Hi all;

- Torcs-NG continue its way with native C/C++ for the core code,
  and ports its tools to native python where interesting and valuable

Blessing Python (and even encouraging it) wherever possible would be great! We are now starting to coalesce around the idea that TrackGen will mostly disappear, but in any case having tools be in Python would make it a lot easier for us to grok and re-use, and easier for you to move forward with. (One of the benefits of removing complexity as we are doing is that it makes it easier to catch up to you.)

- Torcs-NG works at delivering a more and more modular C/C++ API
  for any other project to integrate parts of it the way it likes,
  whatever its language (using SWIG, Boost.Python, or ...)
  This API and its evolutions are negociated with OpenRacing,
  as well as the needed changes to its file formats, provided
  they are fully compatible (no feature loss)) with its current
  art-work base (cars and tracks) and code base (simu engine, robots, ...).

For us, Torcs is modular enough, and we can grab any C files we want.

- OpenRacing explores Python on Mono, as a replacement of C# on Mono,
  (and particularly checks the FPS likely issue)
I will put this on our list! For the next few months we will stick with what we're just getting something working end to end and we don't want to go backwards while trying to add features. But we can switch and we have no specific type of religion here, other than the importance of thinking long-term and using productive tools.

  using Torcs-NG through its API (with SWIG wrappers, or Boost.Python, or ...),
  its code being at the same time a glue between Torcs-NG modules,
  may be full replacements of some of these modules (ssggrap => Ogre3Dgraph,
  Python robot, ...), and likely other features that are not and wont be
  in Torcs-NG.

Right now, the only big chunk of code of Torcs that we plan to call into in the medium term is the simulation engine. But even there we are now considering generalizing to support a different track model, and maybe plug in a new physics engine. We will have a migration path for all of your art and your robots. We are trying to use as much of your code as possible.

- In a medium term (some years ?), Torcs-NG might move smoothly
  to Python on Mono if this platform has prooved to keep FPS
  at a high level, and if the Mono community prooves its ability
  to resist to M$.

2 years is a very long time for us as we only plan work our work out a month or two in advance, and you guys are only a few months old yourselves. I also think you could consider things on a feature by feature and dev by dev basis. At some point, we are going to want to break the robots badly and add a lot of features. Changing the friction of the segments in the middle of the race is the smallest change I could come up with ;-) It would be great to be working together when that day arrives.

Kind regards,