Lucas Berger
197efa1bc6
- Removed stale Backlog entries 999.2/3/6/7/8/9 (already promoted into the active Phase 6 UX Polish) and deleted the leftover 999.2 directory. - Kept genuine backlog: 999.1/4/5/10/11/12/13. - Promoted STATE pending todo to new backlog phase 999.14 (Gitea CI: PR-to-main regression + Docker build/publish). - Archived stale kickoff-new-project todo to completed/.
1.4 KiB
1.4 KiB
title, date, priority
| title | date | priority |
|---|---|---|
| Kick off FamilySync with /gsd:new-project | 2026-06-02 | high |
Kick off FamilySync with /gsd:new-project
Run /gsd:new-project to formalize the architecture decisions into a PROJECT.md, REQUIREMENTS.md,
and ROADMAP.md with phase breakdown.
Architecture is verified and settled (see .planning/notes/familysync-architecture.md).
Pre-work before running
- (Optional) Answer RQ-003: Fastmail calendar API — JMAP vs CalDAV for app read/write
- Review
.planning/notes/familysync-architecture.md— feed this into the project questionnaire
Context to bring into the session
- Family hub modeled on Skylight. Owner on Fastmail/Android; other members Apple. Wife adoption is a hard UX constraint (one slick PWA).
- Calendar: hosted on Fastmail (NOT self-hosted). Custom app brokers read/write via JMAP/CalDAV token. Owner native in Fastmail app; Apple members via PWA.
- Lists + display: one self-hosted custom app — backend + MariaDB, aggregates calendars into a Skylight-style view, owns shared lists, serves React PWA (phones + wall-display kiosk).
- Auth: Authelia OIDC for the app.
- Infra: Unraid + Docker Compose + Pangolin/Newt + split-DNS. MariaDB + Redis.
- Dropped: Baikal, Vikunja, DAVx5 (see architecture note for why).
- Out of scope: email. Excluded: PostgreSQL, App Store/React Native, forking.
- Net: build ONE app on top of existing Fastmail.