Today’s blog is the second video game blog I’ve featured. The first, Tales of the Rampant Coyote, appeared in May with a follow-up interview a couple of weeks later. I found today’s by accident. Curiously, it appeared in a Technorati…

Flash penetration by version:
http://www.adobe.com/products/player_census/flashplayer/version_penetration.html

(94.8% U.S. penetration for version 7. Version 8 is the most recent version)

And here’s the numbers for competitors:
http://www.adobe.com/products/player_census/flashplayer/tech_breakdown.html

Shows Java at 87.3%, which is higher than I expected, but no breakdown by version. My guess is the Java # includes a fair portion of Java 1.1 machines, whereas the latest version is 1.5.

Flash Lite entered the scene a while ago with a version that supports AS 2.0 and other features. However support is still trickling in and it may be a couple years yet before enough phones support it to make it a viable development platform. And then we’ll see how well it ports. It’s interesting that Java is supposed to be “write once, port everywhere” but ends up being just as hard, or harder, to port than BREW because it is implemented so differently on different phones (and because of the many resolutions supported). Flash may end up with the same problems.

Also, I’m still amazed that game development on such limited platforms is really done in Java. I can’t imagine trying to do it in Flash.

The newer versions of Java are pretty good (1.4 and especially 1.5), but penetration of Java versions beyond 1.1 is limited (1.1 was the last version that Microsoft bundled into I.E., though they stopped doing that a few years ago).

So you can hit ~80% penetration with Java 1.1, but with limited features, or a better feature set with 1.4 or 1.5, but terrible penetration.

That’s why Flash has become the defacto choice - ~95% penetration for versions only a generation or two old, with reasonably robust feature sets.

I think you tend to get hit really hard if you have a low-end phone these days because of the politics of creating a mobile title. Game companies want to make good games and so they tend to target the better mobile phones to start building on. However, carriers won’t accept titles unless you can hit a certain %’age of their phones so developers have to make games that work on a lot of phones. Also, some phones that are crappy for games, nonetheless, sell fairly well. And so they want you to hit all those. A lot of the time the result is that the company makes one really good version of their game, for their target phone, then makes some crappier versions of their game for other resolutions and then sends it all off to a porting house. The porting house doesn’t care about quality — it cares about meeting the requirements of its contract as cheaply as possible. Generally the crappier/older phones are also real pains to port to and a lot of corners may be cut to make the games work on these phones. So the result is that the crappiest phones get games that are twice-copied from the original with a marked reduction in quality each time. Get a better phone and you’ll probably get better versions of the same game. Don’t expect anything to compete with your PC yet though.

Why BREW hasn’t taken over:
Mostly it’s a VHS/Beta sorta thing. Java is the standard in Europe and other places where people use their cell phones a lot more than we do here in the US. And so people are used to that. On top of that, the whole infrastructure of BREW is very complicated and has a lot of bizarre externalities attached to it. Java games can be directly downloaded and put on the phone. BREW games are signed and a pain in the ass to get on a phone. The BREW API itself is “ok” but generally I think it is a lot easier to get a Java game up and running (although people I’ve talked to claim that BREW is easier to port). Our games are very heavily connected, talking to a server all the time, and we have found that BREW is a real pain in the ass for what we want to do and in general I think it is going to have a hard time “growing”. Part of the problem is that any API changes to J2ME or BREW aren’t used for years because it takes years for the last phones that use the old API to cycle it. It is too costly to have to port between different versions of the API for different phones. BREW has released 3+ versions but generally developers are still using 2.1. J2ME, for us, is only just now getting to where it is possible to use MIDP2 instead of MIDP1 and MIDP2 is still lacking a lot of functionality. If BREW or J2ME releases new networking libraries, new graphics libraries, etc., it will still be several years before phones catch up and those can be used.

The control mechanism will likely always be a problem, but when you start designing games with that in mind, you can actually get things that are quite fun.

From a marketing standpoint, I assume that the companies running the cell phone business would do things like rentable ring-tones (I’m amazed at the amount of idiots who are willing to pay $3 for a ring tone they can only get for a month), but from a tech standpoint, there’s absolutely no reason why they couldn’t spend more time on better audio/reception technologies.

As a side note, I’ve purchased (well, my wife did) a Nintendo DS this weekend. After playing around with the stylus, there absolutely no way I can ever see me going back to playing any other handheld game without it. Cell phone or not.