2026-03-02 TSC Meeting notes

Review of HPA

• Presentation by Doug and Scott went well

• FFMPEG integration announcement well received

ACES 2 GLES shader compilation issue

• Github issue submitted, should ask Kevin to take a look
  
  

Review of open PRs

• Couple easy PRs for review, please take a look if you have a few moments

• Few bigger PRs from Kevin, we’ll take a look back at those next TSC when Kevin is present

AI coding policy

• LLVM policy as a base, as suggested by LG: https://llvm.org/docs/AIToolPolicy.html

• We have seen both good and not so good AI assisted PRs

• We want to make sure we are encouraging meaningful contributions, not creating barriers to good work

• Doug - pretty sure we’re at the point where AI is used on pretty much every piece of code. Requiring disclosure might be too much of a burden.

• Remi - thinks disclosure is likely a good idea, just from a reviewer perspective.

• Cuneyt - have we talked about using AI code review tools as well?

• Carol: not as a TAC or foundation, but we could certainly try it as a project if we should so desire.

• TSC feels that the rule around good first issues from LLVM in particular is too strict. We don’t want to discourage new contributors from learning however suits them best.

• OCIO in general is a relatively low-volume contribution project, so would prefer to not put more process in place that isn’t strictly needed.

FFmpeg thread

https://docs.google.com/document/d/10PBSdMJfBafHLzSauFpA6NqAvKoMZ6A7E3IO_uLYIFE/edit?usp=sharing

• Summary from thread in OCIO slack from Sam

• Main convo is how to relate OCIO RGB space to CICP / video spaces (YUV, etc). How to do this handoff and convey it

• Zach thinks that the handoff should assume OCIO is RGB full range data. Would be useful to map OCIO display colorspaces to a full CICP tuple in general.

• Please review and comment, and we will post more publicly.