Architecture: What is the correct phasing strategy for the platform genera...

Documentation for Sessions

🏗️ Architecture: What is the correct phasing strategy for the platform genera...

Date: 2026-03-16
Time: 19:25:01 UTC
Session: gspace1-226.12
Agent: architecture
Confidence: 90%
Status: ✅ Documented

📋 Decision Context

Question:

What is the correct phasing strategy for the platform generation v2 overhaul?

Task: S226.2: Platform Generation v2 architecture design

🎯 Decision Made

Chosen:

5-phase incremental approach: Phase I-A (bug fixes, 1-2 sessions), Phase I-B (library consolidation, 2-3 sessions), Phase II (SDK + page-builder + composables, 3-4 sessions), Phase III (backend generation with spike, 2-3 sessions), Phase IV (Vitto orchestration, 3+ sessions, aspirational). Each phase has validation gate.

Options Considered:

  1. All 4 phases in one spec, optimistic timeline (initially proposed)
  2. 5-phase incremental with validation gates and spike requirements (chosen)
  3. Only fix bugs, defer architecture (too conservative)
  4. Full rewrite of generation pipeline (too risky)

💭 Reasoning

Initial proposal was 4 phases with Phase I at '1 session' — spec review correctly identified this as 2-3x optimistic given 75+ consumer files needing import migration. Split Phase I into I-A (bug fixes, independently deployable, low risk) and I-B (library consolidation, incremental with re-export bridges, medium risk). Phase III requires a spike before committing — NestJS DynamicModule loading at runtime is non-trivial (schema registration at module init, route conflict isolation, hot-reload mechanism). Phase IV explicitly marked aspirational — needs dedicated spec after I-III validated. Key constraints: JIT compatibility in preview (signal inputs may not work in blob iframe), ESM SDK interaction (composables using inject() break in blob context), FontAwesome dependency (preview lacks it). Rollback strategy: re-export bridges for consolidation, generic UniverseDataController as permanent backend fallback, incremental widget wrapper refactoring.

📁 Files Involved

🔗 Related

Auto-generated by MCP Smart Consolidation