Anyhow, back to the topic at hand, which is design docs. I’ve seen a few now and have attempted to write a couple, and I think I’m ready to conclude that I don’t like them. At least not for my personal stuff. Maybe if a design doc could be written as a kind of surreal poem. Hmm. Anyway. I think the issue is, for me, the creative process varies and mine doesn’t work that way. Some authors create intricatley detailed outlines and notes and character profiles and world histories in order to keep a story straight in their head. Others just sit down and let the story take them somewhere unexpected. 

And I was thinking about that process for me, and for games. I think what I have, more than anything else, isnt a specific implementation in mind, so much as it is a feeling of what it would be like to play the game. If that makes any sense. And so the rest is crafting a game that creates that feeling. 

That being said, here’s the game design. I decided one day that I really enjoyed playing company of heroes, but only certain parts. The traditional RTS trappings of a build order, a tech tree, using buildings to create units, securing resources, spending resources on units… it all took away from the stuff that was, to me, the fun part – which was using the units. And yeah I know I’m not the first guy to want something more like a real-time-tactical (rather than strategic). In fact, Dawn Of War 2 looks to be exactly what I just mentioned. But regardless. I decided I wanted to make a game where the focus was on the tactics of using the units you’ve got, that is, flanking, encircling, baiting and outmanuevering enemies to win – as opposed to memorizing a build order and then just tossing as many units as you can build at the enemy. And then at some point I came up with the idea of setting this game on the outside of a sphere instead of on normal terrain because, well, I thought it would be cool and the specifics of the camera were a fun engineering problem I wanted to solve. 

Originally my plan was to limit the player to only a very small number of units, something akin to the Heroes in warcraft 3, but I realized that with the essentially planet-wide scale, it wouldnt work very well. The surface area of a large enough sphere (too small and it becomes bitchy to navigate) is too big for only a few units. And so spawned the basis for the current design. The players controls a vast army that acts as a single unit. However, this army can be subdivided into smaller sets for tactical purposes. For example, you could split off a squad to flank around an enemy position and have the remaining units provide cover fire. All sorts of stuff like that. The idea from there evolved into a way of taking a squad of units and splitting them into sub-squads that can be used tactically, and then later, if the player wants, recombinging squads. The UI/controls for this are pretty simplistic although they seem complicated. A set of possible commands given to a squad, say, move to location, fire at position, etc. Then, the player just presses a key to denote that a command is intented to split a squad. So if you select a squad, press E key (E here would denote that you want to subdivide), then give an order to move around the side of an enemy to attack them there, half the squad would split off and follow that order. 

And the real fun comes later, in trying to think of cool/interesting units and weapons that fit a game played on a planet scale. Satellite weapons come to mind here. 

The point though, is that none of this is ever set in stone, nor should it be. You come up with an idea that seems cool, implement it, and if it is, awesome, if not, try another. Rapid prototyping (for some value of ‘rapid’). No other way to do it, I think, because a great idea could suck as gameplay and vice versa. One of my favorite authors once said that he feels good about a book when he’s thrown out way, way more pages than he’s kept. I like that.

So here’s a design issue I’ve been tossing around. Consider the following requirements: the atomic UNIT, which cannot be subdivided. A SET comprised of some N number of units. If the size of the SET is greater than 1, it can be subdivided into two sets. And so on and so forth. Therefore each unit is itself considered a set of 1. 

Simple enough. Lets say we require the ability to on-the-fly subidivide and re-combine sets at will. What kind of data structure can support this? My initial idea was a list of lists. You’d start out with a master list, the set of all units. To subdivide. you create a second list, add half the units to it, remove half from the first list. Two obvious problems with this: 1 – adding and removing from lists is slow, 2 – creating/deleting new lists and the required allocations is no bueno.

So whats a better solution?  Probably best to identify a few constraints. Units need to be able to refer to their set, and set’s need to be able to enumerate their constituent units. Perhaps each unit could store an identifier for its associated set, and this way units are never “moved” and no memory/resources wasted. But then how do you go from a set to its units without searching the list? Could have a set object that stores a list of pointers back to its units but that seems to have the same problems as the first solution. Maybe searching isnt so bad.  If the units can be ID’d and then sorted, O(log n) search time should be plenty fast.  I don’t think I’d ever be dealing with set sizes large enough to make that an issue. 

