|In reply to Comment 63 (anonymous):|
"As for the architectural disparities, before someone jumps in with the endian argument we've heard it many times before -- it can be done and will have to be done if the 'Anywhere' mandate is to be realized."
I think you are underestimating the endian issue. There will be issues not unlike when drivers are ported from the x86 to PPC under Linux. It's also not just about compatibility within an old m68k data, but also files stored in big endian format that are to be used by different pieces of software. I feel that only when the OS is truly written in VP will it make sense to run on x86 and other platforms. I started to get the impression that running an entire OS written in VP isn't the big prize anymore, although I see it as a big plus that's worth the slight hit in performance. Imagine the freedom that CPU designers would have knowing they can change the underlying architecture completely and only have the software people worry about the VP translation layer.
Forget the Amiga for a second and think of the power that this would add to Linux that was compiled to run under VP. Truly portable binaries, the Holy Grail of the IT world. This would be fantastic technology and allow for more competition in the CPU industry and OS industry. Since it would enable competition, that's exactly why I suspect it would be crushed. With the concept of VP, Windows just might run everywhere, but the same could be true for other OS's.