Several questions about Newton functionality

[pHySiQuE]

You could use [code]SoftBodies and Co: It seems that you're not very familiar with NextGenration engine development

Oh look, someone is here to mindlessly repeat Epic's marketing messages as if it were their own opinion.

[JoeJ]

Haha, deutsches Selbstbewusstsein ?

I'm also not familiar with nextgen engine development, because i only do after-the-next-gen engine development to ensure i'm not outdated when i'm done.
...I do this for 20 years now

[FSA]

Oh look, someone is here to mindlessly repeat Epic's marketing messages as if it were their own opinion.

There is a smiley

I always try to put the newest things to our engine or develop own algorithms.

[Julio Jerez]

One of the biggest problems I have with the engine is the big dilemma. On one side too simple demos and people believe the engine is too simple, compels demos, even moderatly complex demos and people find it too hard. Even on this thread people say when you make the cloth demo, the expectation is that is very simple.

I do understand the problem, and I think the solution is having and asset manager system.
I have been thinking and working on this for a while but I never put too much focus on it.

Basically the asset system is the Tool to in the folder Model Editor.
This is not a modeling tool, of a game engine. It is an Authoring tool for creating and editing physic effect

Basically you use to put together effect, and you export then to the format, the game engine expects.
Things that people will be able to create and edit is stuff, like (but no limited)
-Vehicles,
-Players controllers,
-Contraption of articulated bodies.
-Destructions effects. (here goes the structural destruction)
-Authoring collision tress materials.
-Creation scene and compound collision
-authoring cloth
-authoring soft bodies.

These effect will be save on the model format, (the NGD format) this will remove the file format nightmare across different game engines.

For and application to use any asset then can write a plugging expert and importer.
The tool 100% plug-in based, and there are already some plug-in samples.

The weekend I added the autodesk FBX import/export, the already open the doors to handle, FBX, Collada, and OBJ files.

The first effect that I will implement in tool is the creation of Soft bodies
This week I am working on authoring the effect in the tool, and as soon as I fining I will wrote three export plugin. The first for the sandbox demos, the second for the OSG Newton wrapper, and the thrios for the ogreNewton wrapper

I believe that by using this strategy, I can make the effect as complex as they can by, and it will not affect how it is use because the end application will only be concern with how to load the effect in their engine.

[FSA]

The only problem with NewtonMesh on my side is to produce the mesh out from my vertex and texture data. But that is for the future and it will be a lot of work.

[Julio Jerez]

That's why the tool is plug-in based.
you will no need to use the newton mesh, you simply write a plugin for the tool, and you export the effect you want on your own format.

[FSA]

And what does the file contains? The physic effect of course. Vertexdata, Indexdata?

[Julio Jerez]

it is all xml, but you gen an API for parce it. yes each class contain the relevant data.

[elvencode]

The exporter and importer are for importing/exporting data to/from the tool, isn't it? So this means that in the game engine we need to implement another importer (with any logic one wants) to pass the data to the physics engine? Is the NGD format just for the tool or we can import directly to newton?
It's indeed useful to have a tool like that. Probably in the level editor of the game engine one will implement some kind of editor for physics settings but having a separate tool with the full set of options is good, just even to test separately some entity.

About complexity of the examples, i found always useful to have small samples focusing each one to some feature, then more complex examples combining different elements in something more similar to a real program. If the example is complex i think a small article dissecting the main sections and explaining a bit of "theory" so that one can understand what to do is enough

[Bird]

About complexity of the examples, i found always useful to have small samples focusing each one to some feature, then more complex examples combining different elements in something more similar to a real program. If the example is complex i think a small article dissecting the main sections and explaining a bit of "theory" so that one can understand what to do is enough

Me too.

-Bird

[pHySiQuE]

I don't have any problems with samples. I never use them and I think they're a waste of time. I just read the description of what the commands do and start using them. How much simpler can it be?

The structural collapse stuff sounds very interesting. I think it's much better to do something completely new like this that has never been done, rather than chasing after the particle stuff we keep seeing year after year.

[pHySiQuE]

We actually do have the ability now to use the marching cubes algorithm in a geometry shader to handle fluid rendering on the GPU. If I just fed the shader an array of points I can turn that into a fluid volume. That would be interesting.

[FSA]

Any news on

Code: Select all

void NewtonBodyRemoveSimulation(body)
void NewtonBodyAddToSimulation(body)

?

[pHySiQuE]

Why would you need that? Just set the mass to zero and pass on collisions.

[Julio Jerez]

I am still doing it. I was working on some other problems with the mesh an dI have a big change list to commit first.
I will try to get it going this week.

I believe this is a important feature. for example if the small destruction demos, you can see how the engine slow down with so many object has to be crate at the same time.
I can see how for example on can have different version of the same object and swap them under certain game conditions, right that the only way to do that is by doing complete create/destroy sequence.