I don't mind that the PS3 has graphically inferior ports, but I do mind the frame rate issues cropping up in a lot of games. This isn't one of those issues that all devs are going to magically solve -- there will likely continue to be inferior ps3 ports to one extent or another throughout the PS3's lifetime. But they should become less common as time goes on and the installed base justifies additional effort on the PS3 version.
Yet, it's not enough for me to choose the 360 over the PS3, considering the HUGE hardware quality gap between the two consoles in favor of the PS3.
As far as the whole "PS3 vs PS2" difficulty is concerned, MikeB doesn't know what he's talking about and he's just saying things. I don't know if you can directly compare the PS3 and PS2, but one of the big issues with the PS2 was its memory architecture. There weren't any large, fast caches to work with like you typically have with computers (and even consoles) so the big challange was getting the most performance out of the memory you had.
The PS3 issue is different: you're working with an under powered general purpose CPU and trying to augment it by making the best possible use of its SPEs. The SPEs don't necessarily lend themselves to all situations, but you can try to design around them. This makes porting from another platform difficult. The original PS3 design called for a cell with 2 CPU cores instead of 1 and that would've been a lot easier to get Xbox 360-like CPU performance out of. That said, CPU performance is not always the limiting factor and there are many 360 games that don't use multiple cores extensively.
At the most this is a *very* simple comparison looking at the top of the problems. Any programmer worth his salt will tell you that the devil is in the details, not at the most abstract layer. Don't take this as any sort of serious analysis into the PS2 and PS3, but for those unwashed masses this will give you a very generic idea of the biggest issue with the PS2 and the biggest issue with the PS3 in terms of programming difficulty.
The porting issue comes from the set of assumptions one makes when designing a game. If you design a game under the assumptions that you'll have lots of fast general purpose CPU time, or that you'll have large memory caches to work with, you'll make certain decisions. And as you create the engine, you'll make it in a certain way. If you decide to port this game to another platform that doesn't offer one of these things, you'll find yourself trying to compensate for the assumptions you made which are no longer true, or you'll find yourself undoing almost all of the work you did since the assumptions you made were the basis for the top-to-bottom design.
More games are being designed with the PS3 in mind or as PS3-first developments. Because it is the PS3 that has the constraints, it makes sense to develop PS3-first and then port to the 360 since your assumptions about the amount of power with the PS3 won't be a problem for the 360.
