>Mirroring is mirroring and you would hope that what is
>written to mirror1 is the exact same as mirror2, since I wouldn't expect
>the server to think/edit the data that it writes out secondly.

But you're missing the point of what Eric and others have been saying...
The closer mirroring is to the application creating the data, the better
opportunity there is for error correction, via the server options.
Hardware and operating system mirroring have no idea what's going on,
data-wise, and cannot deal with logical errors.
It is far better to avoid database problems than depend upon recovery
to fix them.          Richard Sims, Boston University