glTF 1.1 draft

[lexaknyazev]

• Overview of spec changes: Spec changes from 1.0 to 2.0 #605
• Direct link to 1.1 spec: https://github.com/lexaknyazev/glTF/tree/master-test/specification/1.1
• Direct link to branch: https://github.com/lexaknyazev/glTF/tree/master-test

---

This PR consists of

• rewritten git history of 1.0.1 branch with changed paths to specification/1.1 directory;
• updated JSON snippets inside README.md - now each of them is a correct JSON;
• fixes for typos throughout spec, schemas, and examples;
• JSON schemas are updated to v4;
• all recent updates from master branch.

---

Specification TODOs

• Update Overview / Reference Card (Update reference card for glTF 2.0 #761)
• Tighten up binary data layout, issue isn't finished yet (Trailing zeros storage #749)
• Skinning update (Culling animated models. #723 (comment))

Implementation TODOs

• Cesium, Support for glTF 2.0 CesiumGS/cesium#4009
• three.js, Support for glTF 1.1 mrdoob/three.js#10257
• BabylonJS
• COLLADA2GLTF, glTF 2.0 support COLLADA2GLTF#3
• obj2gltf, Update glTF output to glTF 2.0 CesiumGS/obj2gltf#35
• gltf-pipeline, Support glTF 2.0 CesiumGS/gltf-pipeline#188

Other TODOs

• sample models for glTF 1.1
• Close issues labeled 1.1, all of which should also be labeled resolved

[pjcozzi]

I don't know if there is an easier way to do this with GitHub given the history changes, but copying and pasting the 1.0 and 1.1 spec into https://www.diffchecker.com/diff has worked really well for me.

[lexaknyazev]

Is there anything else we should create a tasklist for that needs to be done before merging this?

Updated top message with TODO. I'd merge it anyway to bring everyone's attention to updated spec (we can state on the main repo page that it's a draft).

Is the validation completely in-sync with the spec?

Command-line version - yes; web is a bit outdated, I'll rebuild it soon. Also we need to finish those two open spec issues.

[pjcozzi]

I'd merge it anyway to bring everyone's attention to updated spec (we can state on the main repo page that it's a draft).

We can discuss today, but I would rather keep master stable and let users refer to this branch while we get a few implementations together.

[pjcozzi]

@lexaknyazev is COLLADA2GLTF updated for glTF 1.1?

[lexaknyazev]

@lexaknyazev is COLLADA2GLTF updated for glTF 1.1?

KhronosGroup/COLLADA2GLTF#3

[pjcozzi]

@lexaknyazev is COLLADA2GLTF updated for glTF 1.1?
KhronosGroup/COLLADA2GLTF#3

Added to tasklist.

[lasalvavida]

I think this is a typo. Shouldn't this be orthographic?

[lexaknyazev]

Sure. Fixed with lexaknyazev@3604243.

[xelatihy]

Is the changes to use top level arrays as opposed to dictionary already being implemented in the draft spec?

[pjcozzi]

The draft spec is not updated yet for the object to array changes.

[lexaknyazev]

@pjcozzi
Please, review 2113534.

[pjcozzi]

@pjcozzi
Please, review 2113534.

Looks good, except this is vague and maybe misleading:

While some implementations could handle misaligned data, not following this recommendation may lead to extra data processing, performance degradation, and/or inability to initialize attributes pipeline.

Perhaps remove it.

[lexaknyazev]

Looks good, except this is vague and maybe misleading...

The point was that, depending on driver / API misaligned data could be rendered (because driver will fix that).

[pjcozzi]

OK, let's be more precise or remove it.

[lexaknyazev]

Continues in #826.