I think the key difference here in terms of hardware for Xbox2 is the fact that Microsoft wanted to play a bigger part in deciding on the hardware. In fact, this time around, they're in full control, whereas the last time, basically nVidia did the entire system. Microsoft definitely makes great developer tools and documentation, and it would be silly to think that XNA will not amount to much. But the thing is that they are clearly without any sense at all when it comes to hardware. The fact that they choose to centralize their FSB or share a single L2 cache among 3 processors shows some real lack of insight. The biggest flub would have to be that 10 MB eDRAM on the GPU -- which I'm told is really MS's idea (both MS and ATI told me that much) -- that just says they didn't even think about resolution.
Hardware-wise Xbox2 is getting disappointing the more I look at it... and I know I shouldn't really be saying that since I'm actually developing on it. But by disappointment, I should say -- it won't suck or anything remotely near sucking... it just won't be anything monstrously groundbreaking -- let's just say it's Moore's Law looking perfectly normal and healthy. The GPU will have some serious power all right, as will the CPU. But if you think of the difference between PS1 and PS2, you should see about the same growth from Xbox to Xbox2, but at the same time, taking into account the difference in resolution, content, shader complexity and everything else put together.
All in all, Xbox2's total system architecture is still very PC-like. Or Mac-like in this case. In that sense, it will probably be easier to develop for. The thing is that SIMD is very important to getting any major performance out of PPC processors these days. Without it, they're basically Celerons. So avoiding pipeline stalls and concerning yourself with *instruction latency* is going to be huge on all 3 consoles with this upcoming generation. In some ways, that actually means we've gone back to the '60s in terms of programming. It's just that it's the 60s with 3 million line codebases.
BTW, I should also note that based on what I'm hearing from every studio I've been to, I'd have to say that, at least for the first generation of games on next-gen consoles, you will not see anything running at 60 fps. There is not one studio I've talked to who hasn't said they're shooting only for 30 fps. Some have even said that for next-gen, they won't shoot higher than 30 fps ever again.
As for PS3... well, it looks as though PS3 will be the hardware king this time around. Just as Xbox had the powerful hardware in current-gen. The question of whether it will be that easy to develop on is still up in the air. Developing a cross-platform engine for Xbox2, PS3, PC, NRev will likely be damn-near impossible. I'd expect a codebase with more #ifdefs than code, the way things look. I don't expect graphical power to be that massively different between PS3 and XBox2, though. They're essentially very similar GPU pipes... PS3's will probably have some features that Xbox2's doesn't and vice versa. Where the difference will lie is going to be in CPU power. How that will manifest itself is still very uncertain. If the graphical power is smaller, that may also mean there is more value for studios to do multi-SKU titles, and we may see a bigger glut of games that are available for all 3 or at least 2 of the 3 consoles.
Then of course, comes the business hand of Microsoft... in this Sony is definitely second, but I wouldn't count it as a gaping hole. I mean, Sony is an 11-figure yearly revenue company, too. If nothing else, PS3 can drive home the point of CELL as an architecture. And in that sense, Sony is playing for the longer term than MS. OTOH, MS is the sort of company that can afford to play out a generation at a time.
That tradeoff, framerate versus complexity, will exist for some time yet.
(or so I hope)
