The wolf is looking good, and thanks for adding the picture on the wiki.
I've never used Git on Windows. If you let me know a few more details, perhaps I'll be able to see what went wrong. There are instructions specifically for Github here
. Did you follow those instructions or did you find a copy of Git for Windows somewhere else? Also what happened when you tried to download FriendlyFruit? Did you get an error or did it just do nothing?
BTW if you fork the FriendlyFruit repository on Github
you'll end up with your own version of FriendlyFruit. You can add .blend files to the media directory in the copy, then send a pull request through Github when you're ready for them to be added to the main FriendlyFruit distribution. (That is the way you are 'supposed' to use Git and Github, but as I said before, it's your decision. You're very welcome to email the .blend files to me if you prefer, and I'll add them to Git.)
Absolute beginners can't contribute, no. Beyond that, I think it's important to distinguish between the core of the game, and developing content. The core of the game involves network protocols, digging around in .blend files, and other difficult stuff. The content is different. 'A player steps on a pressure plate so the monster appears.' The player being in a particular location is an event, and creating a new monster is an API call. Done right, this sort of thing should be a few lines of code only.
You can mod most of the Bethesda games like this, BTW. They all have their own scripting languages—Lua for Oblivion, and I think something custom for Skyrim and Morrowind. Morrowind ran on hardware that isn't so different to a Pi. (I don't know about the Fallout games because I didn't enjoy playing them!
The low-level code is all C++, actually. Panda3D's Python interface is a thin layer over a C++ library, because as you pointed out, Python would be too slow. The Python code is all high level and 'strategic' so the cost of using an interpreted language is not prohibitive. To be honest I'm more worried about memory use. I've done some measurements using Linux cgroups that suggest we could be okay, but until we have an actual port to the Pi, who knows.
I'm inclined to agree that Python isn't the best language for large projects, but there are plenty of examples where it has been done successfully. Blender is the obvious one, given the focus of this project, but you could also look at Django, Mercurial, and various others. The problem is that we didn't have much choice, given the selection of game libraries that were open to us. The Blender Game Engine is also scripted with Python, so that would have been no different. Gamekit uses Lua, which IMHO is worse. jMonkeyEngine would have given us Java, but I wasn't keen on using that toolkit because it's still in beta.
You can read about the protocol here
. It does reuse code, but it doesn't use an RPC library that provides everything out of the box.
Good news about OGLES. The release notes say that it is supported (otherwise Panda wouldn't have made my game engine shortlist) but it's nice to have it confirmed! It wouldn't be the first time something was supported 'theoretically'.
I'm glad professional animators have problems with wonky fur, because it's certainly bitten me, and sometimes isn't easy to fix. (A while ago I tried to create a hairy liquid by combining fur with Blender's fluid simulation. I thought it might look cool—it's something you could never have in real life! Unfortunately something didn't work quite right when the two things were combined, and the fur ended up sticking out in all sorts of funny directions.)
There is a difference between animation and games here, though. For example, the hair in Sintel was a physical simulation, as I understand it. You have a lot of strands, then you simulate the effect of gravity and wind. I'm not sure if anyone has tried to do this in real time for games, but we certainly won't be able to do it on a Pi!