But while the underpinnings of the system might have been Linux, the Oasys didn't allow you to piggyback regular Linux apps or instruments, did it?

I think what we are seeing here is a sort of hybrid between hardware synthesis generation, and perhaps a Linux-like OS to run it. Rather than the OS do everything, including the synthesis (like the MS does), the hardware does most of the work (guaranteeing polyphony count and drop-out free operation) while the OS takes care of operation and data handling.

But, however it is achieved, it is the new benchmark. For working pros in studio and stage alike, bomb-proof operation is the FIRST thing needed. Unfortunately, so far, the 'pure' software instruments have struggled with this, as any combination of VSTi's can pull even a modern multi-core CPU to its knees. Using a more modern OS to handle data streaming, while hardware still takes care of much of the sound generation and manipulation seems, until maybe we have a few more CPU generations with a LOT more power, to be the more reliable route.