[ec2m] Physics engines: MathEngine/Karma/Vortex v. Havok

[Taylor, Tim]

[I've been away for the last few days, but have just
been catching up with the latest posts in this 
thread.]

I'm very interested to hear that the new release of
DynaMechs allows/will allow multiple objects to be
simulated (Richard - do you know when this will happen?
I had a look at the DynaMechs webpage, but couldn't
see it mentioned).

One of the reasons I haven't looked at DynaMechs
(and the same goes for Darwin2K) in much detail is
that I have been wanting to move on to co-evolutionary
simulations between multiple creatures. It was not
possible to do this with previous versions of
DynaMechs, which could only cope with a single
articulated system.

On another note, I have had a further email from
Critical Mass, with more information (from their point
of view, of course!) on the relative merits of
Vortex v. Havok. This is what they had to say:

> Havok has some good features and no doubt they have some
> excellent developers. The product has been developed to
> offer a broad range of features, such as cloth,
> deformable bodies, vehicle dynamics, rope, etc.
> They do a little bit of everything and do them reasonably
> well. However, getting really stable and robust features
> is very challenging and requires a lot of development work.
> What we've done at Critical Mass Labs is to go
> deep into 2 key features: dynamics and collision
> detection. The result is a very stable and robust engine
> in these 2 areas.
> Over time, we will likely develop other functionality, and
> add these features to future versions of the toolkit.
> Another difference between the two packages is the
> interface. The Vortex interface is written in C, and uses
> simple C data-structures. Some people prefer working with
> this than in a particular environment like what Havok
> offers. Another functionality that may be of interest is
> that we offer support for WorldToolKit by Sense8
> (www.sense8.com).
>
> A couple of other points to note:
> we've released a beta version of Vertigo, our physics plugin
> to 3ds max. We are upgrading this and will have a full
> release version available by early November. We are also
> working hard on our next major release since Vortex
> was introduced back in March. It will have bug fixes and new
> features all based on user feedback. (I think I mentioned
> this in my last e-mail to you.)


Tim

Tim Taylor_________________________________________________________
Artificial Life Research Associate, IC-CAVE (http://www.iccave.com)
University of Abertay Dundee, Bell Street, Dundee DD1 1HG, Scotland
mailto:tim.taylor at abertay.ac.uk             tel. +44-(0)1382-308959
http://computing.tay.ac.uk/timtaylor/       fax. +44-(0)1382-308627


> -----Original Message-----
> From: Richard Reeve [mailto:richardr at dai.ed.ac.uk]
> Sent: 20 September 2001 2:45 pm
> To: Taylor, Tim
> Subject: Re: [ec2m] Physics engines: MathEngine/Karma/Vortex v. Havok
> 
[...]
>
> Jut a quick note, thinking of open source dynamic simulators based on
> Featherstone's method - Scott McMillan's DynaMechs
> http://dynamechs.sourceforge.net/ has been around for a while 
> now and is
> in the process of another major release which allows closed loops and
> multiple objects being simulated - I haven't used it for a 
> while, but I
> found it quite useable at the time, and if you're looking for an open
> source platform, this would be a good place to start...
> 
> Cheers,
> 
> Richard Reeve.
>