Anyone have a better solution?

- A fully functional camera which behaves pretty much exactly how I’d imagined it would. You can navigate across the surface of a sphere with WASD as you would in an FPS, except the directions take you around the sphere instead of on a given plane. Pretty simple there. Also, you can look about with the mouse also similar to a FPS camera, except horizontal movement is a rotatition about the camera’s up vector rather than the world up, since there is no “up” while on the surface of a sphere, or at least, “up” is poiting straight out from the center of it. Anyway, it works pretty intuitively. 

- Importing and texturing 3D models. Simple, but yeah, it’s a step. 

- Particle effects. Made my first particle effect as an explosion of orthogonal rings. Check it out. 

- Shaders. Just tip of the iceberg here but I’m learning. Irrlicht makes it pretty easy. 

- Mortar type weapon. You can left click and the selected unit will fire off a sphere, hold down the button before releasing to modify the angle/arc of the mortar (ala gears 2). 

I think this is about it. I’ve got an idea brewing that should be pretty sweet but I won’t spoil it here. 

Now for some games. Fallout 3. Man, what is wrong with me? I just really have no desire to play it right now. I think, maybe, it’s a time investment thing but I’m not sure. Oblivion was so absorbing, I could just get lost in it for hours. I think maybe it’s the quest structure. In Oblivion, it was pretty simple – join a guild (for me it was the fighters guild and then the mages guild), and they give you something to go do (kill some shit), you do it, go back, get money. The money was good too and got better as you went up in ranks at the guild. And some of the guild quests were rather interesting, especially the assasin’s guild (and the rewards there were cool as well). Granted, I’m not very far in fallout, about up to Rivet City now, but I haven’t encountered anything like this. The chick doing the survival guide book has shitty rewards. I don’t know. I may have to get back to it at some point. 

So where does this leave game of the year? Last year it was easy. BIOSHOCK. This year, not so much. Some kind of toss up between Dead Space and Left 4 Dead. Gears 2 was wholly dissapointing. And I still have to give Far Cry 2 a fair shake, and I’m waiting till the holidays to play Prince of Persia, so who knows. 

And now I’m going to spoil the end of the movie Trans-Siberia, which probably nobody has seen or cares to see so whatever. First – it was good. I enjoyed it a lot actually, and I’d recommend it if you’re interested in a slow character study flick. So anyway. End the film, bad guys and good guys are on a train. Bad guy raises his gun, levels the sights on Woody Harrelson and his wife (played by the blonde chick from The Two Towers who wanted to get nasty with Aragorn). The look in their eyes shows defeat. They’re about to get shot in the face. And then the camera cuts to outside the train, and we see it’s on a collision course with another train. So of course, before the bad guy pulls the trigger, the trains collide, then bad guy #1 has a change of heart and kills bad guy #2, and woody and blondie are rescued. This is where the movie loses it for me. Not that it was the worst offense of Deus Ex Machina I’ve ever seen (far from it), but it feels completely unnecessary. Why do writers have to write the script into a corner like that? You might as well just have spider man swing in the side of the train with a pizza in his hand, web up the crazy russian bad guys and be on his way. 

Oh well. For now I’m out.

I think, to be honest with myself, I stuck with it for so long because I had learned it well enough that I could, fairly quickly, prototype up an idea and see if it lasts beyond the infatuation phase. I think, my brain works with ideas in the same way it works with dating – there’s this totally unreasonable lust phase that lasts precicely as long as it takes to get to really know the person, and then things fall quickly back down to earth. Same thing with game ideas, so many times there’s that best-idea-ever feeling, and then you start to work on it a bit, and it’s like, yeah, okay, this is about as fun as getting elbowed square in the chinese lunch-special. So prototypes are useful. But beyond that, there’s some serious drawbacks to both XNA and TorqueX. 

