Hi ORCA Hand maintainers,
URML (urml.dev) is a small, Apache-2.0 language for robot intent: an action becomes a typed primitive, validated against declared capabilities and a safety envelope, and only then dispatched. orca_core is the control core for a 17-DoF tendon-driven hand, which is well past what a single-DoF gripper abstraction can express, so it is a clean target for the part of URML built for dexterous hands. This is a request for comment.
Nothing here asks the project to adopt, host, or maintain anything.
The seam: URML has a dexterous gripper kind with a dexterity declaration (DoF, finger count, supported grasp types, in-hand support) and an optional grasp type on the grasp primitive. A grasp intent (a grasp type, a target, a force limit) is validated against a manifest declaring the hand's degrees of freedom and the grasp types it supports, and against a grasp-force envelope, before orca_core is asked to move joints. orca_core keeps the joint-space control, calibration, and tensioning; URML is the static check before it.
Two real questions: (1) is a typed grasp-intent layer useful above orca_core? (2) Does the ORCA hand's joint structure and grasp-type set map onto URML's dexterity declaration cleanly, or do the tendon-coupling details need something the manifest does not yet express?
Full write-up: https://github.com/URML-MARS/URML/blob/main/docs/rfcs/0625-orca-hand-outreach.md
Thanks for orca_core; an open 17-DoF hand with a clean control core is a great place to test a typed grasp declaration.
Ido Yahalomi (URML, greenvh@gmail.com)
AI-assisted prose, maintainer-reviewed before posting (see https://github.com/URML-MARS/URML/blob/main/VIBE.md). Human-only correspondence available on request.
Hi ORCA Hand maintainers,
URML (urml.dev) is a small, Apache-2.0 language for robot intent: an action becomes a typed primitive, validated against declared capabilities and a safety envelope, and only then dispatched. orca_core is the control core for a 17-DoF tendon-driven hand, which is well past what a single-DoF gripper abstraction can express, so it is a clean target for the part of URML built for dexterous hands. This is a request for comment.
Nothing here asks the project to adopt, host, or maintain anything.
The seam: URML has a dexterous gripper kind with a dexterity declaration (DoF, finger count, supported grasp types, in-hand support) and an optional grasp type on the grasp primitive. A grasp intent (a grasp type, a target, a force limit) is validated against a manifest declaring the hand's degrees of freedom and the grasp types it supports, and against a grasp-force envelope, before orca_core is asked to move joints. orca_core keeps the joint-space control, calibration, and tensioning; URML is the static check before it.
Two real questions: (1) is a typed grasp-intent layer useful above orca_core? (2) Does the ORCA hand's joint structure and grasp-type set map onto URML's dexterity declaration cleanly, or do the tendon-coupling details need something the manifest does not yet express?
Full write-up: https://github.com/URML-MARS/URML/blob/main/docs/rfcs/0625-orca-hand-outreach.md
Thanks for orca_core; an open 17-DoF hand with a clean control core is a great place to test a typed grasp declaration.
Ido Yahalomi (URML, greenvh@gmail.com)
AI-assisted prose, maintainer-reviewed before posting (see https://github.com/URML-MARS/URML/blob/main/VIBE.md). Human-only correspondence available on request.