Slashdot Mirror
Microsoft Says PS3 Linux Not 'Competitive' To XNA
nz17 writes "Gamasutra has a preview of its upcoming interview with Dave Mitchell, Director of Marketing for Microsoft's Game Developer Group. In the interview Mitchell dismisses Linux on the PS3 as a game creators' solution and has said, 'What we [at XBox] are focused on doing is providing great tools at a free or low price point that are going to enable consumers to be absolutely successful at creating games for both the Windows and the Xbox 360 platforms.'"
I see Linux on the PS3 more as an opportunity to get a cheap cell cluster than for game developement.
It may also be an excuse for Sony to avoid customs fees, because now the PS3 is a usable Computer as
compared to just a video game.
***Quis custodiet ipsos custodes***
As the target of XNA seems to be both the professional and the home-brew-market, can the Free Software camp beat this? Well, we already have quite a few game libraries, heaps of engines and a number of IDEs. I'm not aware of any FOSS-'game asset pipeline management tools', and targetting consoles (outside of the Linux-on-the-* projects) has always been something for the big players due to licensing fees.
What is interesting is their idea of having various 'starter kits' for certain types of games (FPS, RTS, platform), all using a common framework. Using them you could quickly get nice results. Is anyone aware of similar FOSS-projects? Might be interesting to build something similar on top of pygame.
This sig is intentionally left blank
I'm currently developing using the PS3 cell. I suspect that the reason that the RSX chip the currently 'locked out' by Sony is because there is a lot of proprietary stuff in there to do with piping data into and out of the RSX, some of which doesn't apply to the PS3 anyway.
It is perfectly possible (and likely) that an OpenGL accelerated library could be released for Linux using the RSX that does not expose the proprietary stuff.
By the way the power of the thing is crazy - you can do 6x real-time HD (1920x1080) MPEG-2 *encodes* (one in each SPU) no problem. There's going to be some pretty cool stuff out there soon.
I've been thinking for a while that this really is the direction GNU/Linux needs to go into if it wants a reasonable selection of games, and it's perhaps something Sony should consider assuming they're serious about homebrew development and haven't just put GNU on the PS3 as a tax dodge.
What's needed is a cross platform framework into which modern games can be developed with relative ease. Preferably this should incorporate some basic engines covering a wide range of game types so the programmer can concentrate on the art and logic rather than the technical details of 3D accelleration. Libraries could exist containing free/adaptable art, and simplifications of certain types of logic (for example, for bots.) Over time, such a system would become increasingly useful as more and more people contribute to the libraries to scratch their own itches or improve games they've obtained and wanted to improve a little more.
I've been really impressed with the Unreal frameworks, and while they concentrate on a specific type of game, it's not difficult to see how the idea can be extended. Buying "Unreal Tournament" (any version) is not buying what's on the box, there's a wealth of homebrew stuff that's freely downloadable and frequently better than the games Epic, and its competitors like id, come up with.
With Java entering the GPL-domain, a significant part of the low level stuff would be implemented (and systems like Jake2 prove that Java is a practical, fast-enough, VM for real time 3D games.) There are enough stillborn projects on SourceForge et al to prove that people like writing game engines, robots, and other components for games - the problems tending to be that the people who write one component get bored when they realise they have to write the other bits. So the skills are out there. To some extent, the technology is out there. What's missing is the integration and the coordination.
You are not alone. This is not normal. None of this is normal.
The Tao project is working on porting XNA for the The Mono project (which is a port of .Net). So theoretically, you could write it in XNA, and run mono on the PS3/Linux and bring any XNA game to the PS3/Linux. At some point in the future. Check it out.
http://www.taoframework.com/Mono.Xna
Would it be possible to start developing XB360MC with XNA and if it is does it cost anything?
The single, solitary, lone reason I own an XBOX is for XBOX Media Center. Nothing else comes close to the TiVo-like integration and Firefox-like expandability.
Unfortunately, the built-in 733Mhz. processor is the limiting factor. Friendtech used to sell a 1.4Ghz. Celeron upgrade, but they don't offer it any longer (if anyone has one of these systems, by the way, I'd be happy to buy it off you). At this point, the XBMC developers are looking at ways to use the onboard GPU to do some of the calculations and take all the heat off the CPU, but this is like looking for breadcrumbs when right next door is a fully stocked kitchen pantry (XBOX 360 or PS3). Instead of wasting time developing for a platform that's basically a dead-end, they could be working on a system that will be able to handle HD content by design.
Would developers have to port MPLAYER to C# in order for this to work? It sounds like the PS3 is a lost cause until someone can whip up some Linux drivers for the graphics system. Does Microsoft have similar restrictions to the hardware?
Well, if you figure that the Xbox 360 is 1 year into a 4 year lift span, it's 25% over (if it really sells well, it'll get a fifth year, overlapping with the Xbox-3). Getting XNA games out now, for hobbyists, let's them tweak and expand their development tools, get paid doing it (the subscription fee probably covers the cost of supporting the project, not much of a profit center, but should make it a bottomless pit of costs), and does so without the expectation of the pro shops.
The big development shops wouldn't tolerate things being imperfect, hobbyists will. For the next generation, expect some XNA derived system to be one of the main development environments for the console.
There is a long tail effect with gaming. People buy consoles for the oddest reasons. If you are a niche system like the Xbox was last round (people wanting HD games, or the absolute most powerful system, not looking for bulk of games like the PS2 was), you can suffer. If you are a shop making an innovative game, that won't sell HUGE numbers, only the PS2 had the volume (unless you were using a Nintendo character license, then the Gamecube was a compelling system) to sell.
Things like puzzle games are popular, light on the graphics, and can be implemented cheaply. If Microsoft figures out a way to sell $20 games online (taking $5-6 as the publisher cut) without involving retailers, then independents will flock to the system, Microsoft will make tons of money, and they'll position themselves for huge market-share gains.
However, the XNA, at this point in the Xbox 360's lifespan, is simply a test. Real development houses with 1-year + dev cycles wouldn't adopt a new system now, but hobbyists and indies will, and what works will roll into the Xbox-3's dev kit.
Alex