Vapourware Codexian Game Development Thread

[chewie]

Something good for programmers to read about FP:

http://randomascii.wordpress.com/2012/02/25/comparing-floating-point-numbers-2012-edition/

Indeed - important read & issue.

 

[SCO]

Edit:
If Mikko's latest line of research pans out, and I understand it correctly, you could make turn based games in dynamic worlds.

How do you mean that? Use small words, you're writing to a functional math retard here.
It seemed to me that the distinction between waypoint and grid is the amount of runtime calculations you are prepared to pay for for the flexibility (no overlapping waypoint grids for different sizes of agent) and realistic movement of a grid (all moveable area passable, dynamic recalculation), vs waypoints.

Whenever i think of the "overlapping grids" that need to be abstracted to 2d from 2 floored structures i get a headache.

But this doesn't seem to have anything to do with turnbased movement? The path replaning is something that should be connected to a more specific "game AI". Like the pathfinder goes "Uh bro, the path we were using is not optimal anymore, your lieutenant shot those rocks there." and the captain goes all "charge". But i don't see what that has to do with waypoints or nav-meshes - especially in a turnbased game, where the agents aren't moving at the same time, so the performance hit of those calculations is minimal - though i guess it could be useful in a ToEE like "enemy turn" where they move at the same time.

Things i'd like to see is a real library of these "replanning" phases. If you consider it, this stuff is a pre-requisite to to larger concepts like flocking squad movement and flanking.
Also, a abstraction of the "create a path" and "obstruct a path" action would be crucial. Implementations for the library user : opening a door, shooting a rock, blowing a wall, blocking stuff with a flame spell. The pathfinder would go "Bro you want to create a new path to this fool? These are the zones i think you should check out", the agent goes "WHAT IS BEST IN LIFE" and destroys that door. Or the opposite situation: "Bro, this steroid-abuser is coming to rape you, and there's a open path", "*casts wall of fire*, *casts buffs*"

Cooperative planning would mean that the agents would need to all be notified and then vote for themselves how to deal with the situation, so a "too simple" model wouldn't be very useful. Maybe agents would give out "blocking areas" and "liberating areas" they could use, and the pathfinder tries to combine them in the next X turns; for X < extrapolated position of the guy we want to avoid/chase at T+X; to create a path?
Something more than simply a event bus anyway.

And it would be nice to see low intelligence critter team putting a wall of ice and a wall of fire next to eachother for example. The ice melting in the next turn would be a nice lulz moment.

If this stuff could be abstracted without going into retardedly specific game code, i think it would be very nice.

 

[SCO]

Ideas are cheap, work is hard.

Story of my life.

 

[Davaris]

Mikko's big idea if I understand his article correctly, is compression of data, so it is realistic to use detailed grids.

...

Too much data, information overload... You are spilling outside the realm of path following into behavioral aspects of AI and sorry I'm no expert on this stuff.

Ideas are cheap, work is hard.

Story of my life.

If you have a ton of money ($10,000 and up), you don't have to do much work at all. Havock can probably do what you have described:

http://gamedev.stackexchange.com/questions/2140/available-ai-middleware

Or try a good AI book. One I can recommend for amateur level people like myself is, Programming Game AI by Example. It starts out with state machines, moves up to scripts, has working source code of a soccer game, working code for an FPS style bot, covers goal driven behaviors and fuzzy logic. Pretty much everything you need to make an enjoyable game.

http://www.ai-junkie.com/books/toc_pgaibe.html

 

[20 Eyes]

I've been too busy with classes to get much done, but I've been going through Nick Gravelyn's tile engine tutorial (http://www.youtube.com/user/nickgravelyn, for anyone interested). Very informative stuff and he explains everything pretty thoroughly and somehow keeps things fairly quick and concise.

 

[tiagocc0]

Or try a good AI book. One I can recommend for amateur level people like myself is, Programming Game AI by Example. It starts out with state machines, moves up to scripts, has working source code of a soccer game, working code for an FPS style bot, covers goal driven behaviors and fuzzy logic. Pretty much everything you need to make an enjoyable game.

http://www.ai-junkie.com/books/toc_pgaibe.html

I've just finished this book, very good indeed!
It's very well written and has lots of examples.

