My question is, from a programming standpoint, will using the ESRAM be as easy as they say? I've seen one developer say it's kind of a pain, which seems likely when compared to unified memory, but is it significant? Will it possibly result in title versions that differ between the two systems? Microsoft says it's easy and that developers had no problems with the 360 setup, which makes sense, but that was in comparison with the convoluted PS3 architecture.
The impression I get is that ESRAM gives quite a boost--if it's used perfectly--but I suspect that in real-world scenarios, Sony's setup is more ideal.
I guess my real question is, will this mean variable results on the XO based on developer time/skill? Will this be something that studios will need time to master or will experience with the 360 translate to developer familiarity with the process? In short, will there again be "lead system" differences because of developers having to go back and deal with split/unified memory differences? I was hoping for completely concurrent development, which I'm sure is largely still going to happen, but the less optimization that needs to be done, the better, in terms of development cost, time, and quality.
