VFX Graphs - API or textual representation?

I’m not taking to node editing very happily. It seems fairly long winded and time consuming, Really I yearn to do things in the way that seems most efficient and productive for me - typing stuff in to a code editor.

For SRP we can avoid Shader Graph and write shaders the old way. (or mix and match the two workflows by writing custom nodes that do most of the work…)

What’s the plan for the VFX Graph? I’ve skimmed the scripting docs and there doesn’t appear to be a way to create a VFX from scratch at the moment. VFX files aren’t sanely human-editable (it’s YAML goddamit. It’s meant to be human readable. Grumble… Grumble…)

Will there ever be support for a declarative or procedural text format for VFX graphs? Is there more stuff possible at the moment that I missed in the scripting docs?

We dont really plan to work on a convenient way to create a vfx asset from code only.

It will however be possible to manipulate the graph API to generate a VFX graph from code without using the nodal UI at all as soon as we make the API public but it will never be the recommended way to do.

Also once API is public it will be possible to extend behaviors by adding new blocks/operators in C#. But again the recommended way to go to extend the vfx library would rather be via nodal UI once subgraphs are there.

As for the YAML, it is standard unity serialization so the readability/verbosity issues applies to any unity asset.