Link Search Menu Expand Document

ASWF TAC Meeting - September 9, 2020

Video Conference Link

Voting member attendance

  • Daniel Heckenberg - Chairperson, Animal Logic Pty Ltd
  • Gordon Bradley, Autodesk
  • Pilar Molina Lopez, Blue Sky Studios, Inc.
  • Michael O’Gorman, Cisco Systems Inc.
  • Henry Vera, DNEG
  • Bill Ballew, DreamWorks Animation
  • Matt Kuhlenschmidt, Epic Games, Inc.
  • Brian Cipriano, Google & OpenCue Representative
  • Sean C McDuffee, Intel Corporation
  • Larry Gritz, Sony Pictures Imageworks
  • Jean-Francois Panisset, VES Technology Committee
  • Cory Omand, The Walt Disney Studios
  • Kimball Thurston, Weta Digital Limited
  • Eric Enderton, NVIDIA
  • Sean Looper, Amazon Web Services
  • Michael Min, Netflix
  • Michael B. Johnson, Apple
  • Dave Fellows, Microsoft
  • Sean O’Connell, Advanced Micro Devices
  • Laurent Ruel, Unity Technologies
  • Ken Museth, OpenVDB Representative
  • Michael Dolan, OpenColorIO Representative
  • Cary Phillips, OpenEXR Representative
  • Joshua Minor, OpenTimelineIO Representative
  • Chris Kulla, Open Shading Language Representative

Other Attendees

  • John Mertic, Linux Foundation
  • Emily Olin, Linux Foundation
  • JT Nelson, Pasadena Open Source consortium / SoCal Blender group
  • Carol Payne, Netflix, Diversity+Inclusion WG
  • Andrew Grimberg, Linux Foundation Release Engineering
  • Jeff Bradley, DreamWorks
  • Mathieu Mazerolle, The Foundry
  • Greg Denton, Microsoft Azure
  • Nick Porcino, Pixar
  • Stephan Steinbach, OTIO
  • Eric Reinecke, OTIO / Netflix

Apologies

Agenda

  • New member introduction: Laurent Ruel, Unity Technologies

    • Welcome!

    • Laurent: started in broadcast, worked at TVA Networks in Montreal, production, control systems, on air automation systems, architecture, coding and design. Election and newsroom systems. Started a company in 97 to make a video server / clip / stills store / character generation on a PC using Matrox cards, became part of Kaydara. Worked on Motion Builder, leading engineering / QA / training / support, became part of Alias, then Autodesk. Product Manager on Maya, then cloud backup at Mozy (sp). Then Grass Valley (monitoring & routers) before joining Unity in 2016, responsible for engineering in Verticals Group: architecture, engineering, construction, M&E, manufacturing and retail.

    • Daniel: have you had any previous engagement with the ASWF projects? Laurent: not so much yet, but working on OTIO integration, USD, OCIO.

  • OpenTimelineIO incubation review: Josh

    • Vote: Project status and next review

    • Josh: A bit over a year since OTIO was accepted in incubation stage in ASWF. Set some internal goals around community and industry awareness. A number of items on the checklist to graduation have been done.

      • Healthy number of committers

      • Substantial ongoing flow of commits

      • Document project owners and committers

      • Technical lead representative to TAC

      • Initial license scan

      • …

    • Community contributions:

      • Netflix & Storm Studios: image sequence

      • Autodesk: spatial coordinate systems schema

      • Rodeo FX: EDL improvements

      • GSoC: Java & C bindings

      • Igalia: xges improvements

      • Upcoming:

      • Blender community (JT Nelson)

      • Frame.io

      • Ahead.io (Cezanne & shott.io)

    • Would be good to have more formal channels to engage with companies

    • OTIO is very please with working with ASWF

    • Outstanding issues:

      • Modified Apache 2.0 license

      • TSC Charter: blocked on license (informal TSC for now)

      • CII best practices badge

        • Currently at 61%

        • Security expertise & guidance needed: what does security look like in a framework that has Python plugins?

        • Hoping to focus on this

    • Still work to be done before graduation stage: would like to remain in incubation stage.

    • Requesting continued assistance moving forward with licensing issues.

    • Daniel: great presentation, gives a roadmap for moving forward. Are there particular areas where TAC can help unblock some issues, offer expertise? Josh: hard to put finger on what needs to happen with licensing. There isn’t a lot of urgency getting these issues straightened out within Disney organization. Setting a timeframe could help to provide momentum. John: it’s a decision of the TAC to renew incubation renewal. ASWF / LF has been working with OTIO for a year, licensing issues have been starting to get unblocked, in communication with Disney legal. There is a path forward, but these things can take time.

    • Josh: what are we trying to “protect against” in the context of security requirements? Daniel: we grappled with those issues with the first set of projects, could be time to revisit. OpenEXR has been the most visible example and gives a good baseline. Some of our other projects have a plugin approach as well (OpenCue).

    • Josh: OTIO does see the value moving forward, but who gets to judge “how secure are we”?

    • Larry: we had this set of discussions with some of the earlier project, a simplified view is that issues related to crypto and login security are typically not relevant to our projects, we don’t have user IDs, logins, PII… So many of these criteria don’t apply to our projects. But because these packages can be embedded into other systems, we don’t want to be the source for security vulnerability. Static Analysis + some Fuzz Testing is sufficient to self certify. Kimball: can also define subsets of the projects that have higher security requirements / a build mode for just the core of the system without the extension mechanism.

    • Daniel: security reporting and address process needs to be in place. Josh: got a notification that a component of the build system had a vulnerability, and was able to address it. John: there are offerings from LF to help with vulnerability management.

    • Stephan: like the idea of being able to disable the plugin system, or enable only internal plugins. Would be good to have partnerships with the community to help adopt processes from the other projects.

