Lots of doors are being closed for Java

Remember that J2ME is more than CDLC/MIDP.

pjt33, what are you suggesting? Java TV, Java Card, CDC, and whatever else are the same shit to me.

To me it’s all the same stuff between J2ME/Android/J2SE. Sure they have some custom classes and even some system class differences, but those changes have little to do with my main code.

There are places where J2ME is still popular, but this is mainly as 1) cheap ad-hoc games provided by phone providers just so they can say their phone ‘supports games’ and 2) in emerging markets where smart phones are too expensive and the networks cannot handle the required data bandwidth. Both of those are in decline.

The reality is that 99.9% of people with a J2ME enabled phone never play games on it.

They certainly never actually buy them. Or maybe they do, but honestly I’m not interested in developing the sort of game that J2ME is capable of.

Cas :slight_smile:

About LLVMs VMKit:

Very interesting! I’d never noticed that feature. Assuming that it works reasonable well, then that’s the hardest part solved (or being solved) by someone else. So this is potentially a viable option for games that avoid the “problematic” classes from SE. I’m not motivated to check, but it would be very interesting to see if someone could get a basic LWJGL demo running in native via the AOT compiler.

@pron: I can’t see any reasonable reason to use Erlang for game programming. I guess one could argue that for MMO games you want a server with 99.999% uptime and that the bulk of the server time will be updating entities (actors). But so what? One can hit these requirements via other languages. I actually “like” erlang, I just don’t think its usage would be reasonable. On concurrency, you have to remember that the term is somewhat overloaded (Java vs. Erlang concurrency is two different animals). For modern hardware the optimal processing throughput it typically hit when the number of active threads (running in a given timeslice) is some small constant multiplier of the number of effective CPUs. Around “2” is a good guess for current hardware.

The new Battlestar Galactica MMO uses Erlang for the server, and I believe (although don’t quote me on this) that it uses Unity for the client. It can handle very high number of threads, it’s concurrency model is far simpler (and easier to use) then Java’s, it has in-built support for deploying across multiple servers, can hot-swap code and has support for coping with failing processes. IMHO there isn’t any reason not to use Erlang for server-side, especially if there is an emphasis on high-numbers of connections and scaling horizontally.

That’s not to say it’s a better choice then Java for the server, just that in some situations it’s at least as good.

I have 8 action games and an RTS game and I am working on a multiplayer game and all are very fast on a fast J2ME phone.

The next step will be putting all 3 together in one game. 3D, RTS, Multiplayer. If that is not capable I don’t know what is.

That Cas, at least, seems to be conflating all of J2ME under CDLC/MIDP. CDC/PP is no more painful to use than J2SE.

That little blog post got a lot more popular that I thought it would ::slight_smile:

HTML5 was indeed a bit too cutting edge for my taste - regardless of performance, just appeasing all the different browsers was a royal pain in the butt.

Last J2ME contract I did was in 2005, and it seemed like it was on the decline back then too.

Maybe it’s still popular in some remote parts of Africa or something, but I would certainly not suggest anyone to get into J2ME dev these days.

Is the Tiobe Index http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html useful at all in terms of gauging a programming language’s health? It seems there is a bit more daylight between Java and C & C++ right now than in a couple years. Also shown: C# & Python on the rise (how much of that is Jython?).

It wouldn’t surprise me if it’s pretty close to reality - java is a very popular language, nobody is disputing that.

It’s just that that’s because most software out there is server/web based in-house stuff.

Java for games has never been popular and I don’t think it’s really getting much better.

Sure no one plays Runescape, AllBinary, Puppy, or other J2SE, J2ME, or Android games… I must be imagining things.

He refers to developer mindshare.

Cas :slight_smile:

Hi guys,

While we’re on this topic… I’ve been wondering lately whether Mono .NET wouldn’t be a valid alternative for gamedevelopment.
As it’s also crossplatform, c# ressembles the java language very much and seems to support all the things we’re missing from Java:

Now I know you guys are prob all in favor of Java, or else you wouldn’t be hanging around here :wink:
But I was wondering if any of you expirimented with this? And if so, what did make you come back?

Thanks!

Martijn

Sure, there are some successful games done in java (I’ve made most of my games with java2d, pulpcore or lwjgl too) but not as many as you’d expect from arguably the most commonly used language in the world.

Indeed, it’s laughable how few games are written in Java. It’s like there has been concerted, co-ordinated effort in preventing it from occurring. All that JOGL bollocks didn’t help, probably deliberately designed to scupper chances of success by dividing the tiny community.

Cas :slight_smile:

I think a big part of the reason is that Applets sucked too much, for way too long (and in some circumstances they still do).

If we’d have the applet experience you can now get with a modern java plugin on chrome under windows or linux (os x is still on the old plugin) 10+ years ago I think you’d see a ton of java games out there today.

What, even with the incredibly obscure rituals you had to perform to get Java2D performing vaguely fast? On some hardware? Java2D is a horribly complicated API, that’s why I turned to OpenGL in the end! More irony. Anyway that ship has sailed, but it still begs the question: why no games being developed in Java now? And the reason always boils down to the same things: no console support, no phone support*. Although I’m not entirely sure why indie developers don’t embrace it a bit more as it’s ideal for desktop games development, as I’ve repeatedly shown.

Cas :slight_smile:

  • sorry tberthel but those games you’ve got there look and play like something from the 80’s not the 90s or even 2000s, minus a decent user interface.

Yes, I think we’d just have gotten something like PulpCore earlier if Applets wouldn’t have (for all practical purposes) died years ago as far as web games are concerned.

The lack of developer interest also means it took a lot longer to get decent game libs and tools (e.g. see how quickly html5 game libs and frameworks popped up within months of canvas/audio becoming usable for 2d games on cutting edge browsers).