VR Hand Interfaces — Replicating a CHI 2022 Interaction Technique on Quest 2
A replication of the CHI 2022 “Hand Interfaces” technique across 11 virtual objects, carried through a full migration from the deprecated Oculus Integration SDK to the Meta XR SDK.
At a glance
“Hand Interfaces” lets a user's own hands become the object they are manipulating — forming the shape of a tool instead of holding a virtual proxy. This project reproduces that technique on Quest 2 hardware and documents what it takes to get a published interaction result running on a current SDK.
The problem
Organizations buy extended reality for training, rehabilitation and public engagement, and then discover the content has a shorter life than the hardware. A school, a museum, a rehabilitation unit or a municipal training center commits to headsets and a set of experiences, and within two or three years the SDK those experiences were built against is deprecated.
- Content built on a retired SDK does not simply carry forward. It stops building, then stops running, and the program around it quietly ends.
- The cost lands where it hurts most: a public body that ran a procurement, spent the budget and trained its staff now owns unusable content on working hardware.
- Controllers add a second barrier. They have to be charged, paired, sanitized between users and physically operated — which excludes some older or less confident users outright and makes walk-up settings like a library or a classroom hard to run.
How this helps
This project is the migration, done and written down. A published CHI 2022 interaction technique was reproduced on current hardware by porting the entire interaction layer from the deprecated Oculus Integration SDK to the Meta XR SDK v83.0.1 — hand-tracking APIs, prefabs and pose data were all reorganized between the two — and the technique itself uses the user's own hands instead of controllers.
- A route for stranded content. The port shows what it actually takes to move an existing XR experience onto a supported SDK, which is the difference between keeping an investment and writing it off.
- Nothing to hand out or charge. Hand tracking removes the controller from the loop, lowering the barrier for users who struggle with them and removing the logistics that make shared installations impractical.
- Written to be reused. Eleven objects and a twelve-section report mean the method is transferable to another team's content, not a one-off demo.
The uncomfortable lesson is worth stating: most of the effort here went into migration, not interaction design. Any organization planning a multi-year XR program should budget for that maintenance up front, because the published technique and a working build are separated by exactly this kind of work.
Scope of the replication
- Implemented the technique across 11 distinct virtual objects, each with its own hand pose mapping.
- Targeted Meta Quest 2 with hand tracking rather than controllers.
- Written up as a 12-section LaTeX report covering method and results.
The SDK migration
The published technique predates Meta's current tooling, so reproducing it meant porting from the legacy Oculus Integration SDK to the Meta XR SDK v83.0.1. Hand-tracking APIs, prefabs, and pose data were all reorganized between the two, so this was closer to a rewrite of the interaction layer than a version bump.
Replication work tends to be judged on the interaction result, but most of the effort here went into the migration. That is the part worth knowing about: a technique published against a deprecated SDK does not simply carry forward, and the gap between the paper and a working build is where the real engineering sits.