For XNA, there’s the distribution problem. If I want my buddy to check out my game, he needs to install the following: .NET framework 2.0, XNA Framework 2.0, DirectX 9 runtime, AND, if my game uses networking, he needs Game Studio 2.0 which includes visual c# 2005 AND the sp1 for that. And then it may still not work. What a fucking nightmare. 

Another problem with XNA is that you’re completely tied to Microsoft’s platform. Initially I didn’t care about this at all and I was excited to make games that could potentially run on the xbox. But as time went on, and as Microsoft changed their pay structure for live arcade, it became less and less apealing.  

As for Torque, I really liked it for a long time and ended up buying the source code, and learning a lot that way. But down the road, the lack of support, the fuck-awful search function on their forum, and the prevalent odd shit that would happen (weird graphical glitches, slowdown, etc), combined with some seriously poor documentation in their code – turned me off of it. 

And now I’ve never been happier. I’m using this, and it’s completely open source, with a huge community. Who’da thunk? 

Also, I’ve dived into 3D for the first, right into the deep end and its buck nasty down here. But I’m liking it. Been working with irrlicht and in bits and pieces, and I’ve got a terrain up and running. Woo hoo, super exciting stuff I know. 

The gravity concept – quick refresh. A puzzle platform game where the “gimmick” is that the player can control the direction of the gravity in order to reach otherwise unreachable places and successfully navigate the environment. How does this work? 

- player can choose among 4 gravity directions (up/down/left/right) at any given time while standing on a platform. this is important because you get into trouble letting the player change gravity mid-air. to see the effects of this, see the indie game featuring this concept the name of which i forget, was some student project at IGF last year and (side note – seeing this idea implemented at the IGF is what originally turned me off from doing the gravity idea – some kind of ‘its been done before mentality working there’) won something, but, really, the mechanic was not fleshed out very well and is not really what i want. there, it was more of rotating the world around the player. gravity on the other hand will affect physics objects and deepen the puzzles

- the environment in question is a room and the goal is to reach the exit. simple as that.

- like the movie Cube, the world consists of various interconnected “rooms”, basically the game world is itself a larger puzzle that the player solves by navigating it and making his way out. possible idea – if the rooms exist on a grid, and can be moved or rotated in their existence in the larger world, then the “larger puzzle” of the world can consist of re-arranging the rooms to find the correct configuration to reach the exit and freedom. more on this another time. 

- the environment will consist of platforming basics such as spiked platforms, moving platforms, and movable physics objects. in addition, the gravity mechanic can be toyed with by having “special” platforms. one example of this is a platform with 1 side glowing pink, indicating that the player can fall through this side of the platform but stand on the others. this can be used to come up with some fairly deep puzzles. more brainstorming will be needed to come up with other special rules for the player to learn and apply. 

- special items. in Cube, the characters use a boot to test for booby traps by tossing it into a room before enterting to see what would happen. if the boot gets incinerated/chopped up/etc, they know that room = death. similarly the player can find special items that can be used to help solve puzzles or rule out rooms altogether. this idea isnt very well formulated right now. will have to think more about this. 

- the visual representation of a gravity change, the player simply starts to fall in the given direction. the camera does NOT rotate automatically, and this is on purpose to prevent dissorienting the player and to reinforce that gravity is changing, rather than the world rotating. the player can choose to tap a button and orient the camera right-side up if they choose. controls will remain relative to the camera, such that, regardless of gravity, left=left, etc. 

- combat. there’s a lot of room to play here. can go simple or deep really, not sure which fits best. 

- story. will get to this next time. too much for the moment. 

And thats it for right now. I’ll keep thinking about it.

for right now the plan is: player controls 3 units, choosing from among various “classes” before the match starts to select his group of 3. switches between them during the gameplay. I need to code up some rudimentary classes and start toying with it.

crunch starts tomorrow. hopefully i’ll make it back here before the apocalypse.

And, why not, throw in some 1upyours formatting style. What have I been playing? Tonight, company of heroes. It’s pretty plain that  I don’t know what the FUCK I’m doing. Infantry doctrine is good, tanks are good, howitzers(sp?) are good, thats about all I know. What a fucking alien genre. I’m trying to learn. Too, coming from consoles, I mean, I think it’s a decent excuse. I think.