3D Selection Vault
An open catalogue of techniques for selecting objects in virtual and mixed reality.
Site: https://3dselectionvault.github.io/
Each technique is coded against a shared set of attributes and linked to the papers that introduce and evaluate it. The goal is a 3D similarity graph: describe a new technique and see which existing ones are closest, and so which to compare against in a paper.
Repository layout
| Path | What it holds |
|---|
data/techniques/ | One JSON file per technique |
data/publications/ | One JSON file per paper |
schema/attributes.json | The attribute scheme: attributes, allowed values, weights. Read by both the validator and the site |
schema/*.schema.json | JSON Schemas for technique and publication files |
scripts/validate-data.mjs | Checks all data against the schemas and cross-references |
src/ | The website (React + TypeScript, built with Vite) |
.github/workflows/deploy.yml | Validates, tests, builds and deploys to GitHub Pages on every push to main |
Run it locally
Needs Node.js 22 or newer and pnpm (corepack enable sets up pnpm).
pnpm install
pnpm dev # local site at http://localhost:5173
pnpm check # validate data, run tests, build (same as CI)
Add or edit a technique
- Add
data/techniques/<id>.json, where <id> is lowercase with hyphens (e.g. bubble-cursor). Copy an existing file as a starting point.
- Add each paper it cites to
data/publications/<id>.json if it isn't there already.
- Fill
attributes with values from schema/attributes.json. Use [] for anything not yet known.
- Set
review.status to draft. A maintainer changes it to reviewed after checking the paper.
- Run
pnpm validate, then open a pull request.
Not comfortable editing JSON? Open an issue with the technique's name and paper instead.
Status
The 45 selection techniques here were imported from the earlier 3D Interaction Vault. They are marked migrated until each one is checked against its paper. The attribute scheme is version 1.0 (see the About page).
Licence
Code and catalogue data are MIT licensed (see LICENSE). Linked papers, figures and videos belong to their authors and publishers.
