i am thinking what other things you would like to see in the future. from small details like making the interface darker to having network support.
having moveable logic bricks like virtools would be a good start.
also, who is programing and playing with the blender source, specially the game engine code? it would be nice to send them our requests to see what they can do.
I’d like a duplicate sensor or acuator within the object, so you could make a duplicate of one sensor and just change one thing. Making about 20 property sensors and having to go add, change it to property, stick in the name, and finally the value, is really anoying. It’d be nice if i could just make a bunch of copies and just change the value.
The stuff in the proposal was nice (see my post on blender.org for more of my comments on it), but there are a couple major problems
It is difficult to turn something like “make aibot[n] follow the terrain to get to waypoint[p]” into code (likely c) that will work for most terrain meshes … It was implied (loudly?) that this could be done
Oh, and more python control would be nice. I would like to see Blender.BGL not be sandboxed (so that I may use glViewport), but I can see the security issues.
Maybe what is needed is ways to have
a) multiple cameras in use in a single scene
b) viewport control (where in the window the things are rendered to)
This would allow split screens (and MechWarrior 3 like zoom-windows)
Is it possible to access individual ipo elements in run-time python? (ColR, ColA … don’t seem accessable other than through an ipo actuatior) If not it could be useful
One feature I’d love to see would be the ability to add dynamic weather(ie rain, fog, etc) and day/night cycles to games.
I thought of an interesting system that could possibly be used to do this, actually…
If you built a model of a game world, and gave it regular, pre-render-type materials and textures with specularity, radiosity etc, could you have a system which pre rendered each face of this fully and isometrically, camera perpendicular to each face, ie with all the pre-render effects that the game engine can’t do, then converted all of the images produced into UV textures and applied them to the object ready for use in the game engine?
One upshot of this would be that you could very easily then animate a sun lamp and stuff, and automatically make an animated texture for the whole level, to do day/night lighting, etc.
I don’t know if anyone would find this useful or not, but it is, I think you’ll agree, an interesting idea.
It’s kind of slow and the lighting is a bit weird. Having the whole scene realtime lit probably slows it down a lot, and using uv animation would be a pretty neat effect:) The easiest and best way to animatie the ground though would be the way I animated the sky: color ipos. You can start at white, then as it gets late go down to black, then go back up.
Anyway, theres a lot of things I would love fixed. Here’s my top 10 blender fixes in no particular order:
More advanced rendering engine with different methods of culling, options for bsp rendering, etc
Speed improvements such that generating 500 super simple objects doesn’t kill blender
Realtime access to everything from python. And I mean EVERYTHING
Better realtime action support
Putting python elsewhere than a controller
The ability to do advanced realtime techniques such as shown in current games: bumpmapping, reflections, and real shadows
More physics options, more collision detection shapes
Being able to load external textures, music, models