On 7/17/07, Marcus Lindblom <macke@yar.nu> wrote:
(Sorry for striking a negative tone here.. but.. )
I don't think it's negative.
Just imagine, what needs to be done to code up a good game today, like your average MMORPRG or something. :)
There are numerous such parts: * content/anim/rendering (best left to others) * networking (asio + levels on top of that) * low-level server work (clustering, load balancing, etc) * high-level server work (accounts, chat, database) * active framework system (game object mgmnt) * auto update/versioning * state distribution * scripting (python? lua? lang-bind?) * AI (state machines, micro-threads, fuzzy-logic? learning? etc)
Do you not think each of those components is useful enough to be its own library? I for one would love a path finding library, a load balancing library, state machines (already being worked on in Boost), and fuzzy-logic library. There's talk of a Boost Database library as well. I agree that a Boost "Game Engine" is not appropriate. What *is* appropriate, IMHO, are all the tools necessary to create a good game engine. You listed some of them above.
Consider also, that development hardware is not accessible to Boost (i.e. GameCube, Wii & PS3) because it's very expensive.
I know of of developers that have access to all of those systems, and also use Boost.
(Let's not forget that the boost communitiy can't find a consensus on how to best design a math lib for 3d-graphics. Now imagine a whole game engine. :)
I alluded to this in one of my other posts on this thread. In summary, my enthusiasm is not for an entire Boost game engine, but rather to stay in the spirit of Boost, and offer libraries that aid in the development of a modern game engine. --Michael Fawcett