New Banner!

Hey! Look at my fancy new banner! Go ahead, click it!

As you can see, I've been working on the look of the blog. I'm not sure how my "aesthetic" sensibilities go over with the general populace, so if the color scheme assaults your eyes, just tell me and I'll try to accrue some taste.

As far as projects go, I'll soon be working with Kevin Chen of geakStudios on expanding "Stealth Robot Programmer." Namely, we want to make a level editor for it. So, that should be cool.

For a while, I was really eager to keep blasting through projects really quickly, but in the grand scheme of things I haven't been working on Ahab's Revenge for very long, and it wouldn't hurt to flesh it out a bit. What I really want to do is demonstrate level variety -- several people have pointed out that the constant upward climb can get a little monotonous. And with full 360-degree firing range and the mobility of a side-scrolling platformer, why not take advantage of the flexibility?

At the moment, there are two primary level concepts that I want to try out. The first, and most obvious one, is to take this in the direction of a platformer. Ahab jumps around dodging and shooting enemies  while moving horizontally toward the end goal, with the destructible blocks thrown in as obstacles along the way. It's pretty simple, but with enough interesting elements added in, it could be fun.

The second concept would lead Ahab's Revenge in a more purist puzzle-game direction. In the "painter" levels, Ahab is in a single room with a solid block of corals in the center, and a fixed number of whales strewn about the stage. The corals are not moving, and the aim of the level in this case is not to destroy all the corals, but to destroy the right corals to make them form a specific pattern. As if you're chiseling out a sculpture. The tricky part is that the colors of the coral are not initially set so that it is possible to achieve your goal; instead you must use a fixed arsenal of "paintbrush harpoons." These are fired from your back like regular harpoons, but are giant paintbrushes that will change one or more of the corals to whatever color it is equipped with. Once you have the colors set up, right, you can switch back to normal harpoons and start clearing the correct corals by hitting them with whales.

If developing Ahab's Revenge has taught me one thing, it's that nobody has any idea what I'm talking about. So here's an example of what such a level might look like:

And you would be given an objective at the beginning of the level to use the right combination of paintbrushes and whales to carve the chunk of corals into, say, an "H" shape.

Anyway, give feedback! Tell me if this is an interesting path to follow, or offer your own ideas.

nathan

Taking Care of Business

NOTICE: All of my games have officially been moved! They are now neatly categorized in the "GAMES" tab that you see above.

A new game level that I made is up on geakStudios! It's called "Stealth Robot Programmer." The level itself sucks and is boring, but try to concentrate on the core mechanic: you get to sneak up on robots and reprogram them to do whatever you want (within reason)! The game itself is not my idea, but one formed before I even started collaborating with geakStudios. However, I sincerely think that it's one of the coolest ideas for a puzzle game imaginable.

I know I've posted this link about five times, but you can play the demo here at the geakStudios site. Make sure to read the instructions!

nathan

This will be the next "Robot Unicorn Attack," I guarantee it.

[SORRY! The game has been moved here: Ahab's Revenge]

In case the start-screen instructions aren't explicit enough, A is left, D is right, W is jump (hold it longer to jump higher), and S makes you drop a level if possible. Oh, also ENTER pauses the game. I accidentally left that out.

It's finally here! My first game coded from scratch, Ahab's Revenge. There are definitely some things I can change up in the future -- I want to improve the scoring system, add some cool scenery underneath the coral, and make the boss a bit more interesting -- but it's finally reached a point where I feel like I can call it a "complete" game. Of course, feel free to leave me feedback. Particularly if it has to do with the speeds of all the game elements or their hit-boxes, since those are things I'm still experimenting with in order to tweak the difficulty.

In case anybody is curious, the background music is "Heiße Lippen" by Cluster. The rest of the audio I made with a combination of my mouth and Audacity.

You may have noticed that the name of the blog changed once again. Aside from choosing to make the title more gaming-related, I also had to decide whether to make it a Bowie or an ABBA reference. Yes, it was stressful. Next order of business: make the blog not look awful.

EDIT: Do you like game prototypes, but hate my taste in music? Then go play Ahab's Revenge at the geakStudios site, where they've been kind enough to publish the game minus the copyright infringement-tastic soundtrack.

nathan

New Projects

Well, it looks like after a week or so of pretty intensive work on the whale-shooter, I'm going to put that on hold for a bit. Don't worry! It will be finished! It's very much my baby, and I will not toss it aside that easily. But, I feel as though I've reached a point where I can take a break for a while and still come back with a good idea of where I left off, and I've got other projects on my plate right now. Some of these projects are for my friend's fledgling game company, and they are not particularly fond of babies:

geakStudios

geakStudios is all about "fail-fast" game design, that is, they release a lot of simple, under-produced core gaming prototypes to the community, and continue development on what sticks. So with any luck, I'll soon be producing bucket-loads of total crap for you to enjoy! Go look at the website and enjoy the alpha releases of "Quicktime Coma" and "RTS Tuttle." I didn't have a hand in either, but when I start developing stuff, maybe I'll give some behind-the-scenes looks at the development process as I have for the whale-shooter, depending on how much geakStudios insider information I'm allowed to reveal.

As far as other personal projects go, my mom has asked me to make a game for her yoga studio's website, so my first idea was "Yoga Yoga Revolution," in which a raga plays in the background and you have to press certain button combinations to perform the correct pose. Also, I've been hoping to collaborate with a couple of my more artistically-inclined Vassar friends, including the curator of this site: Patootchy and Mallow & Koff. She's a computer science major, too! Yay!

