I don't really know that the logistics would work out. Back when I had those non-volatile RAM disks those were really only for the raving loonies, or for people who had serious work to do and needed more-or-less instant-on systems. But the operating systems were so different. These days they're optimized to make a different sort of use of both volatile memory and hard drive. The operating systems we use now were developed, apparently, with the idea that mass storage was always going to mostly be done on hard drives (or something with similar price per unit storage advantages). Even though RAM is very cheap, it still doesn't compete on a dollar for dollar basis with the hard drive. The way the operating systems are designed now, I'm not sure how much of a performance advantage a user would realize with a large RAM disk. So many factors figure into the determination of real world performance. It's not so easy to radically change the behavior of an OS to fit a new performance tuning paradigm as it was a few years ago. Too many variables to take into account in those millions of lines of code. In a way I guess you could say that it's another case of the lowest common denominator winning in the marketplace, instead of the out-and-out best performance solution. It's a pretty common outcome these days in just about any technology. Giving up on the idea of ultlimate performance in favor of wider dissemination and a broader market appeal (fewer denaro for adequate performance rather that big bucks for big bang).
Regards,
Jim