RFC: Restore nebula.gl as a first-class visgl repository

[charlieforward9]

Summary

nebula.gl was the original home of what is now @deck.gl-community/editable-layers. During the deck.gl v9 upgrade cycle, the module was migrated into deck.gl-community and the nebula.gl identity — its domain, npm package, brand, and community — was effectively retired.

This RFC proposes restoring nebula.gl as a first-class repository under the visgl GitHub org, giving editable/drawing capabilities a proper home with its own identity, docs, and release cadence.

---

Motivation

• The nebula.gl domain still has name recognition in the geospatial community
• Users searching for polygon/shape editing on top of deck.gl still land on the old nebula.gl site, which is stale
• @deck.gl-community/editable-layers is a significant, standalone capability — it deserves its own identity, not a subdirectory
• PR #512 attempted to restore nebula.gl design RFCs into the docs tree and was rejected; this RFC proposes the right-sized solution instead
• Issues #38 and #82 track the v9.3 roadmap — a restored nebula.gl repo is the natural milestone target

---

Proposed Scope

1. GitHub Repository

• Create visgl/nebula.gl
• Move modules/editable-layers source into the new repo
• Keep a re-export shim in deck.gl-community during a deprecation window

2. npm Package

• Restore nebula.gl package on npm (currently owned by the original authors)
• Publish nebula.gl as the canonical package, deprecating @deck.gl-community/editable-layers with a migration notice
• Coordinate with current npm package owner for transfer or scoped rename to @nebula.gl/core

3. Domain Transfer

• Transfer nebula.gl domain to the visgl organization (same stewardship as deck.gl, loaders.gl, etc.)
• Redirect current nebula.gl site to new visgl-hosted docs

4. Docs Site

• New docs at nebula.gl (or nebula.gl.dev) using the same vis.gl docusaurus setup
• Migrate existing @deck.gl-community/editable-layers docs
• Restore architectural rationale (EditMode interface, Tentative Features, Guide System, react-map-gl-draw notes)

5. Release Cadence

• tracking deck.gl peer dependency compatibility and major.minor semver, varying on patch versions

---

Open Questions

1. Who currently owns the nebula.gl domain and npm package? Do we have a path to transfer, or do we need a new scope (@nebula.gl/...)? @georgioskarnas
2. Sponsorship from an angel looking to support @charlieforward9 as the core maintainer?

---

Related

• #512 — original PR that motivated this discussion
• Original nebula.gl repo: https://github.com/uber/nebula.gl
• editable-layers npm 30-50k/month

[charlieforward9]

cc @ibgreen @chrisgervang @felixpalmer @georgioskarnas @matt-agintel @ilyabo

[matt-agintel]

Add a quick note of support here.

Charlie has been doing the work around deck.gl editing capabilities in the context of our mapping-heavy project (Agroview). Polygon and shape editing are real workflows for our users, so this aera of the ecosystem matters.

We appreciate the ecosystem and support Charlie contributing upstream where it makes sense. The work he's doing has been useful for us and hopefully other building on deck.gl

[charlieforward9]

it was agreed that isolating this package is not the right direction - instead moving this into deck.glonce it can be proved that it is stable enough to perform deck native operations and view integrations

[chrisgervang]

Agreed, stability and maturity are important for a smooth graduation and sustainable long term maintenance.

Hoping to learn some more context since I haven't been in the conversations about upstreaming and am still hesitant to absorb the API surface.

Here are some open questions I have:

I may be operating with an outdated understanding, but I thought this module needed special hooks on the Deck object to work that introduce compatibility issues with other integrations. Where can I learn more about how the interactive systems integrate with deck?

What gaps are we are hoping to close before upstreaming?

What is the surface area that we're hoping to upstream? Layers, UI controls, etc

[charlieforward9]

Where can I learn more about how the interactive systems integrate with deck?

I haven't worked on the docs very diligently but I'll request your review when I do.

What gaps are we are hoping to close before upstreaming?

All open bugs in here with editable-layer label on them + layer arrowfication?

What is the surface area that we're hoping to upstream?

1 layer: EditableLayer, all modes, a modernized widget pack, 1 modular example.