3.3 KiB
3.3 KiB
TFM aInventory — V2 Requirements
V1 Foundation (Validated & Stable)
All v1 features are locked. These are proven valuable:
- ✓ Item inventory with barcode/part number/quantity tracking
- ✓ Offline QR/barcode scanning
- ✓ AI-powered label extraction from photos
- ✓ LDAP authentication + credential caching for offline work
- ✓ Admin dashboard (users, categories, config, audit logs)
- ✓ PWA with offline sync (IndexedDB + UUID idempotency)
- ✓ Image adjustment (rotation only) in onboarding flow
- ✓ Audit logging (immutable, never deleted)
V2 Scope — Phase 4 & 5 (TBD)
Must-Have (Table Stakes)
- Mobile UX Polish: Responsive gesture handling, small-screen affordance fixes, touch-friendly controls
- Field User Research: Validate assumptions from 1+ deployed locations; capture pain points
- Documentation: Config management guide, deployment runbook, AI provider setup
Should-Have (High Value)
- Batch Operations UI: Multi-item selection + bulk actions (stock adjustment, deletion, relocation)
- Search & Filtering: Find items by name/PN/barcode; filter by category or location
- Export/Reports: CSV export of inventory state, audit trails for compliance
- Box Label Printing: Standardized thermal label layout (improve on current SVG generation)
- Deployment Docs: Runbook for single-instance setup (Docker, LDAP integration, offline mode)
Nice-to-Have (Deferred to V3)
- Advanced analytics (inventory trends, turnover rates)
- Multi-warehouse support (federation, inter-location transfers)
- Custom field schemas (Item type extension)
- Real-time sync (WebSocket updates for concurrent admin sessions)
- Localization (multi-language UI)
Success Criteria
For V2 Launch
- Mobile UX tested with 3+ field users; zero critical usability blockers
- Documentation complete (setup, deployment, config)
- Batch operations reduce bulk work by 50%+ compared to v1
For Ongoing Stability
- Zero data loss incidents in offline sync (100% UUID idempotency)
- Audit logs complete (all CRUD ops logged, deletions traced)
- Performance: Scan-to-save < 2 seconds; bulk sync < 30 seconds
Out of Scope (Rationale)
- Cropping: Rotation covers 90% of rotated/skewed image issues; add if feedback demands it
- Multi-tenancy: Single-instance per organization; federation is v3+
- Real-time collab: Async batch workflow matches field ops; collab adds complexity without value
- Custom field schemas: Predefined Item/Category structure has proven sufficient in deployments
Assumptions
- Field deployments are active: At least 1-2 live sites using v1.14.6
- LDAP available: Enterprise directory integration is mandatory (no guest mode)
- SQLite sufficient: Single-instance, no sharding/replication needed for v2
- AI providers stable: Gemini + Claude APIs continue to be available and affordable
- Offline sync works: UUID idempotency prevents data loss (validated in v1)
Acceptance Criteria Template
For each requirement, specify:
- What: Feature description (1-2 sentences)
- Why: Business/user value
- How: Technical approach (can be refined during phase planning)
- Done: Success metric (tests, UX validation, performance target)
Last updated: 2026-04-22 during reset