Why Sun should make a PS3 JVM

General Discussions
#1

I work for a games company. We have a game based on a (shitty) C++ engine that we know that we will re-write for future releases / new products. We are looking at all sorts of options for developing the new codebase. I’m pushing for Java - and it’s got everything we need for our game to run (and to be way more efficient to develop and modify), BUT the killer issue is the question: “what about running on the PS3?”

Chris has said that people need to develop a killer game in Java to provide motivation for the development of a PS3 JVM, but this is a stupid chicken and egg situation - people won’t develop those games because they won’t write game titles that need to be cross platform in Java in the first place. Consoles are an important platform to the game developer!

This really pisses me off because I’m a Java advocate and I’m the position to dictate the technologies we use, but I can’t push a tech that won’t work on at least one of PS3 / Xbox 360.

IMO, Sun NEED to get this done! Java is unlikely to get large scale support from game developers if you can’t target consoles with it.

#2

The money to be made for Sun in that market is close to zero.

As ‘JavaGaming’ on the desktop has had very few actually released games over the years, dispite many manhours put in JOGL, they might not expect any success on the consoles anytime soon, it’s just a tiny group of people that would (be able to) take advantage of it.

Why would Sun, as a business, invest millions here? Not to mention the license troubles, businessmodels of Sony and Microsoft and new APIs to access the special gfx hardware.

#3

to grow new markets

#4

That won’t yield any cash, only goodwill in a tiny group.

#5

The gaming market is huge. If you can’t make cash out of that, you can’t make cash out of anything.

Sun could make a PS3 JVM and charge for it - there’s a market. Games developers will put out hundreds of thousands for middleware, you could get at least a few bucks for a PS3 JVM. We would certainly pay for one.

#6

Well, since Java is open source you can port it yourself or pay someone to do it. There are also other alternatives using GNU Classpath project, like GCJ (if GCC is able to produce code for Cell, haven’t checked), JC also looks interesting. Using these alternatives won’t be easy though.

Are you planning to create PS3 version for sure? If it’s just some hypothetical future option, then you could stick with Java. You’ll probably find that you need to special version for PS3 anyway. The architecture is different, you’ll probably need more lowlevel access for both speed and memory reasons. Otherwise using C++ is probably better idea :slight_smile:

As for XBOX 360, it can run C#/.NET and there is IKVM that allows running Java on .NET. Don’t know if it’s applicable to XBOX 360 though.

There is also option to use C# directly which could be good compromise between C++ and Java.

#7

wow! I didn’t know about that. It can even run eclipse! Thanks jesek2.

#8

C# basically is Java. I’m surprised nobody got sued. ::slight_smile:

#9

I’d love a proper PS3 JVM (and OpenGL & OpenAL APIs…) because then I’d get my games onto PS3 finally, and actually make some money. As it happens instead I’m having them all ported over to C# they’re going on XBox instead. What a shame.

Cas :slight_smile:

#10

What are you saying? What will your next game be written in? Straight to C#, or are you staying with Java?

#11

I’m staying with Java because the Mac means more to me than XBox. Of course, if the games sell well on XBox I might rethink that position.

Cas :slight_smile:

#12

I’m thinking your games would run excellent on consoles. My kids love them, but I only have laptops so it’s overly complicated for them to play, - they’re more comfortable sitting in the sofa with the xbox. So I think they’d be a big hit :slight_smile:

#13

Mmm. Who’da thought “write once run anywhere” would turn into “write one, run hardly anywhere”.

Cas :slight_smile:

#14

That’s exactly my point - people actually wanting to make money out of games - even in the Indy space - need to be able to hit console targets. By not having a console JVM Sun pretty much ensures that serious money will never be invested in Java games. What’s worse, Sun is driving developers in that space to C#. That’s really crap.

#15

I agree there is no real reason for Sun to make a PS3 JVM atm and in the current situation.

Actually, I think Sony should create a VM, preferreable the JVM of course :slight_smile:

Sony will need to put s.th up against X-BOX360 C# development/XNA eventually. Nice Java based PSN titles would be a great start.

#16

To be honest the arrival of the OpenJDK might prompt someone to get Java on PS3 (and with a bit of LWJGL magic there should be a fairly short path to getting games ported properly to PS3) but it’s undoubtedly quite a lot of work. Having said that I’d pay to license a supported JRE for PS3. Maybe a couple of thousand dollars per title is in my budget.

You’re still then in the position of needing to have a PS3 dev kit and go through Sony approval etc. but at least there’d be very little by way of actual code porting that needed to be done. Well, none at all in fact, that being the point. Maybe a few tweaks here and there.

Cas :slight_smile:

#17

You’d have to make the entire game comply with their TRC though - that’s a decidedly non-trivial task.

#18

There is already the BDJ VM on PS3, which seems quite a good start already. They’d just need to bring it to a full JRE implementation, which seems not such a far off goal.
Performance wise I’m quite sure it could run all puppygames titles at full speed if there would be OpenGL support.

#19

If an important enough company turns up with a good enough game, I’m sure that allowances will be made.

If you want the fastest most effective way of pushing this forwards then I would suggest going and getting a job working for a giant of the casual games space (perhaps PopCap? or one of the new ones, like Playfish?) and trying to get them to go to Sony with proposals.

#20

It’s usually non-negotiable. It’d have to be a very, very, major game.

However that’s not my point - my point is that even if there was a complete JRE there would still be substantial porting work needed. A vanilla java app isn’t going to cut it.