implement KHR_materials_dispersion

1 files changed, 4 insertions, 5 deletions

diff --git a/doc/other.md b/doc/other.md
index b3f8c93..3108e78 100644
--- a/doc/other.md
+++ b/doc/other.md
@@ -17,7 +17,7 @@
 | KHR_lights_punctual             | Supported        |
 | KHR_materials_anisotropy        | Todo             |
 | KHR_materials_clearcoat         | Todo             |
-| KHR_materials_dispersion        | Todo (2)         |
+| KHR_materials_dispersion        | Supported        |
 | KHR_materials_emissive_strength | Supported        |
 | KHR_materials_ior               | Supported        |
 | KHR_materials_iridescence       | Todo             |
@@ -31,15 +31,14 @@
 | KHR_texture_basisu              | Not required (3) |
 | KHR_texture_transform           | Not required (4) |
 | KHR_xmp_json_ld                 | Todo             |
-| EXT_mesh_gpu_instancing         | Not required (6) |
+| EXT_mesh_gpu_instancing         | Not required (2) |
 | EXT_meshopt_compression         | Todo, maybe      |
 | EXT_texture_webp                | Supported        |
 
 1. We can do the same by sharing subresources.
-2. The gltf rust crate does not support that extension.
+2. Shared parts can be optimized to instanced drawing by the client. This would
+   also only be a bandwidth optimization, avoiding gigantic prefabs.
 3. KTX is worse then WebP in terms of size. Transcoding on the client is
    possible.
 4. Automatic atlas packing is also possible on the client.
 5. Introduces too much complexity and interfers with our animation system.
-6. Shared parts can be optimized to instanced drawing by the client. This would
-   also only be a bandwidth optimization, avoiding gigantic prefabs.