SDF VS RGD
SDFormat can describe robots, models, sensors, physics and entire simulated worlds. RGD focuses on the semantic identity and embodied context of the robot itself across physical, operational, capability and lifecycle domains.
RESPONSIBILITY MATRIX
Simulation-oriented robot, model and world description.
Semantic embodiment and AI/runtime grounding context.
XML models, links, joints, sensors, physics, plugins and worlds.
Modular graph/profile across seven semantic domains.
First-class: models, scene, lights, physics and environment.
Not a world/simulator format.
Rich simulator-oriented declarations.
Can record embodiment evidence and constraints, not simulate them itself.
Model identity exists inside simulation description.
Canonical robot/profile identity and integrity are explicit.
May be represented indirectly through plugins/stack-specific semantics.
Explicit capability and action semantics belong to Agency.
Not the primary focus of the format.
Evolution domain represents wear, adaptation and continuity.
Simulation configuration is not real hardware authorization.
RGD context is likewise non-actuating.
OVERLAP
WHAT RGD ADDS
RGD DOES NOT REPLACE
USE BOTH WHEN
OPENRGD FLOW
PRIMARY SOURCES