Implement glTF

[cochrane]

That way the whole thing would be a lot more useful again.

[cochrane]

Rough outline:

• Implement version 1.0 first, including binary format and default material extension
• Each primitive of a mesh becomes a mesh in its own right
• 2.0 as next step after that, with its fancy shaders.

[cochrane]

Immediate next steps (order TBD):

• Support binary files
• Support files without indices

[cochrane]

Gear change: Support only glTF 2.0. Rationale:

• glTF 1 doesn't seem to have widespread adoption; exporters and validators only deal with glTF 2
• Existing models that you can find on the internet all seem to use glTF 2
• glTF 1 and 2 have incompatible schemas (glTF 1 uses objects where glTF 2 uses arrays) which makes supporting both annoying
• glTF 1's default material model of "just put the whole shader file in there" is problematic for e.g. ports to Metal in the far future. The alternative KHR_materials_common is broken (the extension specification does not match the official sample files) and all work at fixing it seems to be targeted at glTF 2, which also has a reasonable (though challenging) materials model.

Will probably also rip out support for glb version 1 because there's no need for it in the glTF 2 world.

[Stylistc86]

Hi, Torsten This is Stylistic86. Here is one of the files in question that I was referring to. Let me know what you think. Thank you, Stylistic86

…

On Fri, Apr 2, 2021 at 4:54 AM Torsten Kammer ***@***.***> wrote: That way the whole thing would be a lot more useful again. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#122>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AI5IN7NIKDC4KXGM2S25A6TTGWA3TANCNFSM42IRI2MA> .