    • Josh: as commercial applications start adopting OTIO, expecting more feedback from vendors.

    • Eric: any guidance on how often fuzzing analysis should be performed? Kimball: OpenEXR already had a fuzz system built into it, but it takes so long that it only gets run once a week, same as static analysis pass. The Google OSS Fuzz service is outside the project’s control.

    • Daniel: for OpenCue we extended a maximum review period of 6 months since they wanted to move out of incubation. What would OTIO want?

    • Josh: 6 months sounds fine, sooner would be better, but that would be a good deadline for now.

    • Larry: should we think about triaging what the holdouts are, and expect the project to have made substantial project on parts of the process it has active control (CII bad, TSC minus the licensing specifics). Josh: that sounds reasonable.

    • Larry: present motion to review in 6 months with expectation that parts of the process under control of the project have been addressed. Second from Kimball.

    • All ayes, no opposed or abstain, unanimous, adopted motion. Congratulations!

  • Follow ups:

    • git branch naming: Kimball

      • Example of how it was implemented for OpenEXR

      • Kimball: haven’t actually adopted the new scheme yet due to other priorities, EXR TSC in unanimous agreement, for a development / code project, having branch names as “development” and “release” was clearer than just “main”. Even if not actually released (yet), code in the “release” branch will be less in flux than on the “development” branch. But could be fine to use the “main” branch for the EXR Image repo (no development process there).

      • Larry: the possible source of contention is that almost all other projects are settling on “main”, agrees on “rel / dev” is better, but being different than anyone else is a “tax on the future”, newcomers to the project will have to face a (small) hurdle. Also might be worth delaying since GitHub is offering guidance that they will offer features in a few months to help deal with branch naming issues. So if we postpone until GitHub is ready, we might avoid some initial complexity.

      • Kimball: more generally does it make sense to have a discussion about branching strategies across all ASWF projects?

      • Nick: GitHub project pages can be used to document the branching scheme

      • Kimball: happy to consider the support from GitHub

    • slack archiving: JF

      • Matrix and other LF supported paths: Andrew

      • Daniel: most of discussion / advice was that although the cost of moving to commercial Slack instance (about $30K/year), the value is limited, and we should make sure to not depend on Slack as a long term archiving solution. Also a lot of our Slack use is private IMs.

      • Not going to request funding from Governing Board to switch to paid instance.

      • Can someone articulate the contrary view? Or should the LF consider alternatives to sidestep some of the limitations.

      • Kimball: should be an ephemeral communication, anything long term should be recorded in documentation / code.

  • Working Groups

    • CI

    • Diversity & Inclusion

      • Carol: had great meeting on mentorship process, working with LF

      • Monthly D&I meeting today.

      • Will reach out to TSC chairs to look at what mentorship projects would look like for those projects.

      • Daniel: lots of interest in mentorship at the TAC level

      • Larry: since GSoC just wrapped, 3 TSC representatives should present on successes and challenges of GSoC projects from this summer at next TAC meeting. Although led the GSoC effort in general, don’t have specific perspective of the work performed at the project level. Daniel: will schedule this for next TAC meeting and make sure the relevant people are available.

    • Python 3

    • USD

  • Technical Project updates

  • Update on candidate projects

    • WG on Collaborative Remote Review

      • Last round of outward facing events, Collaborative Remote Review was identified as important industry project.

      • Looking for someone to take up that project: sponsor and lead the effort. Kimball: have we had previous discussion on this? Is interest, interesting but potentially challenging project.

      • Daniel: in all the surveys since even before ASWF creation, an open source media player / review tool was identified as a need. Clear development path, ability to embed, ease of integration… Good candidate for open source, some studios have internal tech they may want to open source. Presentation by DJV. Haven’t found a way to move beyond that stage.

      • WFH requirements in last 6 months has added to this need. How do you do collaborative “theater” reviews when you can’t put everyone in a theater.

      • Kimball: wouldn’t necessarily want to lead this group, there are people at WETA who are much more knowledgeable about in-house playback system. Daniel: sponsor does not necessarily need to be the lead, can delegate.

      • Daniel: up to the WG to define the scope. Kimball: will try to get something started.

      • Michael: there’s a bit of a chicken and egg. In this case it’s more a topic for discussion. Every studio has customized their own or third party player. So very open ended topic.

      • Kimball: seems the scope wouldn’t be to define an actual project, but a baseline of features required to meet this need.

      • Sean: are we also looking at existing projects? Kimball: yes, seems appropriate. Daniel: good topics for discussion. JT: meeting with Darby from DJV tomorrow, he would like to have more clarity as to what it would mean for a project like DJV to be involved with ASWF. So could make sense for him to get involved in this working group.

    • Datasets for testing and/or machine learning ( John M has survey )

      • https://forms.gle/3Fgg9QkjRufcYGkp6

      • John: every time a survey comes up, content / data asset repository keeps coming up and floating up the list of desired. Talking to David Morin, we should create a survey to help better define what a Creative Commons data repository would look like. Link above, please complete this short survey before EOD today. Trying to find examples out there of not just “pull only” repositories, but also ones where you can contribute.

      • A good topic for the creation of a working group.

      • A new repository of ASWF assets / logos: https://artwork.aswf.io/

      • Is a curated list of links to existing datasets be a good start? Google Doc as a way to start? Some discussion on the topic on the OTIO Slack channel.

      • John: trying to get an idea of the scope and focus of such a repository. “A data driven approach to a data repository”

  • Next meeting

    • Sept 23, 2020

Action Items (AIs)

  • Daniel, Larry: Arrange brief GSoc student presentations and retrospective for next TAC meeting

  • Daniel: Formalise decision to remain on free Slack tier with guidance to move important discussions / decisions to other formats.

Notes

Chat