← Back to team overview

ubuntu-gaming team mailing list archive

Re: Fwd: identifying obstacles in ubuntu gaming and pushing it forwards

 

2009/6/17 Haroon Khalid <haroon@xxxxxxxxxxxxxxxx>:
> did someone say making a direct X type of lib is a bad idea? What else is
> there you can do? The opengame art site has hardly any solid art.

Yes I think that a DX type lib is a bad idea because we have the
equivalent libraries already. I fail to see what benefit developing an
all encompassing API will actually bring. Look at it this way, if I'm
developing a networked desktop application with a database backend I'd
research the most suitable GUI API, the most suitable networking API
and most suitable database connection API for my language and
circumstances. It doesn't matter that all three will use a different
syntax because chances are they will take up a minimal part of the
entire code of the application. Also that kind of modularity lets me
choose one network API over another for my particular situation.

Games are exactly the same, DirectX or OpenGL/SDL/OpenAL only make up
a small portion of any reasonably complex game. The majority goes on
game code: AI, file IO, logic, scene managers, UI elements etc. etc.
and the most reasonably designed games will have abstracted the lot
away coz (for example) you won't find DX on a Wii. There is very
little benefit writing some kind of wrapper for the different
component APIs (OpenGL et al.) we have because it's not really an
issue.

And then there is the age old saying.. "Write games, not engines"
although you can obviously write a game using the APIs directly, the
work has already been done a zillion times in various open source
engines (Q3, Ogre, Irrlicht, Horde3D, Crystal Space etc. etc. etc.) so
just use one of those instead of reinventing the wheel.

That said, if you aren't using an engine, there are *some* gaps we
could fill with C libraries to fit in with the GL/AL/SDL/DevIL stack
to better match DX in features, including:

1. A 3D font output library
2. A 3D math library ( I tried this one: http://www.kazade.co.uk/kazmath/ )
3. An FX framework library
4. A 3D model library (like DevIL for models)

DX has those above 4 things included. But I reiterate I don't believe
code libraries are the problem really, I think it's a lack of
collaboration of resources and a lack of a central place on the
Internet for the open games community for managing
code/models/texture/sounds/teams and games.

Luke.



Follow ups

References