ASWF TAC Meeting - 05 May, 2022

Video Conference Link

Voting member attendance

Other Attendees

David Morin, ASWF
JT Nelson, Pasadena Open Source consortium / SoCal Blender group
Scott Wilson, Rust WG
Deke Kincaid, Digital Domain
Doug Walker, OCIO / Autodesk
Lee Kerley, Sony Imageworks
Morgan Prygrocki, Adobe
Robin Rowe, CinePaint
Sean Wallitsch, Dreamworks
Jim Helman, MovieLabs
Mallory King, Spire Animation
Robert Vinluan, SideFX
Jason Scott, FuseFX / Rising Sun
Jeff Bradley, Dreamworks
Nick Porcino, Pixar Animation Studios

Apologies

Cary Philips, OpenEXR Representative
Sean Looper, Amazon Web Services

Agenda

Sync up with VFX Reference Platform is delayed
- But can still discuss it amongst ourselves
- Round table around how this affects projects. Have all the projects made a proposed candidate version for the VFX Platform 2023?
  - Carol: OCIO will be 2.2. Doug: provided feedback in retrospective we presented, we get feedback at SIGGRAPH, so ideally would be to present things at SIGGRAPH and have a bit of time to have time to finalize things. OpenVDB did October last year, that would also work out for OCIO. Last year we did 2.1.0 on August 31st to comply with VFX Platform 2022, but ideally would prefer October if that’s possible. Sounds like OCIO will apply for this “extension”.
  - Ken: OpenVDB has done it for the last 2-3 years and will do it again. Aiming for a major release, version 10. Kimball: October timeframe? Ken: yes, releasing 9.1 in a month, then going for version 10, which we want for VFX Platform 2023, aiming early October. Seem to have this conversation every year. Hopefully would be better if it was in writing, but there’s reluctance to do that?
  - Kimball: interesting they are looking at changing the compiler version but not the C++ year / level, would seem to be an easy(er?) thing to do? Larry: we could bump up to C++20? Kimball: not sure we’re ready for that! Do we know what C++20 support is link in Visual Studio? Chris: it’s a subset, doesn’t seem any compiler fully implements C++20 yet, so 17 seems like a more conservative choice. Ken: and less exciting! Kimball: really want the structure init stuff in C++ 20. Larry: we all have CI, we test against all the compilers. Chris: C++ 20 is actually fairly well covered across compilers: C++ Compiler Support
  - Kimball: would be good to add text to VFX Platform to deal with Clang, using libstdc++, but also want -fgnuc-version=9.3.1 to have headers behave correctly. How many are compiling with Clang? Larry: didn’t know about that. Test against gcc versions has to happen correctly, and don’t get treated as if you are a very old version of gcc and lose more recent features. Libstd++ s not technically but effectively part of gcc, they have checks in their header files on the gcc version. There are other separate packages that have tests against gcc version in their header files, for instance test for built in intrinsics. Could be added as an “asterisk” on the VFX Platform page.
  - OpenEXR: Kimball likely 3.2
  - Nick: OTIO will be 1.0, but not part of VFX Platform yet
Carol: quick plug for mentor volunteers for Summer Learning Program, D&I effort, completely in the ASWF. Entry level technical interest in VFX / animation industry. Looking for technically oriented students, window closed last week, will accept 20 people, so need 20 mentors. Lots of people did it last year, will put a bit more structure around what it means to be a mentor. Any TAC members are more than qualified if you have time and interest. No requirement for people to be students, just 18 and over. Got a great range of applicants, a lot are focussed on technical aspect of the industry. Also changed the learning platforms they get access to, one of which is fxphd. Will also use LinkedIn Learning. Program is June / July, 1-2h per week commitment. Focus last year was women, this year is people of color. Spreadsheet for Summer Learning Program Mentor Signup
- Sean: is reviewing done? Carol: yes, only need mentors at this point, already have enough volunteers as team leads. Larry: how many students? Carol: 20, same as last year, do a repeat of last year. May expand last year.
- Wave: had good experience last year, was pretty easy, the person was motivated.
- Carol: try to match applicants stated interests with interest of the mentor
Larry: cross group OpenEXR / OCIO meeting tomorrow, should we try to connect colorspace with OpenEXR header metadata, properly record how to do this and how to connect OpenEXR files to OCIO colorspaces.
- An issue that’s come up in a lot of places, not sure we can make progress, but worth discussing. Is there a bridge that makes sense? Might be interesting to have someone from MaterialX also show up. Jonathan: relevant to MaterialX and USD, will try to attend. Will be important to deal with potential conflicts, what if a texture specifies a color space, and USD / MaterialX specifies another one?
Discussion on security issues in projects
- OpenEXR got an email from RedHat, are you assigning CVEs to reports from OSS Fuzz. Any other projects using OSS Fuzz? Otherwise you are lucky.
- Larry: only assigning CVEs for things that are exploitable? Kimball: question from a security person, we typically don’t report everything unless it’s exploitable, otherwise it’s a “bug report about a bug report”.
- Have had discussions around security reporting in the part.
- Nick: had a rash of CVEs filed a couple of years ago, where were those coming from? RedHat / Debian directly? Maybe a CVE is a marker that you need to adopt an upgrade sooner rather than later.
David: we had earmarked next meeting to do a poll / design session to help figure out strategy. We do this on the board from time to time, we choose a topic to dive into, and Rob Bredow is adamant we only pick one topic at a time so we can go deeper. So would like to do a poll to suggest topics that we would like to discuss in this format. https://pollev.com/aswf
- Board has approved succession rules for itself, including term limits. Some applies to the TAC, some doesn’t.
  - Continuity planning: projects / WGs can still greatly suffer if a single person “retires”
    - This can happen inside a facility, can have a single expert on any specific project. High “bus factor”.
  - Increasing long-term contributor base
  - Is it easy enough for people at member companies to get permission to contribute to projects?
