2026-02-24T09-41-23_auto_memory/memories.db-wal, memory/memories.db, memory

memory/2026-02-24-svelte-migration-feasibility-analysis-for-compass-.md

@@ -0,0 +1,21 @@
+# 2026-02-24 Session Notes
+
+## Svelte Migration Feasibility Analysis for Compass Project
+
+Nicholai asked whether converting the Compass project from Next.js 15/React 19 to Svelte would yield performance improvements. The assistant provided a candid architectural assessment after analyzing the project structure.
+
+**Performance Benefits Identified**
+Svelte would provide real wins in three areas: Cloudflare Workers cold start times (smaller bundles, faster startup), complex interactive views like the Gantt chart and data tables (direct DOM updates vs React reconciliation), and overall bundle size reduction. Svelte 5 runes also offer a simpler reactivity model compared to React hooks and useEffect chains.
+
+**Migration Cost Analysis**
+The project contains 50+ components, 32 server action files, 19 hooks, and heavy dependencies on shadcn/ui, TipTap, and Recharts—all React-centric. A full migration would require:
+- Replacing all shadcn/ui components with shadcn-svelte (different API surface)
+- Removing Recharts entirely (no Svelte equivalent; would use LayerCake or Pancake)
+- Testing less battle-tested TipTap Svelte bindings
+- AI SDK v6 has weaker Svelte support than React
+- Converting 32 server actions to SvelteKit form actions (different mental model)
+
+Conservative estimate: months of rewriting for 15-20% better TTI and smaller bundles, with regression risk across every feature.
+
+**Recommendation**
+The ROI doesn't justify conversion now. More pragmatic immediate steps: profile actual Cloudflare Workers cold start times, investigate tree-shaking and route-level code splitting optimizations, and monitor OpenNext improvements. This approach yields 80% of the benefits at 5% of the cost.