[ODE] ODE .NET Bindings Performance Issues, Unsafe code, etc...

[Gonçalo Lopes]

Dear Daniel,

Thanks for the positive feedback. I'd love to unite efforts with the
other teams, but right now there are some divergence of opinions at
just what should be used to guarantee performance under the .NET
framework (marshalling vs. unsafe pointers), as Jason as suggested.

Right now there at least 3 different options to wrapping ODE to .NET.
There is the P/Invoke w/ Marshalling approach, the Managed C++
approach (James Rayne) and the P/Invoke with unsafe calls which is my
approach. During the next week I'll design several performance tests
to compare these 3 approaches, using the untouched C-API of ODE as a
control test, so we can also understand better just what are we losing
by just trying to use ODE in .NET.

I'll post my findings as soon as I have them and maybe then we can all
take more informed decisions regarding the approach we should take. My
early release of the wrapper was so I could get as much feedback as
possible early on.

As soon as I have some results, I'll let you all know. In the
meanwhile, if somebody gets to try my approach, be sure to mail me the
feedback.

Thanks to all,

Gonçalo


On 11/1/06, Daniel Huser <dhuser at 1eeurope.ch> wrote:
> Gonçalo,
>
> Great work! Sounds just like the thing, the .NET community was waiting for!
> Too bad we can't get that all under the hood of OdeDotNet! If we somehow could get your efforts and Terrys efforts into one and the same wrapper/binding, we'd have THE Ode.Net Bindings! :)
>
> I hope to get a chance to test it soon!
>
> -Daniel
>
>