Addressing Oracle's lacking attention to #javaforgames

You should read this (yes, the whole thing. Grab a drink and sit down).

And then you should reply to this, including the #javaforgames hashtag:

2 Likes

Also perhaps vote for the article on hacker news =)

http://news.ycombinator.com/item?id=5347071

I don’t get this.

I’ve seen retweets, read badlogicgames’ post, but I still don’t get what this is.

Is there some kind of demand going around, asking something out of Oracle? Because if that’s so, there’s no manifesto I could find, nor is there any hint as what it is exactly this hash is suppose to push Oracle towards.

I am sure they noticed some games were made with Java. (If they don’t then they all need to be fired. Right now.)

So, what EXACTLY is this supposed to entail? If you want to make demands, just be unequivocally clear and direct. As I see it, the message isn’t coming through. All I read is “I develop in Java because… [some reason]” and I’m still searching what all this have to do with Oracle…

Basically making what we do with avian (compiling java code to native) supported by oracle, or rather the prerequisitions for that.

Don’t think that compiling to native is THAT important.
It’s eclipsed by Oracle’s decisions about what the platform can and cannot do.
My interpretation of the article is that Oracle should make their mind up what they’re gonna do with Java. Because it hurts the gaming community if you don’t know whether the platform will still be viable in two or five years.
I guess other communities have the same problem. Of course, since the gaming community is relatively small (compared to the corporate world anyway), the risk is higher than for, say, the Spring and Hibernate people. So the REAL thrust of the article seems to be “hey look Oracle, you might want to consider the gaming community, too, because that could become another important community to entrench the platform further”.

I doubt it’s getting much traction with Oracle though. Oracle’s mission, first and foremost, is making money. Making the platform applicable for more communities is just a small building block for that, and will always take a back seat.
I think we can be happy if they don’t discontinue JOGL.
A lot will also depend on how much traction the Java gaming libraries can get in the gaming world. Getting more traction there is going to help JME’s case more than appeals to Oracle IMHO.

Just my 2 cents.

@toolforger said: I doubt it's getting much traction with Oracle though. Oracle's mission, first and foremost, is making money. Making the platform applicable for more communities is just a small building block for that, and will always take a back seat.
That's why compiling to native is important... Right now that' our only way to iOS
@nehon said: That's why compiling to native is important... Right now that' our only way to iOS
Imho, the right way would be to support projects like VMKit(http://vmkit.llvm.org/), because they basically solve your problem. The point with native compilation is, that you basically loose all the nice debug abilities you have with the vm. Plus the actual problem with java is building a garbage collector which is fast enough to allow for real time applications like games.

Honestly, I still dont get why native compilation is even an option. I mean you run into problems like guarantees of your memory model versus the memory model of C and stuff like that and I dont see why one can not use OpenJDK.

You would use OpenJDK and oracle would think about this use case, thats the point. And btw you debug while you test an app, not when its deployed.

@kardinal said: Honestly, I still dont get why native compilation is even an option. I mean you run into problems like guarantees of your memory model versus the memory model of C and stuff like that and I dont see why one can not use OpenJDK.

Native compilation (AOT - Ahead of Time) can follow same memory model guarantees like normal JVM/JIT. Major benefits of AOT are:

  • lower footprint (especially important for embedded devices/phones)
  • easier, self-contained distribution, independent on specific installation on user machine
  • considerably faster startup on embedded devices (no need to waste precious battery power on JIT compiling things over and over)
  • political (Apple does not allow interpreted stuff on iOS)
  • theoretically better compilation for extremly cpu-hungry optimalizations (vector code etc)

Only thing you really lose is adaptive compilation - but this really happens only on server Hotspot and with smart AOT you could get 95% of that benefits with profile-driven compilation anyway. Plus possibly on-demand class loading if you are running on iOS (in normal systems AOT jvms allow lazy loading).

1 Like
@madjack said: I don't get this.
The letter and my little twitter campaign were sort of two separate things that conveniently happened at the same time. Both have been very successful. The campaign allowed me to connect with several Java game developers, small and large, and the letter has gotten us in touch with the right people and useful discussions are happening - unfortunately where companies are involved, so are NDAs.

Glad it worked.