How much the virtualization could be considered a safe alternative for develop applications on "problematic" systems?
For problematic systems i means the eventual impossibilities to use new hardware directly on linux enterprise distro and different OS systems (as BSD).
Someone have already make some experiments or, even better, try this approach for long period of time?
My hypothetical scenario could be a system with a new Intel Cpu with a desktop linux distro at 64 bit with 8 gb and two virtualitation systems (one based on equivalent RedHat distro and one based on a BSD system). Using only one virtualization each time.
Intel push a lot the UEFI on their motherboard but i read on the phoronix reviews of differents brands that recent desktop distro should not have problem while, theoretically, enterprise distro should work the same with some backport, but i'm not sure at 100 % if some issue remain the same, while about BSD system, even with the actual intel sandy bridge, the situation remain bad, so i wonder if the virtualization could be a viable solution for "resolve" these issue (but perhaps should be better said put a sort of "patch" at that problem).
But remain to me many concern about the robustness of similar solution.
For example if happen a power loss, the data present on the virtualizated system, remain safe or not?