  - Resource allocation and scalability
  - large asset storage for DPEL and other groups
  - New projects / tools needed to keep adapting to new business requirements?
  - Bringing in new members from outside of the ASWF
- JF: would be good to learn from other LF foundations.
- David: we’ll pick the “continuity” topic. Are we doing this at the next meeting in two weeks? Kimball: yes. Is there a project update scheduled for next meeting? No: it would be a full hour. David: so may pick a few of the topics to have enough to fill the hour. Think about these topics, bring so discussion points and potential solutions.
David: last board meeting discussed a succession plan. Board was presented with a vote on the plan. Came from the work of D&I group that brought up lack of diversity on the board. D&I did research and presented a number of ways board have dealt with this issue. AMPAS has also gone through that process a few years ago. This was discussed, board created a subcommittee to see what to do and presented a plan.
- Term limits: 4 year maximum from start of GB membership
  - 2 year breaks between terms
  - No lifetime limits for board members
- Phase-in plan:
  - By december 1, 2022, rotate out a minimum of 25% of qualifying GB members annually on a voluntary basis, with subsequent minimum 25% rotation targets each December 1 thereafter until all founding GB members are rotated out.
  - Reach annual 25% rotation goals by lottery if needed.
  - GB Alternates:
- GB Alternates:
  - GB members select an alternate within first 2 years of membership
  - Alternates serve as voting GB member when primary GB member is unavailable
  - Alternates can be changed at any time
  - There are no term limits for alternates
  - Phase-in: start introducing alternates in March 2022, all to be in place by 1-july-2022
- Diversity Goals:
  - …
- A deck has been put together
  - Defines what a board member is
  - Not just people who show up at the meetings and vote on topics, but also ambassadors for the ASWF
  - Growth of foundation has hit a plateau, $1m run rate of membership, which allows us to operate, but proving difficult to get new members. Every new prospective member isn’t part of core member of companies that understands and uses open source, we have most of the companies in our industry who are OS savvy. Have talked to a company that has a lot of software and they use our projects, but they don’t think in an “open source way”. So requires education, “a longer sales cycle”.
  - Lots of companies present at the Open Source Forum, helping to enroll board members to help bring these companies in.
  - Room for new people to come in, will keep an eye on our D&I ratios and goals.
- Scott: what happens if it’s someone to retire from the board, but they don’t have anyone to take their spot? David: as long we didn’t have term limits, we didn’t think about it much. Now that we have those term limits, a board member knows they are only there for a certain amount of time, so they can plan accordingly, and be motivated to look for these people, and start by looking for an Alternate. So it’s not as big of an ask to ask someone to attend at first without initial responsibilities. Starts a process of internal communication to get ahead of the curve. We are looking for 25% rotation each year, when someone is getting closer to their 4 year limit, GB can help if a board member has trouble finding someone, can help the board member promote ASWF within their company. It sets a process in place. Similar to issue with project maintainers: when there’s no rule, people tend to stay for a long time, and don’t think about a succession plan until there’s an emergency. Having Alternates and Term Limits on the board forces you to think about succession. Without input from D&I, board could have gone on for quite a while as it was.
- Michael: are there conversations around member retention and membership happiness? Haven’t heard that type of focus, as we get new membership, haven’t heard about membership happiness? David: no, we don’t have a “meter” of member happiness, but a great question. There’s a membership fee every year, we have a discussion with our board members and general members, “happiness” is a sub topic in that discussion, is there anything ASWF can change. Have lost some members over the years for different reasons, a general member decided their fee would be better spent somewhere, some misconceptions at first, some companies think that at the general member level, they’ll meet a lot of studios and have PR opportunities. Have lost some Premier members since the champion who originally sold membership internally left the company, some companies went through more difficult financial times. Other reasons not necessarily related to “happiness”, but a great metric to have.
- Larry: for Premier members, some decided it was worth spending the money for risk mitigation given their level of dependence on these projects. My employer was very pro open source, but there are now projects that we are critically dependent on than 3 years ago (MaterialX, USD, OTIO…). So we’re even more risk exposed. So maybe one way to measure the importance of ASWF.
- David: should we do a survey of some of these parameters? Larry: Which of these projects would be in big trouble if it went away? David: could also serve in recruiting of external vendors.
- Shout out to Animal Logic’s post about The Open Source Obligation, why they do Open Source, their ALab initiative. Encourage people in our member companies to make such statements to the outside world, state internally and externally the commitment to open source and the value proposition of being an ASWF member. Other companies will take notice, and I can use that to promote the foundation. “Happiness” and “Dependencies” are good benefits.
- Carol: decided to focus on the Governing Board first, don’t want to take rash action with the TAC, how do we start talking about these issues in the context of the TAC, strategies around succession, long term success of projects is influenced by diversity. It wasn’t easy to do with the board, but it will be more complicated with the TAC. Larry: more leery of forced term limits for projects and TAC, but a good idea to designate an alternate. Is anyone else up on these topics, a useful exercise, maybe forcing everyone to designate alternates would be good.