Is There Anything Wrong w/ Playstation 2 API? 9
adubey asks: "Over and over I've heard game programmers say that it is difficult to write games for PlayStation 2. A lot of these quotes came from gaming sites or the mainstream press, where there probably aren't as many programming-literate readers as there are on Slashdot. So perhaps Slashdot might be a better place to ask ``what is wrong with the PlayStation 2 API?'' Are there any PSX2 programmers here who might be able to give some insight? Can this even be discussed without breaking any NDAs?" This is the first time I've heard of such a thing. Where did this rumor get started?
Re:Sedcondhand report (Score:1)
I'm an advid gamer but only a entry level programmer, IMHO(and god is it humble) there are a lot of tittles thats seems superb for the ps2 (MGS2, FFX, ect..). Sony is throwing a lot of money into the production of the ps2 and sony has a good track racord for user end support (they still supply part for beta.). The technical difficulties can be overcome if the user base grows and more componies familiarize themselves with the new achitecture. I have heard that many companies have avoided the problems with jaggies by simply rendering at higher resolutions(with the resulting drop in frame rate ect..). The documentation for the developer might be horrible but I hear(so much second hand info) that sony is reviewing and redistributing new docutmentation and developers kits. The texture ram problem is problably the worst.. I can't think of any way they'd overcome that but then again the quality of the graphics aren't always intergral to the quality of the game.
Reference (Score:1)
Playstation 2 and Linux (Score:1)
Re:Hardware Managment (Score:1)
Hardware Managment (Score:2)
Ars had a great technical article long ago... (Score:2)
Have fun!
Poor documentation and bizarre architecture (Score:2)
Anyway, his beefs were three-fold. First, the documentation given to them by Sony was bloody awful. The translation wasn't so hot, and some parts ranged from confusing to flat out wrong. Some documented functions just didn't behave the way the docs said they would. It's hard coding blind.
Secondly, the architecture is overly strange whereby established procedures just couldn't be used. There are multiple paths one can take to accomplish and no one really knows which ones work yet since the design is so foreign. I got the feeling that prior console or PC programming experience was probably a bad thing, since the way you previously would have done things was no longer valid. These sentiments were echoed by Dr. Buchanan, head of Research at Electronic Arts at a presentation at the University of Alberta last year.
Oh, and the third thing? And I quote: "It doesn't have enough texture memory to render my ass, let alone a good game."
Different (Score:1)
From what I've seen of the various platforms, programming the DC isn't much different from programming a PC. You move polygons with the CPU, the download the textures to the PowerVR (PVR hereafter) chip and the send the polys to the PVR and say go. Essentially, it is like having a 3D API, except you have to use it via assembly calls instead of a nice DLL. This isn't a hard adjustment for someone used to OpenGL.
The PS1 doesn't have a nice 3D accelerator to use. Instead your have to write your own renderer, but you are given a chip that will help with some of the 3D math involved. This is not unlike programming on a PC before hardware acceleration was common. However, Sony supplied a library (like OpenGL I guess, but it isn't available for public inspection) that made life simpler for the first generation games, then later generations scrapped the library and essentially worked on bare metal.
Now, I don't have any personal experience here, but it seems to me that programming a PS2 is somewhat like programming the PS1, except the architecture is a lot more complex. Instead of 1 CPU, we now have 3, one of which is a 1/10 the speed of the others. The pipelines are more complex and the two main CPUs are similar, but they aren't even really the same.
And, unlike the PS1, the PS2 reportedly didn't come with OpenGL like libraries. Instead, the first generation games must figure out the poorly documented internals of the system all at once instead of over several generations of games like they did for the PS1. While this may be heaven for some guys, the average programmer would find the a nightmare, and all but the best will struggle here.
Previously, systems like the PS1 were a lot more flexible than systems like the N64. However, looking towards the future, vertex and pixel shaders implemented in hardware like Nvidea is doing should give programmers nearly as much flexibility while remaining much simpler to figure out. While I think the PS2 will suceed, I doubt that many more platforms will persue such a complex architucture in the future.
Is just a matter of time (Score:1)
About the vram limitation, if you read some of the articles about the architecture you are supposed to use all that bandwidth to feed texture bits to the GS, unlike a pc where you have obscene amounts of DDR memory to play with because the bus speed is still a joke.