Director of M&E at Microsoft Azure, covers from scripts to post. Came from industry, worked at Pixar and ILM, excited to work on projects with industry partners.
Follow ups
CI documentation and sample project? (JF)
Nothing to share yet, but still committed to it. Will take shape of personal GitHub repo, then to be moved to ASWF when fleshed out.
Daniel suggest that the sample project (“teapot project”) could be used to demonstrate the use of existing ASWF projects (EXR, OCIO…). Show cross project dependencies, CI integration… Should this be a separate project to avoid mixing a project that demonstrates structure vs a project that demonstrates cross project integration.
Gordon: components designed to work within a larger framework, how compeling an example can you build by gluing these components together?
Sean: can we justify this project by connecting the dots between the ASWF projects and new projects that would fill the gaps?
Michael Min: how far do we want to scope this effort?
Cary: haven’t heard discussions about an objective to build a complete pipeline within the context of ASWF.
Daniel: there are other groups that have as a goal a fully open source production pipeline. Was suggesting a more modest goal to show a tutorial for developers.
Cary: had feedback about issues building OpenEXR as a sub-project of another CMake-based project. So more examples of CMake integration between projects would help.
Eric: a project that links all other projects is the hardest thing to compile
Member survey to guide TAC and CI efforts (Gordon)
Take our ad hoc surveys a step further, survey done at SIGGRAPH was not easy to determine “which voices are we hearing”.
Propose as a first step to do a deeper dive on the raw data that was gathered.
What are the questions we want to answer?
What would be the right survey questions to answer that?
We should have a post mortem of how well we’ve delivered on the expected value of the member companies in the first year.
Asking these questions seems to be larger than just the TAC, should this topic involve / be driven by the Governing Board? Do such discussions exist?
Daniel: first year strategy nominally runs until end of tenure of David Morin, in the process (with Daniel) of formulating goals for next year. Interested in hearing feedback from the TAC: next steps for CI, project needs… So it’s a two way interaction.
Gordon: can work with David, Emily, John and Daniel to help define this to help put this together.
Daniel’s tenure ends in September, discussed with John a process to elect next TAC char, will send out a call for nominations via email list. If there are more than one candidate we will vote.
Anyone interested should reach out to Daniel to discuss what was involved. Estimate 4-5 hrs per week.
Healthy committer base across vendors, VFX companies and individuals
Healthy flow of merges and commits
Completed CII badge
License scan completed (initial issues with embedded dependencies and their licenses)
ASWF has really helped with organization of governance and structure, documentation, contribution, security… Highly beneficial to the value of the project, and help new participation.
Value of OCIO to ASWF mission: project is well adopted in the industry, including outside the VFX industry (games). Speaks for itself. Looking forward to OCIO v2 next summer.
Daniel: project has been excellent at contributing to TAC process, CI system, setting of direction.
Andrew: project not yet migrated to ASWF GitHub, but will be done soon.
Quorum is met
Daniel proposes motion for graduation, Gordon seconds.
Many ayes, no abstains, no against. Motion passes unanimously, will be recommended to Governing Board for graduated status.
Goals for TAC: Year 2
CI
Dependency management
Security expertise and processes
Documentation standards and standardization
Vector / matrix library
Technical Project updates
OpenVDB
OpenColorIO
OpenEXR
Cary: beta release is out, feedback about CMake issues when building in context of larger CMake-based project.
Will be working on more disruptive changes.
Moving repo to ASWF GitHub waiting on CLA setup (business / legal)
Eager to move on splitting imath to separate project. Will touch on issues of whether it’s a separate, independent project, in a different repo with different license, CLA… Or is there a notion of a “dependant” project that shares infrastructure?
Could ASWF project inherit from a base project for some infrastructure (licensing / doc)?
OpenCue
OpenTimelineIO
CI updates
GPU build agents – hosting?
Not covered by free Azure build agents
Andrew: Azure Pipelines does not support a free pool of GPU builders. In talk with representatives at Azure Pipelines, where is Microsoft on roadmap with dynamic pool?
We (LF) have found a way to pretty easily deal with the cost model to manage Azure resources. Would be an option to spin up some managed pools.
Brings up issue of Jenkins environment: recommend to turn down that environment to save those costs. There’s a non trivial amount of work to spin up that infrastructure, takes about 2 days of work to spin up that environment.
Kimball: do we see ever needing Jenkins in the future? Andrew: OpenCue went to Azure Pipelines directly.
Kimball: move to decommission Jenkins, Larry seconds.
All Ayes, no opposed or abstain. Vote carries. Andrew to stand down the ASWF Jenkins infrastructure costs.
Sean: AWS would be happy to provide funding for GPU instances. Can use any resources for Azure Pipelines builds.
May need to use Mac Stadium for Mac GPU, but OCIO is happy with just Linux for now for their test suite. Doug: would be happy to test on Mac and Windows as well.
Update on candidate projects
Rez
Blog
Emily: launching ASWF blogs, call for participation went out on mailing list