a synced field
  • CURRENT_SCHEMA_VERSION is raised (record what earned it)
  • A desktop auto-updater ships — it changes the fleet assumption this rests on

9. Calendar Writes Live in Plugins, Behind Per-Provider Opt-In

Status: ✅ Active (recorded 2026-08; describes the boundary that shipped 2026-03 in 3e2265fa57 / 020fd56504)

Decision: Super Productivity is never the authority for calendar state. It reads calendars to show the day's commitments, and it may write a mirror of a scheduled task back — but only through a plugin issue provider that opts into the timeBlock contract, and only when the user has enabled that provider's auto-time-blocking setting. Core code contains no calendar write path.

The boundary today:

SurfaceWrites?
Built-in iCal/CalDAV URL feeds (src/app/features/schedule/ical/)No — poll and parse only
Built-in issue providers (src/app/features/issue/providers/*)No — none implement timeBlock
Plugin providers implementing timeBlock (google-calendar-provider, caldav-calendar-provider)Yes, when isAutoTimeBlock is on

Rationale:

  • A time block is a projection, not a synced entity. TimeBlockSyncEffects pushes task state one way — schedule, reschedule, title, estimate, done, delete — into an event the app itself created. It never reconciles a user's edit of that event back into the task, and it never touches events the app did not create. That keeps the flow one-directional even though it writes, which is what avoids the sync loop a true bidirectional design has to solve.
  • Off by default, per provider. isAutoTimeBlock is an unchecked box on the provider's config form. Writing into someone's calendar is not something to infer from an integration merely being connected (manifesto: opt-in, quiet by default).
  • Plugins are the right home. Every write path needs OAuth, per-vendor event shapes, and vendor-specific throttling. Keeping that in packages/plugin-dev/ behind the timeBlock contract means core carries no vendor API surface, and a broken provider degrades to read-only rather than breaking the app.
  • What is still excluded. No reconciliation of external event edits into task state, no adoption of pre-existing calendar events, and no per-occurrence recurring-event editing (RECURRENCE-ID/EXDATE, #8148). These are the parts that would require answering conflict resolution between vector clocks and ETags, and they remain unbuilt — see #5001 for the open bidirectional request.

Implementation:

When to Update This Pattern:

  • A core (non-plugin) calendar write path is proposed — that crosses the boundary this record draws
  • Reconciling external event edits back into task state is proposed (#5001) — that needs the ETag-vs-vector-clock conflict story written down first
  • Per-occurrence recurring edits land (#8148)

10. Vector Clocks over Server-Side Entity Versioning

Status: ✅ Active (recorded 2026-08; the alternative design is git show 07511ab45c:docs/long-term-plans/server-side-entity-versioning.md)

Decision: Conflict detection stays on vector clocks, pruned to MAX_VECTOR_CLOCK_SIZE = 20. Server-side per-entity version counters (optimistic concurrency control, the shape every centralized API uses) were designed in full and are not being built. The design was never rejected on its merits by a maintainer decision — it is recorded here as declined-by-default, because nothing has yet justified its cost.

Rationale:

  • The problem it solves is not observed. Pruning only discards causal information once 21+ distinct client IDs have touched a clock. For a personal deep-work tool that is not a realistic fleet, and the one edge case that did bite — an import client mispruning against its own ops — is already handled by a same-client check.
  • It would make the server the source of truth, not just the referee. The SuperSync server already detects conflicts (detectConflict in packages/super-sync-server/src/sync/conflict.ts), but it does so by comparing clocks the clients authored — the causal history stays client-owned. Entity versioning moves that authority into the server. File-based providers (WebDAV, Dropbox, local file) have no server to run it, so the vector-clock path must survive regardless, and we would maintain two conflict systems instead of one.
  • The migration is the expensive half. It needs a new server table, a wire protocol change, a backfill for every existing entity, and a mixed-fleet window where old (clock-only) and new (version-carrying) clients edit the same entity. The design did address this — each step was independently deployable and backward compatible — but "correct on paper, across eight steps, in the subsystem where mistakes silently destroy user data" is precisely the cost being weighed, and there is no failure it currently buys us out of.
  • Encryption boundary. The design assumed entity versions are non-sensitive and would ride outside E2EE. Plausible, but it adds another plaintext channel to reason about and was never re-derived against the current threat model — see supersync-encryption-architecture.md.

Implementation: unchanged — see docs/sync-and-op-log/vector-clocks.md. The server prunes after conflict detection, before storage.

