|
Is Random RandomExperiments in game design |
|
Work on this has been paused kind of indefinitely - my home programming time has been taken up with a paid project. Tis a pity, but I am pleased with what I’ve learned. Whether I pick this project up or not (I have other ideas perculating!), I will be picking up another game project once this paid stuff is done.
Been a week or so since I got anything done, but I got in about 1 hr, 50 minutes, duly entered into my time tracking thing (which is still way overkill)
Today was all about events. Events, events, events. Trying to come up with a good event model - and I’ve got a pretty good handle on it. Going to make some notes here, as I realized some things just as I closed my window (and because it’ll be awhile before I get back to this project, as I have a paying project to slam through in the next few weeks) Full Story »
I found an easy to install project tool, complete overkill, but that’s in and I’m tracking those hours now. Can divide up by projects, etc. Of course, it has lots of collaboration stuff built in that I have no need of, but.. enh. It works.
First time trying to work with a blog tool in the other window. Let’s note - starting again at 9:45, though some of that time is going to get sucked up in the blogging. I haven’t the faintest if anyone is interested in this process - maybe we’ll see.
First up, a note on the map architecture - essentially, along with obvious things like the dimensions, my maps have a concept of a default tile, the tile that is there if nothing else is. Everything else is noted on an exception basis. Currently, the map files are .csv files, since those were stupidly easy. I’m mixed on changing that - yeah, if I leave it like that, the players can change the map. On the other hand, if I leave it like that, the players can CHANGE THE MAP. Cool. I think I’ll go with the idea that I’m not competing against the players - if they want to ‘cheat’ and that makes it funner, they can go for it, with my blessing. So, csv it is, barring technical snag.
And made an update to the scrolling code. It now does nothing until you get close to the edge and then centers. A better way might be to smooth scroll to center, but we’ll see… in any event, this method is FAR better then what I had going.
+20 minutes to the work. Might get to a map editor tonight
So, uh, if I’m not just writing a tech demo (for unimpressive tech, at that), I need some game mechanics behind it. To some extent, one is going to influence the other.. but not completely.  I had been thinking for a long time that I wanted to do a city simulator, take the AI in Oblivion and kick it up a notch (since, hey, I can’t really beat the graphics). I don’t think I’ll go quite there yet, I’ll do a dungeon crawl, sorta like the Rampant Coyote’s Game-In-A-Week Hackenslash. I’m not doing this in a week, but I want something solid I can play around with in, say, two weeks. I’m defining ‘week’ the same way he did - 40 hours of work. I’m at… 20 hours thus far (ok, from now on, I note hours in these blogs), I think? Slower then his pace, but, hey, I’m learning a new language here and I’ve never done a lot of this graphics/game specific stuff, either. I’m not committing to dropping the game at 80 hours, I’m just committing to having something that people can play and look at by then.
Ah! All caught up, I’m now blogging my last development evening, a few nights back.
The next step was a big one - supporting a map bigger then fit on the screen. That means a minimap. That means sprites appearing and disappearing. That means scrolling the display window. Oh, my. TONS of off-by-one errors, what with the whole keeping track of the dimensions of map and the window into that and translating to and from relative positions to absolute ones and then relative positions to pixel positions and tracking when people left the screen or returned to it… eeesh
It also meant actually defining what the screen was going to look like and where things would be. So I decided to surround the main map window with a 20 px wide border, so I could put manually scrolling arrows on it (which still aren’t there). I tucked the mini-map in the top right. I left lots of space for status information for.. something. More musing on game design in a later post, but I have some ideas. And it meant making all that work.
Ok. I am NOT an artist. At ALL. But today was programmer art. Ick. Ugly, ugly. A top down sprite view of a character. The ugly green colour is transparent, at least. Even worse, there’s two more of those - one of them pink and one green. As part of that, the tile size went to 30×30, since 20×20 was just useless. And that brings me to my next topic - I subclassed BaseMover in this one. I started with the UserCharacter subclass - it is a basemover that just gets told what to do by mouse clicks. Not much change - it doesn’t even override Update(), it just gets ordered around by the event loop, which just sets its destination.
Pathfinding problems. Screwing around, I ran into serious pathfinding performance problems. Basically, I took away the graphics for awhile and just did pathfinding tests. And my current algorithm worked great for 10×10 or 20×20 mapboard.   But I tried for 100×100 and it coughed up blood. Hum.
More reading, more googling. And came across http://www.policyalmanac.org/games/aStarTutorial.htm
And it says that the algorithm I used last time WASN’T A*. Huh. So I implemented this one and it does work better. We’ll have to see if it works well enough when I actually get going with a real map.
Hrm. So what was next. The problem with writing this blog a week after the sessions - hopefully that won’t happen from now on. The next day was starting to bring in something other then just tiles and splitting things out in more detail.
At this point, I’m thinking of a game that’s a city simulation. Don’t really know where the GAME is in there, running through a number of ideas, including being a city watchman, being a city dictator, a conquerer… a thief. Lots of options. At this point, the only thing I’m fairly sure of is that I want the AI to have some depth in it.  I’ve done a whole bunch of reading about AI over the years and I think I’m going with a goal oriented system and one in which the AI characters interact with each other and the player will be able to see that. How? Don’t know yet. I have some ideas about need queues, goal queues, etc, but I’ll have to work them out over time.