On Thursday 12 January 2006 23:10, Ovidiu Sabou wrote: > 2. to be able to sell some of my games. > 3. a completely FREE engine for my FREE games or some way to pay a > REASONABLE FEE if I want to sell non-free games. Of course that a > completely free engine(with a non-copylefted free software license) > wouldn't hurt me at all.
I do think GPL hurts the progress of soya quite a bit with people being put off. > 4. a game engine which integrates well with 3d modeling > applications(like blender and 3dsmax). I REALLY wish I can ask for > artwork from a 3dsmax user or from a blender user or whatever and just > use it without too much hassle. Well this is where Verse should come into play.. but as that might be a while.. Id really like people to check out the CystalSpace project. In particular the blender2crystal blender plugin. The script works with your entire Blender project. Scene's are exported as different sectors, meshes are exported, duplicated meshes (alt+d) are correctly handled by just creating a reference to the mesh, lights exported, blender materials mostly supported. Also the script has an extra interface for setting CS specific options like collision detection. And then if you use CEL ( library ontop of CS for game oriented stuff) you can also start to set triggers and quest objects. If this wasn't alreadyamazing, CS comes with two utils, walktest and viewmesh. You can from the script hit the walktest button and a few seconds later you are walking around your level. These tools make working with CS a joy! Nothing I've mentioned CS does is that hard and its easy to extend with more functionality. Ogre too has similar tools but compiling it and setting everything up is just a pain. Both Ogre and CS use an XML based format - I guess outputting XML from any 3d modelling program is reasonably easy. Of course an editor built into a game engine has advantages too - the preview of what you see *is* what you get - but writing blender again doesnt sound that appealing to me! > 5. I'd also like a system like battle.net with a good Python API in > order to have internet gaming enabled in my game. The metaserver > should be hosted for public use for free games, and non-free game > developers should pay a fee to host their game rooms on the > metaserver, or run it on their own systems. I dont really know much about networking.. > 6. an engine that is EXTENSIBLE: I want to be able to add features to > the engine to suit my needs for my type of game if Python is too slow > or if I simply cannot do it because of lack of low level access to > system resources. As an example for extending the engine I can give > this one: If I want to implement my own tools for modeling and level > design, I might also want to add a new type of model and level to the > engine without modifying the sourcecode of the engine(it should work > by just plugging new modules for new features). Yes! But this is difficult! And also leads us to language choice etc.. > 7. an engine that just has the features that a game engine should have > > :) (see soya's features + some more ! :D) heh - reading through the Ogre docs they seem to have the philosophy "here are a bunch of helper classes - use them if you want but a serious game will reimplement them themselves" - which I see as the major flaw in any 3D engine stuff. We can make any number of nice classes but if your making something big or heavy on graphics you will want to write your own code specialised for the task. Of course if your still writing games for GeForce2 cards this isnt going to bother you much.. > 8. the software should be actively maintained of course I think this also comes down to an active website, documented source and generally getting people to think "I could help with that". > One way of achieving this is by writing a 3d engine from scratch > strongly oriented towards python(though I don't know what's there to > orient). This is the soya approach. > Another way is to use an existing GOOD engine and write a GOOD wrapper > with GOOD documentation for Python. I don't know if such engine > exists. > The third option, if the first version is not realistic(there is no > fun in forcing the use of Python in a game engine I believe) or if the > second option doesn't exist, is to write a new GAME ENGINE(and not > just a good looking but useless 3d engine) in a popular(many libraries > to choose from) compiled language with good performance(C++ eh ?:) ) > with Python in mind(this way ruby or other interpreted languages may > have wrappers for it) that has all(or some, the most important) of the > above mentioned features. swig and c++ must be what your talking about ;) I dont think pyrex is good enough at the end of day. The language is very nice to use but I think it brings more problems than it solves with a large project like soya. perhaps swig, c++ and some dynamicly linked plugins? ;) > There are many more things to say but that should be enough :) lol - yes - im exausted