When to Update This Pattern:

  • Real fleets are observed exceeding ~20 distinct client IDs per user, or pruning is traced to an actual user-visible conflict (rule: start from a reproducible problem)
  • SuperSync becomes the only supported backend, removing the "file providers need the clock path regardless" constraint

Decisions Recorded Elsewhere

These carry the same authority as the numbered records above. They live outside this file because they are long enough to stand alone, or because they are enforced as contributor/agent rules that must be read before touching the subsystem. Keep this table complete — if you record a decision somewhere else, add a row here.

DecisionWhere it lives
SuperSync database encryption at rest — no project-managed volume encryption; the LUKS and PostgreSQL-TDE attempts are retired as OpenVZ-incompatibledocs/supersync-encryption-at-rest-decision.md
Schema-version bump policy — default to NOT bumping CURRENT_SCHEMA_VERSION; a bump never protects the released fleet and cannot be revertedoperation-log-architecture.md §A.7.11 Bump Policy, schema-version.ts, AGENTS.md sync rule 10
Required fields on persisted models — a new field on a persisted model is optional (?) plus a runtime default, never requireddocs/sync-and-op-log/persisted-model-fields.md, AGENTS.md sync rule 11
One user intent = one op — effects inject LOCAL_ACTIONS; a multi-entity change is a meta-reducer, not an effect fan-outdocs/sync-and-op-log/contributor-sync-model.md, AGENTS.md sync rules 1–3 and 6
src/app layer boundarycore/ and ui/ must not import features/; lint-enforced, with a shrink-only grandfathered listsrc/app/README.md, eslint.config.js (FEATURE_LAYER_FENCE)

How to Use This Document

When Making Architectural Changes

  1. Before implementing: Check if your change affects any active pattern
  2. During implementation: Follow the documented patterns
  3. After implementation: Update this document if you've:
    • Changed an existing pattern
    • Added a new architectural pattern
    • Made a decision that affects future development

When to Add a New Decision

Add a new decision record when:

  • The decision affects multiple files/modules
  • Future developers need to understand "why" not just "what"
  • The pattern needs to be followed consistently across the codebase
  • The decision prevents a specific class of bugs

When a Decision Changes

Do not rewrite a record's rationale in place when the answer itself is reversed. The reasoning history — why the old answer looked right, and what evidence flipped it — is the thing that stops the same idea being re-proposed a year later, and it is invisible in git blame. Instead:

  1. Set the old record's status to ❌ Superseded by #N and leave it where it is. Numbering must stay stable: ~30 code comments cite decisions by number.
  2. Strip the superseded record down to Decision + Rationale — its Implementation and Key Files go stale the moment the code is gone — and add one line stating what new evidence or cost made it wrong.
  3. Write the replacement as a new numbered decision whose rationale names the record it supersedes.

No record has been superseded yet, so there is no worked example. Decision #5 is the closest model for the content of step 2 — it records a rejected design, what it actually cost (~1,565 LOC of cross-cutting machinery for a single producer), why the headline benefit was never realized, and the commit (0893a86162) preserving the prior implementation. It is itself ✅ Active and does not demonstrate the status/replacement mechanics of steps 1 and 3.

This applies only when the answer changes. Fixing wording, adding a key file, or clarifying an existing decision is ordinary editing.

Decision Record Template

markdown
### N. [Pattern/Decision Name]

**Status**: ✅ Active | 🚧 Draft | ⚠️ Deprecated | ❌ Superseded by #N

**Decision**: [One-sentence summary of the decision]

**Rationale**:

- [Why was this decision made?]
- [What problems does it solve?]

**Implementation**:

- [How is it implemented?]
- [Key techniques or patterns used]

**Documentation**: [Link to detailed docs]

**Key Files**: [List of primary files implementing this pattern]

**When to Update This Pattern**: [Scenarios when someone should review/update this]

Why One File

This log deliberately does not use one-file-per-decision (docs/adr/NNNN-*.md):

  • The numbered records are one read for a contributor or an agent. Unlike the lint-enforced rules in AGENTS.md, a decision record's only teeth are being read — spreading them over 30 files means nobody reads all of them.
  • docs/ already separates plans, long-term-plans, research, sync-and-op-log and wiki. A further location makes decisions harder to find, not easier.

Note this is "one index, many locations", not "one file": Decisions Recorded Elsewhere deliberately sanctions authoritative decisions living in their own documents. What stays consolidated is the entry point, so that ~30 in-code citations (// See: ARCHITECTURE-DECISIONS.md Decision #2) resolve to one place.

Revisit when supersession chains actually accumulate, or when this file passes ~1000 lines. Then split by subsystem, not one file per decision, and keep the existing numbering so those citations stay valid.



Commit Reference

When committing changes related to these patterns, reference this document and the specific decision:

feat(tasks): implement feature X

Uses dueDay/dueWithTime mutual exclusivity pattern (ARCHITECTURE-DECISIONS.md #1)