List view
Goal: Evolve the Dynamic Context Core from a reactive system into a sophisticated, developer-guided partner. This milestone focuses on implementing advanced decision-making models while providing developers with explicit tools to influence and constrain the SAA's behavior, ensuring that autonomy always serves artistic intent. Philosophy: A truly symbiotic architecture cannot be a black box. While DCC v1 establishes the engine's awareness and basic adaptation, v2 focuses on building trust and partnership. The core idea is that the developer's intent is the most valuable piece of contextual information the engine can have. This milestone provides the tools for developers to express that intent—not through low-level tuning, but through high-level hints, priorities, and constraints. The engine's autonomy is thus framed not as absolute, but as a powerful capability operating within the clear boundaries and goals set by the creator. This transforms the SAA from a simple optimizer into an intelligent assistant that actively collaborates to realize the developer's vision.
No due date•0/2 issues closedGoal: Integrate a powerful scripting language to enable rapid development and iteration of game logic. This milestone goes beyond simple integration by treating the scripting runtime as a first-class, adaptive citizen of the SAA, ensuring that even user-defined logic is performance-aware and budget-controlled. Philosophy: While Rust provides the performance and safety for the core engine, game development productivity hinges on a flexible, high-level language for defining behaviors, UI interactions, and gameplay sequences. However, scripting is often a source of unpredictable performance costs. In Khora, we address this by making the scripting virtual machine (VM) itself an Intelligent Subsystem Agent (ISA). This unique approach allows the engine to monitor, budget, and even constrain the CPU resources consumed by game logic, preventing runaway scripts from compromising the user experience and making performance management a holistic process.
No due date•0/3 issues closedPhase for stabilization, final optimization, comprehensive documentation (including specific SAA concepts), and preparing the application (the editor based on the SAA engine) for potential distribution.
No due date•0/3 issues closedExplore the most advanced and experimental aspects of SAA, such as adaptive game data flow (AGDF), richer interfaces (semantic contracts), and the potential use of specialized hardware or AI.
No due date•0/6 issues closedImplement the core of SAA: resource negotiation. The DCC will dynamically allocate "budgets" (CPU time, memory) to ISAs based on global goals and agent requests.
No due date•0/5 issues closedGive ISAs the ability to operate according to different strategies and implement initial simple adaptation logic in the DCC, allowing the engine to change behavior automatically based on context.
No due date•0/9 issues closedBuild the first version of the SAA "brain" (the DCC), capable of aggregating the previously collected context information, and begin structuring subsystems as rudimentary Agents (ISAs).
No due date•0/10 issues closedEnhance the editor with key features (asset browser, gizmos), add basic networking capabilities, and introduce manual controls for system parameters as a prototype for future DCC control.
No due date•0/9 issues closedIntegrate basic Extended Reality (XR) features via OpenXR, ensuring that XR-specific information (tracking, latency) is also captured as contextual data for SAA.
No due date•0/7 issues closedCreate the basic graphical user interface for the editor, including a rendering viewport, implement an initial in-engine UI system, and create a dedicated panel for visualizing SAA contextual data.
No due date•0/11 issues closedEnhance rendering capabilities, add core simulation features (Physics, Animation, basic AI), and begin actively exploring alternative rendering strategies (SAA preparation), while improving performance visibility.
No due date•4/15 issues closed