Re: [boost] [rfc] Boost Graphic/Game Engine

From: Michael Fawcett (michael.fawcett_at_[hidden])
Date: 2007-07-17 14:17:06

On 7/17/07, David Bergman <David.Bergman_at_[hidden]> wrote:
> That would be interesting, but, honestly, what - if any - pieces
> would be relevant to Boost?
>
> In general, there are a few areas that might be concrete and
> reapplicable enough to warrant Boost inclusion (or at least discussion):
>
> 1. Graphical user interface library. Such as win32gui for all (well,
> at least Linux, OS X and Win32..) platforms. There are a lot of
> potential input/inspiration in the game developer communities here.
> The GUI of a game is general enough (buttons, overlays, text,
> scrollbar...) that it might be isolated.
>
> 2. Geometry. We already have uBLAS, but there are definitely other
> matrix operations that would be needed for a proper 3D (sub) engine,
> and other non-matrix geometric operations.
>
> 3. 3D Rendering. Would it be possible to build an Ogre-like (in
> functionality and scope) layer using modern C++? I would probably
> start with wrapping one of these "Ogre" libraries.
>
> 4. AI. Well, the AI in games do not really have time to use advanced
> AI, but there are a host of interesting AI models one can Boostify,
> such as support vector learning (or other kernel methods), GA or at
> least a basic perceptron library :-)
>
> I have a hard time visualizing a Boost.GameEngine library, but it
> could be a separate effort, using Boost and its principles, and
> *perhaps* being able to borrow the name "Boost" somehow, but that is
> obviously up to Management.

I think that's a very good, obvious break up of what is involved.
Each of those pieces is very applicable to areas other than games.
For instance, I use path planning, but not for games.  I also use 3D
rendering, but not for games.  Same with geometry and matrix
operations...I'm pretty sure anyone in Robotics or Aerospace & Defense
would be able to make use of the above components.

--Michael Fawcett