features
- OpenGL based
- 3d models with skeletal animation
- raytraced shadows for lights
screenshots
Some old pix here of the 3Dfx version.
download
You can get the old Internet demo here. It's really basic compared to
what's cool these days. Keep in mind this was written for a Pentium 133/3Dfx v1
system.
useful links
|
| | | 05-mar-2004
friday
| trying to get opengl to do the heavy lifting
Yep, until now I was doing all the 3d math in my code and passing the final
coords to OpenGL. So now I'm converting over to letting opengl do it (and
therefore only having to calculate my 3d into global coords)
I found this article very helpful.
Oi, it's sad though... my homebrewed 3d isn't quite coordinates-compatible
with OpenGL. I have the Y axis as positive going south, instead of negative.
It's just more intuitive that way to me, especially since MW started out
as a 2d tiled scroller design.
I guess I need to dig in there and make everything conform. I miss the
flexibility of doing my own projection. (I don't miss the performance hit tho ;)
|
linkme |
|
29-feb-2004
sunday
| refactoring
Just been refactoring things still, converting structs to classes with
real constructors/destructors etc. Now that 'Room' is a class with
virtual void render() I can support both my old map format and whatever
new one I come up with. I can even make it a plugin. Yay polymorphism!
complexo
Check out this
graph. (from this article). Heh. I have like (maybe) half of them
started or done, so that's kinda cool. Some of them I won't need, either.
The biggest hole, what I'm really missing, is the persistent db type shit.
Luckily for that I can crib ideas from MUSHes etc., which I'm pretty familiar
with. I'll probably do the same global DB for all attributes, with flags and
such. It's still gonna be a chunk of work. I'm thinking of using sqlite, although it seems like overkill for
now.
|
linkme |
|
21-feb-2004
saturday
| cg rOXO
I just got and read through "The Cg Tutorial." Cg is awesome! It lets you
program the whole 3d pipeline, but with half of it happening directly on
the graphics card, and with single-instruction matrix ops and whatnot.
Very exciting! I can't wait to get that working.
It seems to do a very good job of letting the programmer
think in C while he's actually working on a vector architecture.
You just write a little nifty program that does whatever it wants to
either a vertex or a fragment (pixel, if it passes the clip test). Newer
cards can even parallel process these programs.
The only real downside is that it's kind of nvidia-centric, although it really
tries not to be. Well, mostly in that there's no bundled profile for the
high-end stuff in other cards, like my FireGL.
Yes, I'm writing MADWORLD to target FireGL.
They sure seem impressive now, but in 3 years it'll just be middle of the road,
and I'll be half done ;)
|
linkme |
|
17-feb-2004
tuesday
| spherical harmonics
I stayed up late trying to grok this stuff via this
paper (PDF format, 1.8 megs, local mirror). Not there yet.
|
linkme |
|
16-feb-2004
monday
| update
Also GL_EXT_fog_coord. I now have the beginning of my 'fog-of-war' effect.
Oh man, I found some seriously stupid old bugs. They've been in there forever!
Nice to be rid of them.
|
linkme |
previous articles...
| | |
|
|
