2022-08-09 OpenAssetIO Meeting @ Siggraph

Write-up of a meeting with Mathieu Mazerolle (Foundry) on OpenAssetIO.

Participants: Jeroen Bakker (Blender Institute), Julian Eisel (Blender Institute), Mathieu Mazerolle (Foundry Research)

Context:

  • Basically an informal meeting to get to know each other, the projects and what further steps to take.
  • OpenAssetIO acts as a common bridge between a DCC (Blender, Modo, Maya, etc) and an asset management system (such as Kitsu [1], ftrack or ShotGrid).
  • It is an intentionally thin library, focused on a few core aspects to exchange information about assets, and does not define asset storage, contents, types, etc. It’s “violently unopinionated”, as they like to say.
  • OpenAssetIO is an ASWF project now, not a Foundry project anymore. That means the project becomes more of a community project that needs contributors.
  • Sybren and Julian already looked into OpenAssetIO some months ago. They found that it fits very nicely with the asset system design, but can wait until looking more into the core asset system implementation + BPY integration.
  • There have been a few conversations with the OpenAssetIO team since then.

[1] Kitsu is more of an asset tracer than a fully fledged asset management system. For the sake of OpenAssetIO, it can still serve as asset management system though.

OpenAssetIO & Blender:

  • Julian: Basically OpenAssetIO isn’t necessary for Blender - we could just do things our own in the asset system and BPY. But it’s a great opportunity to adopt an open standard that will make integration with other systems easier.
  • Again, the design seems to fit nicely with the one for Blender’s asset systems: Assets as thin abstractions with characteristics and metadata, assets that can be but don’t have to be files, URIs as references, … We also don’t want to tell asset management systems what to do, we might steal that “violently unopinionated” slogan :slight_smile:
  • The URI based asset resolving sounds like something we were thinking of doing anyway. Depending on the exact functionality OpenAssetIO has for this, we may be able to use that at the asset system core.
  • So all things considered, it would be a big missed opportunity to not support OpenAssetIO.

Contributing to OpenAssetIO

  • OpenAssetIO needs contributors and demo use-cases.
  • Since resources for asset related topics are scarce at Blender (no employed dev can actively work on the asset system currently), we wouldn’t be able to contribute much unfortunately.
  • Good thing is, because of the intentionally small scope of OpenAssetIO, there shouldn’t be a lot needed from the Blender side. If OpenAssetIO manages to keep the scope, it can be very helpful more or less as-is.
  • So focus should be on making the best use of OpenAssetIO in Blender, while keeping the communication between both projects active.

Next steps:

  • Involved Blender devs can join the monthly ASWF OpenAssetIO meetings.
  • Julian will talk to the Blender Studio about making their asset tooling (such as watchtower) and asset management system (Kitsu Based) OpenAssetIO compatible. This would be a nice use-case to present OpenAssetIO.
  • How exactly an integration into the asset system and BPY should be done needs to be figured out. Important topic to look into for the further design.
  • Once this is better defined, a code.blender.org blog post could show the benefits of OpenAssetIO in Blender, and how we plan to integrate it. This is a way to show support for such open initiatives.
13 Likes