The UK based magazine Edge who go to the four corners of the earth in pursuit of games (hey we've only just got the Xbox and GC isn't out till May 3rd FFS!!!) is probably the most respected games mag in the country and is the choice read of all the UK games developers themselves so it's opinions hold a lot of sway and are respected as fair, honest and above all independent from format bias. Like the majority of multi-format mags they were in no doubt that the Xbox has the best line up of launch titles of any console ever, including the GC. It can't be argued then that 'there's only Halo'. But the battle has only just begun and we're going to have to wait at least years before anyone can say which console has the best games as at this stage any claim either way is ridiculous and without foundation due to the development time of A-list games.
But this thread isn't about games, the topic was who's hardware is best? It's an interesting one and it's almost impossible to answer. Why? Well because the consoles all use completely different hardware bases. The Xbox is based on the PC, cisc, windows, nvidia, etc. The PS2 is based on supercomputing technology and vector processing (I could be making this bit up but I think it's similar to tech used in Hitachi's supercomputers), incredibly powerful but until recently not seen much outside of Japan and ironically shares similarities in concept with the N64! And then there's the GC (i thought I'd bring it in just because it shows the full facet of computing hardware) that is based on Power risc architecture as seen in IBM's RS9000 UNIX systems and Apple's desktops.
That's 3 consoles, all vying for the same audience but taking there foundations from the different sides of the IT industry (maybe i'm just sad for finding that interesting!) How can you compare a supercomputer to a personal computer or a mid-range Unix box? You can't unless of course you're talking about the end result, the actual real world ability of programmers to make the box do something. There's no doubting the awesome performance of the EE as the most complex processor in a desktop machine anywhere. It is a collosus but what I could never understand is how they could couple it up to such a weedy front end (I'm on about graphics here). Despite it's superior performance the PS2 was never able to outshine the Dreamcast when put in a head to head for graphical flare and compared to the NV2A in the Xbox it doesn't stand a chance.
That is where the glory lies in a games console and it always has been. It's not what's under the hood but the end user feeling that shows a console's true performance and the Xbox has the better graphics and also lets not forget sound thanks to the MCP and integrated GPU into the Northbridge of the N'vidia chipset running the machine which is perfectly balanced with the performance of the PIII and system bandwidth, that's the key 'balance'.
The EE simply costs too much to manufacture because of its complexity which means Sony has to skimp on the peripheral components in order to sell the console at a reasonable price wihtout making a loss that is unrecoupable in games sales. 4MB in a graphics card is laughable. No matter how you cut it textures take up space, say one texture using a 32 bit colour pallete (real colour) is 256 pixels by 256 pixels (As would be seen in todays cutting edge gaming) That one texture would take up 2Mb of space (Be careful to remember the difference between bits and bytes here!) Which means that the PS2 would only be able to hold 16 different textures in its memory! Think of 16 different objects in a scene (roof, floor, walls, table) and now think how pathetic a total that makes. And the PS2 has no way of back buffering those textures in main memory. What isn't being stored locally has to come from the DVD which only has a very restrictive data bandwidth relatively speaking, there's no partionable system RAM and certainly no bufferable hard drive to fall back on when space gets short. When you cosider that and the fact that the DVD drive in the Xbox is considerably faster than the one in the PS2 it all stacks up against the machine.
No amount of vertex manipulation is going to make up for throwing about glorified wire frame models. That might be suitable for graphics artists but not for computer games.
The PS2 is underpowered and Sony knows it. When discussing die shrinks of the machines processors Sony revealed that with the associated cost savings it would be possible to increase the integrated graphics cache on the graphics synthesiser from the original 4Mb to a much more respectable 16Mb. The extra texture cache would enable for the PS2 to stand on much better terms with the Xbox and then considerations of AI and physics might come into the equation, but who wants brilliantly intelligent computer controlled players with the most realistic Newtonian physics model ever conceived when the graphics look like the original elite? (I may be exaggerating a little here)
And all this discussion is missing one vital point games designers have to ask. How am I going to use it? Well the simple answer to that is through API's. The development of the GameCube has focused heavily around two things, 1/ The control pad and 2/ It must be the easiest format for a coder to programme and do what they want to do on it. The GC has fabulous API's and is a joy to work on, as is the Xbox which uses DirectX and also versions of PC hardware familiar to coders everywhere whether they're from a Linux, Windows, BSD or whatever background. They know the Xbox format and they know how to programme for it and make them special. How many programmers, on the other hand, have programmed for supercomputers? Worked directly with them? Not many that's for sure. When the PS2 came out developers went 'WTF?' The VU code that is so important to getting the most from the machine was a completely new venture for most coders new to the PS2 and only now, over 2 years since PS2 development kits were first sent out are the developers truely able to say ' yeah i think i know what I'm doing here' with any real conviction or honesty as demonstrated by the lack of any really gob-smacking titles until the release of Gran Turismo 3.
The Xbox has been blighted by comparisons to the PC from all quaters, but if you're a programmer and you're making a living on the PC and you want to go into console games for the first time with your revolutionary product that's going to change the way we play games (and the PC is unarguably the best place to find creativity and originality of thought as it's free of the developer restrictions imposed by all console manufacturers, for all those that don't know it was almost impossible to get a licence from Nintendo to publish games on the N64), what format are you going to choose? Xbox with it's familiar friendly API's or PS2 with it's phenominal potential for intricate and diverse game dynamics but only after 18 months experience (,... or of course there's always GameCube, friendliest of all consoles, focused strategy, power above PS2 but slightly lower than Xbox, but just try getting that licence ;-) ) ?