That's where we are now. And I'm moving up to New York in about a month! So exciting. Later, folks.

nathan

In Dark Trees

Okay, Flash got me this time. I didn't get much programming done today because I'm still trying to wrap my head around the coral problem. I've got some sort of algorithm for collisions involving the giant bitmap, but I'm not sure how well it's actually going to perform in practice, just in terms of accuracy. Also, a new layer of confusion has just been added. I tried storing each individual coral piece as its own bitmap and updating the position of each bitmap every frame, just out of curiosity.

And... it turns out that I really underestimated how fast bitmapping is compared to spriting. There is NO noticeable slowdown in the game-play, and I made about 45 rows of them! Oddly, though there's no lagging in player controls or whale movement, the bitmaps get spaced incorrectly as you get higher up, which may indicate that Flash is having more trouble iterating through the giant array storing all the bitmaps than copying the pixels. Of course, this is just me guessing, and I could probably get around the problem by dynamically slicing off pieces of the coral array and sticking them back on when they're needed again. since the position of the coral relative to the rest of the coral is always the same, there's no point in updating the position of coral that's off-screen every frame when I can just stick it back on again if it's needed.

So, I'm torn. The large bitmap approach seems undeniably more efficient, but I only have moderate faith in my algorithm. The small bitmap approach makes collisions a lot simpler for me, but possibly a lot more tedious for Flash, and I'm wondering if there's a way to do this that wouldn't result in too much overhead code. I could really use a personal Flash guru right now. If anybody has any suggestions, they are totally welcome. For now, I think I'm gonna take a break for tonight and get the rest I am sorely lacking.

Here's a rushed implementation of the small bitmap method. There's some unfinished code, so it might glitch out a little if you shoot the whales!

Log Entry #1, 8/5/2010: The coral is sort of working! Not very well, but sometimes when you shoot whales at coral of the same color, it disappears! The game's not giving me any errors. Basically I just have to refine the algorithm, which will be a lot of work, but it's still reassuring.

nathan

[EDIT:Moved it again.]

Moment of Truth

The game is reaching a crucial point in its development, and I thought this was worthy of its own post. Basically, everything's running pretty smoothly; there's some mostly trivial collision stuff that I want to add at the end, but the last main obstacle is the barrage of "bubbles" that are supposed to descend from the heavens. As I mentioned two posts ago, Flash doesn't seem to be very good at rendering large numbers of moving sprites. So, I finally decided that it was time to stop re-inventing the wheel and go do some online research. What I came across was this:

Arrow Test

To quote that guy from The Wire, sheeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeet. That's a lot of sprites.

Well, actually it's not. The test came from one Mike Grundvig's blog, and it seems to be a clever use of the BitmapData Object type's copyPixel function. A fast way to move non-animated sprites across the screen? Could it be my salvation? Could it be?! Well, I'm certainly going to try it. In the mean time, here's the game as it stands. Steer with the WASD keys, point and click to shoot. you can drop down levels by pressing down (S). Bon appetit.

Log Update #1, 8/3/2010: Today is performance day! Performance, performance, performance! The results aren't immediately noticeable, but I've made the game better at cleaning up stuff that's off-screen, so now it's never updating more than 8 whale objects at a time. I've also been goofing around with bitmaps, so hopefully my bitmap-bubble plan will pan out by the end of the day.

Log Update #2, 8/3/2010: Unmitigated success! Well, so far. I initialized my block of bubbles (which are more akin to cubes of coral now), made it creep down the stage, and there was no noticeable slowing. I didn't explain it so well earlier in this post because I hadn't tried it and I didn't know exactly what I was talking about, so here's a better explanation for anybody interested:

When you make game sprites from Flash's built in drawing engine, to move these sprites in the game it literally has to redraw them in a different position every frame. In an earlier post, I mentioned getting around this by making it into one giant sprite, but this apparently doesn't help because Flash still has to redraw the damn thing every time. I opted out of Flash's vector drawing entirely by just copying everything to a giant bitmap image in memory, and for reasons that are not entirely clear to me, moving this giant bitmap is much easier for Flash than moving a ton of sprite information. Now when the screen scrolls up, I can make a bigger bitmap, copy all the pixels from the old bitmap to its bottom half, and fill out the rest.So, let's get going with that.

Log Update #3, 8/3/2010: Wheeee! Gigantic bitmap resizing is working surprisingly well. Of course, now I have to program collisions with it, which may prove difficult. In other news, look how trippy the game is right now! Sadly, it can't stay that way. :^(

Log Update #4, 8/4/2010: Here's what's on the agenda for today, just so I can stay focused: first I'm going to fix the level-scrolling glitch that I discovered last night, which should be easy, and then I'm going to try one method of collision detection between speared whales and coral cubes. Basically, I use ActionScript's built in hit-test function to see if the whale is intersecting with any part of the bitmap. Then I use another built in function to convert the stage coordinates of the Whale's center to its coordinates relative to the bitmap. At any given time one whale should not be intersecting with more than four sectors of the bitmap, so I'll cycle through them in order of closest to the Whale's center to the farthest, and if any of them are non-empty, the pieces of the collision will react accordingly. The awesome thing about this is that it's insanely efficient; instead of iterating through a giant list of corals and testing collisions on each, I just grab one by its grid coordinates. The problem is that I have to do math. Sigh...

nathan

[EDIT: Game has been moved to next post.]