CY2022 Draft Published

[Nick Cannon]

The Draft CY2022 VFX Reference Platform has been published at https://vfxplatform.com.

With Python 3 migration requiring significant ongoing work at studios, our guidance for CY2022 has been to avoid disruptive changes to help ensure that important migration work is completed next year. We do expect bigger changes in CY2023 and will be sharing some advisory guidance on that later this summer.

Please send any feedback on this Draft Platform either to feed...@vfxplatform.com or reply here at vfx-platform-discuss.

We are looking to lock down the Final version at the start of September.

Nick

[Kevin Wheatley]

On Sun, May 16, 2021 at 6:12 PM Nick Cannon <nick....@gmail.com> wrote:
> Please send any feedback on this Draft Platform either to feed...@vfxplatform.com or reply here at vfx-platform-discuss.

I'm sure this will have been considered but for a number of
applications, the full ACES 1.3 version requirement will likely depend
on OCIO 2.1.x due to the gamut compression operation introduced in
ACES 1.3. The gamut compression being one of the key reasons for the
bump in the OCIO version.

Kevin

[Sean Wallitsch]

Thank you for the Python 3.9 upgrade. This will position us well. 2022 will see the last bugfix release for 3.9, but we'll get security patches until 2025.

RE: Targeted OSX version.

It might be worth noting in the matrix that the 10.15 bump means that apps are required to be 64 bit, with support for 32 bit applications being dropped.

--
You received this message because you are subscribed to the Google Groups "vfx-platform-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to vfx-platform-dis...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/vfx-platform-discuss/abc070e8-a682-4696-b882-ccef6d6bbb5dn%40googlegroups.com.

[Bastien Montagne]

Hi,

Thanks indeed for the update to Python 3.9. However, this still means that more than half of the 2022 year will see VFX platform recommending a tool that does not get bugfixes anymore (besides security related issues)... On the other hand, it is definitely a good thing that you ensure expected versions are released before the new platform version becomes active.

So I was wondering what exactly is your policy regarding the limit date of release for libraries and tools to be included in the VFX platform for the next year? Python gets released in October, and has been reliable on their release dates (and general quality of the new versions) for a long time now. Is this already considered as too late for inclusion in the VFX cycle? Or is there another reason to stick to an already relatively old version of Python?

For context, for Blender we had to switch to Python 3.9 already this year because 3.7 had a relatively important bug [1] and we needed the fix in our releases. This is not a common situation, but when it happens it means we might have to break VFX platform compatibility.

Bastien

[1] https://developer.blender.org/T84752

[David Aguilar]

Hi Nick, thanks for publishing this out.

On Sun, May 16, 2021 at 10:12 AM Nick Cannon <nick....@gmail.com> wrote:
>

Ptex 2.4.0 would be nice to include in CY2022.

The changes in 2.4.0 are about symbol visibility for Windows (thanks
to Autodesk for the sug) and have no observable impact on API or ABI
compatibility.

Ptex 2.3 and 2.4 are source (and binary-compatible) but we bumped the
minor since we changed how we expose symbols in libPtex. I suspect
that it'd help out any vendor that might have internal patches for the
same.

cheers,
--
David

[Nick Cannon]

Thank you for all the feedback so far. We'll make some updates in the next few dates to add the suggested notes, and also bump up the ptex version.

On Tue, May 18, 2021 at 8:13 AM Bastien Montagne <bas...@blender.org> wrote:

So I was wondering what exactly is your policy regarding the limit date of release for libraries and tools to be included in the VFX platform for the next year? Python gets released in October, and has been reliable on their release dates (and general quality of the new versions) for a long time now. Is this already considered as too late for inclusion in the VFX cycle? Or is there another reason to stick to an already relatively old version of Python?

From this year we are more strictly applying the policy that a version must be released by September 1st to be included in the Platform for the following year. The reason for this is to allow software providers to have time for testing and quality assurance with the final Platform before the window for publishing major releases against it.

We will still consider later change requests which we will share first with the Working Group, and then here on vfx-platform-discuss for feedback. Only if there are no objections raised will we adopt a late change into a Platform.

So an October release of Python is too late for our regular schedule, but it can be requested for a late change at which point we would solicit the community for feedback.

Nick

 

[Sean Wallitsch]

I think this falls under the same reason that we're still waiting to see if we're going to get a new version of OCIO/ACES. Python 3.10 isn't out yet, and vendors need time to work against the build target.

On Tue, May 18, 2021 at 8:13 AM Bastien Montagne <bas...@blender.org> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/vfx-platform-discuss/e4eac044-c811-45ba-a7ac-7c4b018a04f1n%40googlegroups.com.