 

[GordonHalfman]

I haven't found all that many useful books about AI. This was very good though:

http://www.amazon.co.uk/Behavioral-...r_1_11?s=books&ie=UTF8&qid=1334310261&sr=1-11

Doesn't waste time on the implementation details, source code etc, and actually gets into the nitty gritty of good practical decision making tools.

 

[Zed]

Okay so I'm gonna attempt to move away from tile-based movement and make free movement. Tile-based movement doesn't make a lot of sense as it is. World will still be 16x16 tile based.. might up it to 32x32 but it will remain tiled nonetheless. Hum hum.

 

[20 Eyes]

Okay so I'm gonna attempt to move away from tile-based movement and make free movement. Tile-based movement doesn't make a lot of sense as it is. World will still be 16x16 tile based.. might up it to 32x32 but it will remain tiled nonetheless. Hum hum.

I've been weighing the pros and cons of both systems. It seems like it'd be easier to make AI for tile-based combat but free movement removes the restrictions (obviously). Free movement would also be better at having subtle range differences actually matter, like in ToEE.

 

[Zed]

Yeah but the thing is I'm gonna use a separate screen for combat.
But I just remembered that I had some puzzle plans for a tile-based system.
Fuck it I'll just keep it like it is and start on combat.

 

[chewie]

-> Disco!

Fleshing out lightsource properties of game entities - and to have some fun while doing so, I created some crappy programmer-art-light-animations.

 

[DakaSha]

I actually finished this shit a long time ago but just never though about posting it. May as well so that it wasnt a total waste of time. Not like ill ever use it in a "real" project but it is game related

A*Star in HaXe

First time i ever designed a UI so my bad

 

[20 Eyes]

Haven't posted here in a while, been busy with summer classes + other real life stuff. I've been going through Head First C#, I'm about 2/3rd done with the book and I think I'm going to make a serious attempt at a simple game when I finish. I'm just about done with the second of the three major labs, which happens to be a Windows Form turn-based RPG (features AAA graphics and serious choices-and-consequences such as 'do I attack with mace or sword?'):

I finished most of the tricky parts (tricky for a beginner, at least), it's just a matter of filling in a couple of things. The cool thing is they don't give you the solution code, so if you're lazy like me it's not so easy to cheat yourself and just check the book for the answer. I also got bored at some point this week and started writing a tile engine, this time with as little references as possible. It's funny how much better you understand something when you build it yourself. Right now it just builds a layer of X and Y of the same tile, next up is allowing different tile textures.

I actually finished this shit a long time ago but just never though about posting it. May as well so that it wasnt a total waste of time. Not like ill ever use it in a "real" project but it is game related

A*Star in HaXe

First time i ever designed a UI so my bad

I might seek some guidance if/when it comes time to put A* into my tile engine, but that's a ways away.

 

[DakaSha]

Yeah just let me know

 

[20 Eyes]

Tile engine is coming along, but there is still lots to do. It's starting to resemble a game, which feels pretty good. Dark grey rectangle on the bottom is actually the GUI, though you really can't tell by looking at this point. Next on the agenda:

• Sprite movement
• collision detection
• pathfinding
• rewrite the tile's texture setter so it loads from a single tileset image (right now each texture is a single file)
• Do something with the GUI

 

[shihonage]

Trying to give Monsterland walls a superficial appearance of depth without killing performance.

 

[20 Eyes]

Got collision and pathfinding down. Collision took about 10 minutes, thanks to my super simple tile engine, and pathfinding went in really easily to. If you skim back several pages in this thread, you'll see I had a hell of time getting pathfinding to work so I was quite surprised when I got it going so easily. Of course I had no idea what I was doing back then (I have just the faintest right now). Thankfully, I had a dozen A* tutorials and samples laying around. Mouse values are off in the screenshot because I haven't clamped the mouse to the screen yet. If you can't tell (I know it's pretty basic looking), the grey tiles are walls and the shading is the pathfinder finding the best route between two tiles on either side of the wall.

