Ragdoll Pack Is SHIPPED!!

Well, yesterday was my birthday, and on the same day I saw my very first, completely independent, game-related product go out the door. Seems like a good time to look back and reflect a bit. It does tend to make one think. I just spent something approaching six months of my life working on a $20 product, that probably won't ever pay me back for more than a decent month in any real job. Ha ha. Life can be funny that way.

But, hey, nobody ever told me to be an indie game programmer. :-)

And, even though I watched not just the numbers, but the numbers of digits in my bank account plummet during that time, I can't feel too terribly bad about it. Six months ago the entire engine/ts directory, as well as all that gobbledy-gook about ghosts and bitstreams, and of course everything about the RTS kit, were pretty much "black boxes" as far as I was concerned. Were it not for Ben Garney, Tim Gift, and the other wise and generous souls in the GG office, I'd still be spinning my wheels on that stuff. (Thanks you guys!)

So, it wasn't really six months on the $20 Ragdoll Animation Pack. It was six months on the pipeline from the Open Dynamics Engine into Torque, including: RTS pack basics, Torque client/server architecture, DTS format fundamentals, and don't forget the constant ODE shock therapy!

Anyway, enough excuses. The important thing is, I DID IT! Today, my life is beautiful, birds are singing, flowers are blooming, and if the proverbial bus takes me out tomorrow, at least I'll die happy!

So what did I do? Ever since Tim Gift dropped the magic words "Open Dynamics Engine" at my feet, at my first IGC in 2003, I've been absolutely possessed by the need to link it up to the Torque Engine, and more than that, to create animation with it.

I should back up a minute, though. I've been trying to make my own game since at least 1997. Along the way I've had a lot of great ideas, but one that just would not ever leave my head is this: Trying to model character motion, without a system of gravity and collision detection to handle the realism part, just sucks. I don't care how many people I see out there doing it, with IK or whatever, if we are creating a run or jump or fall animation by dragging a character's limbs through what we think are realistic trajectories, but we are not modeling the actual gravity and forces of the imaginary situation, then in my humble opinion, we're just wasting time. No offense to all the artists out there who do a miraculously good job of pulling this off believably. When it comes to animation in my game, I don't want art, I want a simulation.

If I had a simulation, not only of "natural world" forces like collisions and gravity, but also including some kind of AI and motor forces in the character, then someday I could theoretically generate ALL of my game animation without an artist having to lift a finger, except maybe to make corrections and improve the feel. I'd also have a pipeline with which to effortlessly create movies and animations, and, ultimately, to get rid of the "saved sequences" animation model entirely, instead having live virtual creatures animating themselves in realtime.

This is the primary will-o-wisp that I've been chasing after with glazed eyes for all these years. Unfortunately, before Tim told me about ODE, I was going to have to write the entire physics engine before I could even start my real project, which is generating the animation. While that goal could still be considered within the bounds of "humanly possible", it definitely comes under a different time frame. (Like, in Brian Ramage's words, "go back to school and get a Ph.D. in physics.")

But even with ODE in the picture, it's still not exactly cake. ODE is far from simple, and neither is Torque. Especially on the ODE side, I've found all kinds of stability issues and unexplained behavior that has required bizarre workarounds, and all I've done is the simplest little bit I could think of. I'm not doing any AI here, I'm just modeling some more-or-less realistic joints and collisions objects to match Kork's skeleton, and then flinging him. Later I'll be putting muscles on him and handing him a genetic algorithm with which to learn how to do what I want him to do. (In case you're wondering, this second part of the project is prototyped already, that's what I was working on when I hit IGC last year.)

My most immediate future plans for the project are to deal with the obvious, glaring, painful issue that people have their own player skeletons, and probably aren't using Kork or Elf Girl or Blue Guy in their games. I can't make dsqs for models I've never seen, and I can't very well go through hundreds of skeletons and custom generate animations for each one, either. Fortunately, right about the time this realization started to really sink in, Jeff Gran happened to pop up right behind me (literally) in the Garage Games office, with a very cool idea. If you missed it, check out his plan, I'll let him tell you about it. But trust me, we're working to solve this problem.

But anyway, this plan is fast getting out of hand. Just wanted to touch base and say "Thank You!" to everybody who helped me wrap my dim-witted brain around all the problems it's had to wrap itself around so far. I'll shut up now and give you time to get your credit card out. Go check out the Ragdoll Animation Pack Product Page, and if you have any interest in seeing another struggling indie pull his rent together for another month and keep working on radically cool lines of development for Torque, you know what to do!

Till next time,

Chris Calef