Slashdot Mirror
Developing for the Playstation 2?
Kentai asks: "Im thinking of building my final year CS project around developing a piece of software for the Playstation 2 and I am wondering what various pitfalls I should expect. Bringing our own software to the living room on nice hardware seems like a whole new computer medium that would be fun to exploit. Problems I am expecting so far include: issues like obtaining a compiler (pretty important), the whole Playstation 2 CDR issue and legal issues. What compilers can one use to develop PS2 software, and how much does each solution cost? Also, has anyone had any luck or placed any hope on a Java VM for this platform?"
You might want to keep an eye on www.javagaming.org . A few of the Sun Guys Frequent that site, and will sometimes post little bits of information.
Less Talk, More Beer.
From my own senior design projects, let me discourage you from designing a game. We (3 of us) designed a PC game that used DirectX7, used 2D graphics, sound, and was online only (no AI). After working and designing it all year, we could only get people on the screen fighting one another. The amount of work we put in was the most out of any of the projects.
I don't want to discourage you too much, but designing a game with graphics and everything is a huge task. Most people have no idea what they are getting themselves into when they do it, and it can pay dearly on the other classes you will be taking (I dunno about you, but the last year of college was the hardest classes, all requiring multiple programs to be done all year long). If you want to design a game, may I suggest something similar to a MUD or MOO? You can put more features into it, and people can have fun playing it. But getting a game working (with graphics) on a PS2 will require a ton of time, and a large team of people.
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
I am wondering what various pitfalls I should expect.
Don't worry, Pitfall hasn't been released for the PS2.
Python has been used for commercial games on the PlayStation 2. I don't think that the development tools are on the Web because of Sony licensing. But if you know who to talk to and where to look, you can probably find them.
The dreamcast has an EXCELLENT community built up around homebrew.
There is an awesome library/miniOS called KOS which gives you hooks into everything, you compile with an SH4/ARM crosscompiled gcc, and it's easy to do once you're set up.
http://dcdev.allusion.net/
The PS2 homebrew crowd is JUST STARTING to get off the ground, but there's still problems of having to hack your ps2 with a soldering iron to get it to boot a cdr/etc, as well as very weak library support.
Homebrew console programming is very fun, but go for dreamcast, ESPECIALLY with how cheap dc's are right now.
There's even ways to get a serial cable hooked up, so you get debug info and can upload your program without burning a disc every revision.
Better support will not be found!
It'll take you ages. Especially if you want to do it on a PS2 (even if you had access to professional tools). Especially if you've got no previous games programming experience.
If you're going to do a game...
a) Write it on a PC - much easier to debug etc..
b) Use as much middleware as you can - if possible, use someone else's engine.
I wrote a game last year called Hunt for the Red Baron in 9 months. I was the only progammer and I had an Artist and a part-time games designer working with me. We used Renderware and a load of our own libraries, and without them I'd never have had time to do it, and I'd already got several years of games programming experience behind me. Believe me, all that boring stuff to load in bitmaps, handle input, etc etc takes ages. Go for a existing engine if you can.
The other option is just to write an engine for your project and forget the game side of things, and just end up with a demo.
cheers,
Tim
That would be me :)
The main thing to expect from the PS2 is that you have minimal to no library support. If you want to get polygons or bitmaps on the screen, you have to manage video memory yourself, you have to talk to the graphics hardware directly, and so on. If you want to really get performance, then you have to write custom assembly code for the vector processors.
The bottom lines are: (A) the learning curve is very steep; and (B) expect to do very embedded-systems type of work, and not be able to rely on fancy C++, Java, or the like.