This is spot on. The Emacs example is excellent. Now, of course you realize that, because refining, validating, fixing, polishing and maintaining compatibility is expensive and considered very boring by developers, all this requires a major corporation and major funds which means the universal Linux Desktop will never happen. In the very best case it will be something like the "Ubuntu Desktop" but not just "the Linux desktop".