This past month I’ve been working on a variety of things. I recently rediscovered some version control history visualization software called Gource, which I decided to try out on a couple of my source code repositories.

This resulted in a couple of fairly short, but interesting (for me) videos that really brought back memories of how those projects progressed.

This first one is a visualization of the history of Be Tiny, World! which was originally a Ludum Dare entry. I continued to work on it after the game jam ended, adding more features, porting it to Android, and doing a major graphical overhaul.

The next one is a visualization of Blasteroids, which was originally made for NANY 2014. I’ve continued to work on it over the years, and I still have more plans for it. So I guess while this visualization is current as of right now, it will be out of date once I add even more features and upgrades to the game.

Another thing I’ve been working on this month is related to GitBook. GitBook is a service I first heard about a couple of years ago but I didn’t think I had a use for it so I promptly forgot about it. But I recently rediscovered it and decided to see how it compared to Google Docs. I’ve been working on converting the cInput Reference Manual to GitBook format. As of this moment, I’ve already converted all the easy parts, and I only have the more complex scripting reference section left to do. But once it’s finished I intend to download the ePub and PDF formats of both the Google Docs and GitBook versions and see how they compare to each other and to how I think they should look/work. As of this moment I’m thinking I’ll probably stick with the GitBook version once it’s done.

And of course I’ve also been working on Cubiq this month. Last time I said I hoped to have a website or trailer video to share with you. I’ve done some work on the website, but it’s still very much an incomplete work in progress, with no real details about Cubiq as of yet. But if you’ve been paying attention, you may have seen a preview of it already.

I think we’re almost ready to start the limited beta testing, but we’re still trying to work out the details of the best way to go about doing that. Itch.io recently announced their take on Early Access, called Refinery. Which seems well timed, but I’m still confused on some of the details, and the more I look into it, the less it seems useful for our needs. It looks really great for PC early access, but not-so-useful for mobile (Android) beta testing.

On the other hand, Google Play already has options for beta testing, but didn’t provide any kind of community aspect of it, such as discussion boards for feedback. So I’m still looking into the possibility of combing Itch’s features with Google Play’s. Or we may just post around (not in a spammy way, of course) on various game related forums with links to the beta and see what sticks.

As a reward for sticking with this blog post for so long, I’ll finally get to something specific about the progress on Cubiq this month:

While testing the game, I noticed that a particular stage ran significantly slower than the other stages, and felt sluggish and unresponsive. After looking further into it we found that that stage in particular had about 4 times as many draw calls as other stages, and also about 4 times as many triangles. This and other issues really slowed the stage down on mobile devices. After some smoke and mirror trickery, we got performance on that stage to be comparable to the other stages.

Cubiq Forest - Forest Theme Compared
Performance statistics comparison.

How? Instead of having a full 3D scene with all the bells and whistles, we took a screenshot of the 3D scene, then deleted most of the 3D objects out of the scene and now display only the 2D image with a few 3D objects in front of it to give the feeling of depth with a lot less for the mobile CPU/GPU to process.

The 3D scene before the change:

The simplified scene after the change:

Admittedly it’s a bit lower quality, but most of the time that view will be mostly obscured by other gameplay and UI elements. All in all, I think the loss in visual quality is worth the performance increase.