I guess it's time to put some thought into the specifics of my game. I've got some work to do deciding things like number of units, how much strategy vs how much RPG, how much game I actually want to make, ect. I have a basic idea for the setting and the combat mechanics, but that's about it. And I still need a camera class and to rewrite the tile texture setter to take single-image tilesets.

Trying to give Monsterland walls a superficial appearance of depth without killing performance.

The tiles look nice. It used to be ASCII, right?

 

[shihonage]

Yes, but I'm trying to get away from the ASCII stigma and bring it to the level of VVVVVV visuals-wise.

Task probably achieved.

 

[eric__s]

I've accomplished a lot in the past few weeks. I feel like this is the level of productivity I should try to maintain, if not surpass, for the next project I work on. I've written two pretty good 16-bit FM songs for a platformer that's winding down production. I think I only need to make one or two more songs at most.

Area Song
Boss Song

Here is a map I made for a roguelike I'm working on.

It's kind of weird that a roguelike is using predefined maps instead of randomly generated maps, but that's what the lead designer wanted and I think we're able to get better results out of them, although it's very time consuming. You can't really tell from the picture, but these maps are gigantic and if you put together all the maps I've made, it would be 4 or 5 times bigger than Wizardry 7.

I've also finished some songs for some other games, a lot of design document stuff for an RPG and have been talking to a lawyer and accountant about setting up an actual company. This is really intimidating and a little over my head, but also pretty exciting. All I can think is 'this is real, this is really happening'. Big things are in store!

 

[20 Eyes]

After some bug squashing, pathfinding and movement are both working smoothly. Movement is completely mouse-based. I added a camera that scrolls with the mouse when the mouse gets to the edge of the screen or with the arrow keys, but it has a minor kink or two. I started on the GUI, here is a (very) roughdraft. This screen probably looks eerily similar to an older one I posted months ago, because I'm using the same placeholders. Since I took this, I added an End Turn button to the upper right of the GUI bar. All it does right now is exit the scree, because I haven't started on the turn-based engine.

To Do:
Knock kinks out of camera
More GUI stuff
Start on turn-based engine
Rewrite tile texture loader

 

[asper]

I love those graphics cboyardee, did you do them yourself? Cool music!

 

[OuterSpace]

Yeah cboyardee, that looks really awesome.
Are you using any third party mapping software like Tiled to create the maps?

 

[eric__s]

No, I'm not the graphics guy on that project. I can't do graphics at all, I'm mostly just design, music and sometimes very shitty programming. The graphics were done by a super talented guy named David Nyari. If you like them, you should check out some of the other games he's worked on, 78641, Vidiot Game and and a million more. Also, the mapping system was made by the programmer, Jesse Ceranowicz. I keep telling him it would be a really good idea to include it with the game, but he doesn't want to : (

 

[Destroid]

I made this little prototype last night. It'd a 4D puzzleish game, and I think it's a bit of fun.

https://dl.dropbox.com/s/6i3xbv5j6h1vqjc/4drl.swf

Pretty raw atm, but you get score and you can lose. The objective is to move through the portals and escape the Cosmic Horror that pursues you through space (and time?).

It's also a bit of a headfuck being 4D. What you see is 4 columns, each of which represent a view down the entirety of a single axis. You are able to move back and forth along these axis, using the Q, W, E, R and A, S, D, F to move up and down respectively (space to wait or restart after you die, you won't want to wait though). The purple dots represent your position on that axis, and a position is given at the bottom. The asteroids are impassable terrain.

To get points, you must escape through the portal (a blue triangle), and it's distance (in moves) is given at the bottom to help you find it. On the way you must evade the Cosmic Horror, which will seek you out every level. If it moves onto your square you immediately lose. The Horror moves towards you each time you take a move, but if it is blocked by asteroids it will remain stationary.

This is pretty damn hard at present, the best score I've seen from anyone is 8.

 

[OuterSpace]

Nice Destroid. I'm not exactly sure what's going on when I'm playing but I like the visual style. What did you use to create this and how are you hosting it out of dropbox?