It’s been a busy week at Radiangames. Here’s a quick recap of the latest happenings…
Android 4.4.2 Issues
I’ve been getting an increasing number of reports from players that new devices with Android 4.4.2 won’t run my games. In most cases they crash instantly, though in some cases they run for a couple minutes first. This is an awful problem for a developer, because it’s a bug that involves lots of different people (Unity, Google, whoever wrote the graphics driver). In a nutshell: The OpenGL driver for new Andreno graphics chipsets crashes when you do a memcpy on vertices (the way that Unity 4.X does).
You can’t rely on that to be fixed by drivers, as some people may be on 4.4.2 for a long time. Unity 3.5 doesn’t crash, but I can’t convert all my projects back to that without it being a massive headache. And Unity doesn’t update their builds too often, so if you find a problem like this, you’re stuck for a while. In fact, they just fixed the bug (according to their Issue Tracker), but who knows when they’ll have a download for a new build. It could be a week or 3 months. I tried all sorts of supposed fixes and workarounds, like previous versions of Unity (4.1 and 4.2), shader changes, and using a 32-bit color buffer. Nothing worked.
I found one other person who had worked around the issue after giving up that there was a Unity-based solution. It was the guy who wrote NGUI (the most popular UI plug-in for Unity), Michael Lyashenko. I didn’t entirely understand his solution at first, but he helped me figure out what the fundamental issue was. Basically when you create a custom mesh in Unity, you can only change the size of the triangle indices list, but not the vertices lists (position, color, UVs, etc).
I changed my custom mesh rendering system to fit those requirements and sent a new build of Inferno+ to one extremely helpful fan (Jeffrey Pead), who confirmed that Inferno+ no longer crashed with the fix. Problem solved? Well, solution found at least!
One Fix, 16 Builds
So the fix for my games and Android 4.4.2 has been found, but now I need to update 8 projects with the new rendering system, and make 2 builds for each game (1 each for Google Play and Amazon) and upload all of those. That’s one downside of making so many games. When something goes wrong, it has a huge impact.
Localization Is Hard/Easy
Other stuff keeping me busy: I’ve been doing some more localization work for The Powerpuff Girls. More specifically, we’re adding Turkish, Russian, Arabic, Korean, Chinese, and Japanese to the language options. I’ve been dreading doing this task for a long time, and have put it off as long as I could. It’s not been as bad I thought in many ways, though it’s still something I hope I never do again. I’ll post a quick guide on Gamasutra before I start working on Inferno 2.
Not Bad For a 4-Year Old
JoyJoy turned 4 years old last Thursday. At least the original XBLIG version did. So I didn’t have high expectations for the launch of the mobile version. I hoped it would do better than SideSwype, which started decently strong on iOS, though it died off pretty dramatically since then. Let’s just say JoyJoy surpassed my expectations, as it’s been my best iOS launch so far. Not by a major amount, but better than Inferno+, which I didn’t expect. It’s also done well on Google Play (roughly the same as Inferno+).
Reviews haven’t been particularly amazing or anything (it’s just an arena-based twin-stick shooter after all). It got decent placement by Apple–the best I’ve had for my self-published games, though not as good as Bombcats or Super Crossfire HD–and the really fortunate thing is the timing of the game. There just haven’t been many twin-stick shooters lately. That doesn’t mean I’m going to focus exclusively on twin-stick shooters, but it does make me feel better about Inferno 2 coming in July and a Ballistic spinoff planned for later this year.
Windows Phone 8 Update
So far, WP8 sales have been not-so-great, but it’s still early. They’re pretty similar to Amazon sales the past week (aside from Inferno+ being Amazon’s FAOTD), which is to say the games are not selling a lot. I’m experimenting a bit to try to increase them and just understand the store and its lists better. Bombcats SE and CRUSH are free on WP8 for a short while, I’m using some free AdDuplex impressions to get some extra downloads too, and I’ve removed the trial mode from the couple games that had it. The storefront favors free games heavily already, and free games are mixed with paid ones in one of the major lists (New + Rising). Having a trial really doesn’t help anything because of how neglected it is in the store UI. Of the 4 major stores, Windows Phone’s is only just behind Amazon’s for how terrible it is (though to be fair, I only have an OG Kindle Fire, maybe it’s better on the newer ones).
That said, Inferno+ will be featured on WP8 in the US tomorrow, so I’m curious if that will cause a sales spike or not (I have no idea where in the store it’ll be featured). If so, the slow trickle of sales combined with the chance for a feature will still make it worth doing WP8 versions from now on.
Fluid SE is going very well before the recent issues popped up. It’s still 3-4 solid days of work from being done, and it controls amazingly well with the virtual controls. Better than the XBLIG version in some ways, due to some neat little tricks I’ll explain later. The level improvements and additions have also made it that much better.
I still have no idea how to describe the game. It’s like an analog Pac-Man time trial. Is that an arcade game, a racing one, an action game? There are some puzzle elements too (figuring out the optimal path). Seriously, if you have any ideas on how to describe it, please share…