Newton Dynamics is now Open Source!!!

[TheMrCerebro]

A few days, into the website of Newton Game Dynamics, i enter to download the latest version of its physics engine and I have found that now is open source and best of all ... zlib licensed.

We can modify or improve it as we like, that if, without breaking the license.

In addition, it is now compatible with the license of Irrlicht...
possible future release of Irrlicht with physics engine?

Sounds good

Link:
http://newtondynamics.com/forum/downloads.php

[serengeor]

In addition, it is now compatible with the license of Irrlicht...
possible future release of Irrlicht with physics engine?

I don't think that Irrlicht really needs a built in physics engine, as some people prefer other physics engines, and also it would make irrlicht harder to maintain.

[Virion]

this is really a good news!!

[pippy3]

I'd rather have irrlicht have a wrapper, and perhaps some optimizations for the newton wrapper.


But this is great news.

[bitplane]

This is really good news.

I'm not sure about Irrlicht integration, but a unified interface between the different physics wrappers would be a good step towards that though (unlikely unless everyone cooperates to define it)

[REDDemon]

I had the same idea of a unified physics integration via a unique header.
(just see irrsolid thread if i'm lying XD).

The problem is that if you are going to integrate physics with an interface for all the physics engine first you must learn from all existing wrappers and create a middle way (so people are much more comfortable with the new interface). That's a big work and requires much programming for every physics engine.(of course just providing a interface is really the simplest part, then some people need to write phys drivers also) Integrating this with irrlicht animators and AI steering behaviours would increase the final speed but this requires adding also a bit of AI to physics (soon some AI/physics classes will be released with irrsolid).

[bitplane]

of course just providing a interface is really the simplest part, then some people need to write phys drivers also

Providing an interface is the hard part, it's much, much easier to write an implementation for a special case than design a good, clean interface that works for every case.

You'd have to research every available physics library and create the simplest, most lightweight and intuitive interface that supports the broadest range of features possible, yet remains extensible, without having loads of special cases. If it were to eventually replace Irrlicht's collision manager then it would also need to simplify rather than complicate Irrlicht's design. That's a pretty tough problem.

[Granyte]

a wrapper or even a tutorial that works would be a good start cause for now every tutorial availible on the web with irrlicht use irrlicht 1.31

[Lonesome Ducky]

This is good news, but I think I'm going to stick with Bullet still.

[pippy3]

This is good news, but I think I'm going to stick with Bullet still.

+1 for bullet.

Though now newton is open source I may have to give it a go.

[REDDemon]

bullet is THE physics engine now. but others engines are still good.

[Radikalizm]

If only Havok would follow suit..
That's probably never gonna happen

[randomMesh]

Pretty much too late. People use other physic engines now.
To be a weathercock sucks.

[spock]

it's not too late. newton still has its qualities over other engines like bullet. and now the newton author will have more time to concentrate on new features like softbodies instead of wasting time with multiplatform builds and so on.

i think this is great news!

[lpersona]

bullet is THE physics engine now.

irrlicht + bullet + windows xp sp2 64bit = CRASH