Mantaflow, not a nodal workflow? Wait...let's take time to think before it lands in Blender

I think it’s important to start discussing about Mantaflow Blender integration now, because later will be too late.

Blender aims at being “everything nodes” and currently we have the shaders being nodal, animation nodes, and particles nodes being designed, things are going the good way.

…But I can’t hide my surprise when I see that the new Mantaflow implementation being put in Blender is not nodal workflow at all.

I want to say stop right now about going this way and ask that devs and community to talk all together first about how we want it to work inside blender. Having an old school mesh attribute based workflow and not a procedural nodal system is a big no go for me !

And I don’t think it’s good neither to implement it a in way that will be depreciated very soon because nodal will come soon after, better to do things the right way from scratch.

■■■■, no!! Implement in blender now, we have been waiting years, we don’t need to wait other year or two more.


No it should be implemented in any way the implementer wants, in the easiest way for them to do so. Once in, and working, then the interface can be improved, changed, or replaced with nodes. In fact it might be a different person who adds a nodes UI to it.

Don’t worry Mon. Every little thing will be alright in the end.

Users who just want features fast and who don’t care if it’s being added the good way or the whatever way are often users that won’t actually use the feature on projects.

Sofwares used as examples of quality are the ones that don’t fast add features just to feed the crave of some impatient nerds.

So no. I can wait for blender 2.83 even 2.85 if it means that mantaflow Fx will share a polished workflow that follows the same logic as the particles one and that can interact with it within the same graph. And I don’t want crappy extra attributes everywhere because of a depreciated workflow that lasted 3 versions before being abandoned in favour of the nodal one.

Seriously, that is absolutely not what happened.
Mantaflow integration took years. First GSOC was in 2014.
In 2016, there was a serious work done on UI of modifier to make it understandable by Blender users used to its smoke and fluid simulations modifiers.
In 2017, it was made compatible with current particle systems.
in 2018 and 2019, it was redesigned to respect 2.8 UI principles.
And now, it works.

There is a problem with its display that don’t want to show a solid mesh in solid view.
But except that, it is mainly ready.

The fact that this UI can end up in 2.82 does not mean that nobody will work on nodal version of it for modifiers nodes release.
People will be happy to use Mantaflow in 2.82, 2.83, 2.84, 2.85 and at same time, discussing its modifiers nodes UI version that will be developed for a 2.86, 2.87 , 2.88 or 2.89 release.

We will not rush things. But currently, particle nodes are not ready to be merged in 2.82. Maybe they will not end up in 2.83.
And modifier nodes are supposed to be a WIP only when particle nodes will have been done.

Anyways, we would to have modify mantaflow UI to work with particle nodes UI. It could happen in particles node branch or in mantaflow branch.
There is no reason to delay integration in master of what works and don’t continue work on a nodal interface in parallel in another branch during several years.
If you don’t want to use it, just don’t use it. But don’t annoy everybody because of your preference.

Current attributes are not supposed to be crappy but useful.
They probably will still be useful in 3 or 4 years.
If they are replaced in mantaflow code during upcoming years, there is no reason that mantaflow maintainer don’t do the work to replace them.
It does not matter what UI is used.
If mantaflow code changes, somebody will have to work on an update in blender.