Over the year we’ve accumulated a lot of validations from various use cases for models, lookdev, rigging, etc.
As much they are important in production, they also introduce quite a high barrier for people who just want to try the basic functionality. For example getting most models published requires quite some work to comply with the requirements. Naming, UVs, transforms, etc.
I’m wondering whether we should by default disable most of these advanced validators to make it easier to start and then introduce them to production if the studio wishes. With pype 3.0 settings GUI this is as easy as ticking a bunch of boxes and saving defaults so. That would keep them available for anyone, but would simplify the start quite a lot.
Interesting! The model requirements are quite strict, but it has set a high standard of publishes so we can rely on models.
Are there any cases where the validators were stopping/hindering the production?
Nope. I would actually strongly recommend everyone to use them as much as possible in production. This is more of a case that when pype is deployed its always a mayhem of
my models are not going through and hours of explaining to artists before everyone gets on board.
While this improves production quite a bit, it is also a friction point on adoption. We’ve had to turn them off anyway and introduce more slowly for studio, simply because it was slow down production too much from the get-go.
+1 for making it configurable. Potentially each validator should be its own setting, so a studio can pick and choose as they go along.
I agree, for me to ‘persuade’ validators for quick publish test in makeshift Maya scene could always be a tad difficult.
Another thing is that many validation messages for poor artist leave a lot to be desired from the UX perspective.
I usually recommend that if one really needs to push through a model rapidly or a test publish to temporarily use a
pointcache publish over a
Similarly, for those who are in need of getting something shared with someone else (in Maya) with something complex they’ve already set up in Maya they can use a
mayaAscii publish - which is a close to raw “export selected” of the contents. This would really be a last resort, but could be a solution for when there’s the need to push something out NOW rather than in an hour if you get stuck.
But I agree. We’ve been using the strict models for a long time now and even here newcomers tend to get “bitten” the first few weeks by how strict it is. It’s similarly worse if we get models from elsewhere and we just need to propagate them but then notice assets from a library or another (even high-end) studio to not have those strict rules.
I’d set up the default so that “out of the box” model production is considered safe for a regular production.
I wonder if we could somehow set up a “poll” like list here of what everyone feels what kind of validations make sense. Just to get a sense of where the boundaries are for everyone.
Maybe having something like “settings template” that will trigger bunch of changes across pipeline could help. Leave defaults for strict and safe mode, but have something like easy mode that will disable more strict validators and set other thing to less demanding settings. Btw that is idea for additional Pype 3 Settings feature - with that it would be possible to quickly apply settings for project. For example color settings, etc. For 3d pipelines working with cad based models, you could apply “CAD settings” and it will configure specific options on project settings for example. Or difference between mocap animation sources and hand animated could result in different validators. Or different input/intermediate/output, burnin, slate settings, all applied at once by one click.
This is on the short term roadmap for the settings GUI in Pype 3 actually. With the reworked triggering of the plugins which is now fully live from the settings (meaning whatever you do in global or project settings is immediately applied to your scene publishing). It will be easy to add these configurable
collections of settings and have them applied to a project quickly.