Category: Projects

As stated before, I was aiming to have BattleGrid playable on my iPad by PAX so I could get feedback from strangers. While the game is certainly playable on my iPad, it still suffers from framerate problems when things get hectic, so I’m probably not going to be showing it to strangers – it’s simply not ready for that.

I think I’ve figured out how to fix the framerate problem, I just don’t know exactly how to do it, and the only tools I’ve found to help either don’t work well or are a bit complex.

The problem is that iOS simply can’t handle drawing a lot at once. Unity uses “dynamic batching”, which means that each time it needs to redraw, it batches everything together that’s using the same material (textures + shader) and sends it as a single draw call. My game uses about 100 – 140 draw calls in busy scenes. Everything I’ve read says I should keep it below 30 draw calls for good performance. So I’m using over 3 times as many draw calls as I should.

I think I know what needs to be done to fix it – I need to use a texture atlas. This means taking all the individual textures and cramming them into a single big texture. This is the easy part.

The hard part is that the UV mapping on all the models also needs to change. The UV mapping defines what part of the texture is displayed on each face of the model. When all the textures are combined, I need to scale and offset the UV coordinates for each model, and I can’t find a tool to make this easy. In the end, I’ll likely just write a tool to do it myself, since working with meshes in code isn’t too difficult. Writing a tool to do the work for me will take some time, though, so there’s no time to do that before PAX.

I’ll install BattleGrid on my iPad 3 and see if it runs better, but I’m not getting my hopes up.

Wow, two months since the last update. Well, the last update I actually published to the blog – I’ve had about a half dozen posts I’ve started but never finished because the things I was writing about changed before I was finished. And because I wanted to add pictures I never got around to grabbing. So this post is going to sum up everything from those half-finished posts.

I spent a lot of time last month tweaking things, figuring stuff out, and updating graphics. It was a good month of work, and resulted in most of the frustrating bugs being fixed.

Toward the beginning of this month, I sent links to a few friends to start getting some feedback. As expected, the feedback has been extremely valuable. The in-game UI has changed to something that works much better – something between the new and old UIs. I created a series of help screens based on feedback about what confuses people. I also got the idea to create a map for the sole purpose of testing the frame rate when things are hectic, and that one map is the map I play all the time now as I improve things.

A week or two ago I installed BattleGrid to my iPad. It runs, but as of the last version I installed, the controls weren’t working, and the game ran terribly. On my test map, the FPS would drop to 3. Yikes. It’s good that it’s running, though.

Unity also decided to give away their “Mobile Basic” kit, which includes iPhone and Android publishing, so BattleGrid will probably end up on Android tablets at some point. I’d need an Android tablet to test with, so the iOS version will still come first and long before the Android version.

Some of the developers in the Unity Asset Store decided to put their stuff on sale with Unity giving away the mobile kit, so I’ve spent about $200 on development tools to make my life easier. I’ve been pretty pleased with most of what I’ve bought, but there’s a downside… I bought a new UI framework, and I’ve been moving my entire UI over to use the new framework. The framework is fantastic – it’s much more efficient than what I had and much more flexible, but it has taken time to figure out its quirks and how to use it properly. I also got a tool that handles gestures for me, so I’m replacing a lot of my control code with updated code that uses the gesture framework.

My productivity has also suffered because Mass Effect 3 was released and I’ve spent a ton of time playing it.

So, I might not hit my PAX deadline for having something playable… I’ll see what I can get done this weekend and next and cross my fingers.

I’ll also try to post some more updates soon about the changes I’ve made.

It’s been a while since I’ve posted any updates about my game, and that’s mainly because I’ve been spending a lot of time working on it and things have been changing pretty quickly. There’s about a half-dozen posts I’ve started in the past month about things that have already changed.

Anyway, toward the beginning of this month, I sent links to the game to a few people to start getting feedback. As expected, the feedback has been extremely valuable. The in-game UI has changed to something that works much better, and I’ve started work on help screens based on feedback about what confuses people. I also got the idea to create a map for the sole purpose of testing the frame rate when things are hectic, and this one map has been very useful as I figure things out.

What people don’t know yet (including my few testers) is that the game is running on my iPad now. Granted, it runs terribly, but it runs. I’ve only got about three weeks before PAX, so having it running on the iPad is a huge step toward my goal.

The other big news is that Unity made their Android basic kit free earlier this month, so there’s a good chance BattleGrid ends up on Android tablets at some point. Of course, I need an Android tablet to test with, so the iOS version will still come first, and long before the Android version, since I need to see the iOS version make some money first…

Oh god, I hope it makes some money. Enough to pay off the development tools and licenses, at least.

I’ve been searching for a clever solution to an issue I’ve had with factories since I introduced the “5 active units per factory” limit: how do I show the player how many units are active?

The straightforward solution is to display some sort of bar above the factory, like a 5-segment progress bar that fills up depending on how many units are active. I didn’t really want to do this, though. Mostly because my experiments in getting UI elements to line up with 3D objects have been unsuccessful.

One of my other ideas was some sort of antenna on the factory that sent out little pulses to the units periodically. I liked the idea, but it doesn’t solve the problem – there’s still no indication of how many units are active unless a player waits for the next pulse and counts the pulses. I liked the idea, though, so it’s been sitting in the back of my head a while.

Then today, a neat idea struck me that I thought I’d try out. Enter the new factory:

The new factory is a bit more compact than the old, but for good reason. It needs some room for the new drones that hover around it:

A new drone is spawned for each unit the factory builds, so you can easily see how many units are active by counting the number of drones. And every few seconds, the drones send out one of those “pulses” I mentioned:

The pulses seek out the unit the drone is attached to. These pulses look fantastic (a screenshot doesn’t do them justice) and add a little new animation to the game.

The factories are likely to get another change later since I’m not totally satisfied with the new look, and the pulses, while looking pretty cool in-game, are still a little drab, so I’ll have to dress them up a bit.

The switch to the “select a cell, then a structure” UI when building structures looks and works great. There’s only one problem: not all my structures fit in a 1×1 cell. Some structures are 2×2, some are 3×3. The new UI still worked for these sizes of structures, but didn’t feel right. For 3×3 structures, the tile you picked is in the center of the structure, which is OK. For 2×2 structures, though, the tile you picked is the upper-right of the 4 tiles. I lack a screenshot, so this will have to do:

I liked the new UI too much to switch back, though. It felt right, even if it wasn’t working great yet. So I did some more thinking, added a bunch of code, and a few iterations later I have things working pretty well. The only downside is that building requires one extra tap/click. However, requiring that extra tap works out pretty well.

When you tap a cell, the build menu appears. When you select a structure, the game finds the appropriate position closest to the tile you selected and places the structure there. Then the camera re-centers on the structure. For 1×1 (or 3×3) structures, there’s no change (though I might add a slight zoom-out if you’re too close for 3×3 structures), but for 2×2 structures, the camera will line up on the intersection of the 4 cells:

If the location is valid, the cells with light up green. If not, the invalid cells will light up red:

If all the cells are green, you can tap the structure, and it will start building. The part that works out well with the switch is that you can select a structure, and if the location is invalid, you can use the standard camera control gestures (swiping and pinching) to pan and zoom the camera. The structure you’re working on will stay centered the whole time, and when you’ve moved to a good position, you can tap it to build it. This is an improvement over the old UI, where you’d tap a structure and you could drag it around, but you lost control of the camera, so the structure had to go somewhere within the camera’s view when you selected the structure.