Unit 20 · Language valence: Swift under policy · g12b

Okay, gut reaction: **Massive undertaking. Brutal.** Swift for *everything* is a significant constraint, especially with the plugin requirement. It's doable, but demands extreme discipline and architectural foresight. Feels like a recipe for technical debt if not handled *perfectly*. **Top 3 Technical Risks:** 1. **Plugin Interoperability (Swift-Only):** Getting plugins written by external developers to reliably and securely integrate with a Swift-centric core, especially across desktop/mobile, is *hard*. Need a robust, well-defined, and *secure* plugin API. Swift's dynamic capabilities will be pushed to the limit. 2. **Async Concurrency & UI Responsiveness:** Unpredictable async APIs feeding into a 3D viewport and a responsive UI is a concurrency nightmare. Swift's concurrency model is relatively new. Requires meticulous error handling, cancellation strategies, and UI thread management to avoid freezes and deadlocks. 3. **Cross-Platform 3D Viewport Performance:** Achieving acceptable 3 (greedy generation)

Okay, Linux CI and scriptable plugins *really* change the game. 120fps on mid-range Android is a brutal target too. This is escalating the difficulty. **What Breaks First (and Why):** **The Plugin System.** Here's the breakdown: * **Why:** The combination of "scriptable by end-users" and "Swift-only" is a *massive* conflict. End-users want flexibility – they'll likely want Python, Lua, JavaScript, *something* other than Swift. Forcing them into Swift dramatically limits adoption and usability. The Linux CI adds another layer – how do you safely execute user-provided scripts in a containerized environment? Security becomes paramount and incredibly complex. * **The Problem:** You're essentially asking users to become Swift developers to extend your app. It's a non-starter for most. The security implications of allowing arbitrary Swift code execution are terrifying. **The Solution (Multi-Phased, with Trade-offs):** (greedy generation)