I haven't gotten around to updating the skelengine page yet. The screenshots you see are from one of our very early builds. We have since added Quake3 map and Doom3 model loading capabilities. Of course, when I put it that way, it doesn't sound like much. However, a lot of features were added in the process. Such as:
- An abstracted filesystem, which allows us to read from zip files transparantly
- Bezier Patches for for the Quake3 Curved surfaces
- Static lighting via lightmaps
- Quake3 shader parsing and rendering (we convert them to GPU shaders)
- Basic post process effects, such as blooming
- Skeletal animation
- Normal and specular mapping
- Volumetric stencil shadows for models (we don't have a unified approach yet)
- Quake3 brush geometry collisions, including proper hadling of stairs
- Animated textures
All in all, I'm pretty happy with our progress. We are currently overhauling the materials system to be more shader friendly. The goal is to be able to generate shaders on the fly based on a material definition. There are all kinds of sticky issues inherent in a system like that. Such as being able to splice together a handful of shaders without causing naming collisions. There is also the issue of how and what to pass from DirectX to the shader.
This week saw the release of another game that I was involved in, Hanna Montana: Music Jam. This is the game that I was working on when I left Gorilla back in December. I couldn't mention it before since it was still an unannounced, in-progress title. To be honest, I am still a bit hesitant to claim it since I wasn't with the company for the entire development cycle. However, I was very involved in the shaping of this game and in the initial heavy development. Because of that, I can't help but feel a sense of "ownership" for this game.
Geoff and I have been working on skelengine again. We use STL for a lot of things in the engine, but we never really considered the performance impact of using a third-party implementation. Supposedly, the Dinkumware's version of STL performs well enough for most things, so we stuck with it. Eventually, I ran into a weird performance quirk where it was taking longer than I thought it should to iterate a 5000 element vector. So for kicks, I integrated STLPort into the project. I half expected it to perform on par with Dinkumware, or maybe slightly faster. What I didn't expect was a 4.5X speed increase. Moral of the Story? Don't believe everything you read on the Internet.
A lot of things have happened in my life since my last update. The most important of which, is that I am now a maried man. The evidence is at http://shallowadventures.com. Yes, that's me. Yes, I am wearing a tux.
I want to start updating this site more frequently, so keep tuned.
Well, it's a new year and I have a new job. My stint as a game programmer was fun while it lasted, but sadly it is over
for now. I did manage to be involved in the development of three published titles while I was there. They are
Disney's The Little Mermaid: Ariel's Underseas Adventure, Strawberry Shortcake: Strawberryland Games, and The Cheetah Girls.
I did the most work on the Little Mermaid, out of the three. You can click on the images of the games to get more info
about them. While you're doing that, you should check out Opera Slinger. One of my
buddies from college worked on it, and it's in the running for the indie best of show at this years GDC!
I am really looking forward to The Dresden Files, a new TV show on SciFi channel. If it's anything like the books, it should be a damned fine show. Of course I am also looking forward to the return of Stargate SG-1 in March, and I suspect most of you know how cheesy that show is.
In parting, I feel that I must talk about Eragon the movie. Do not see it. It sucked, and it sucked bad. I do recommend you read the books though. They were quite good.
