I don't get it. It seems straighforward to extract a single message
out of an "RFC 822 email archive" and serve it up.
Or heck, give me a pointer to the whole archive. That's fine by me.
That at least makes it _possible_ for me to get my job done.
>Many other documents may also be derived (e.g. from data bases, server
>side includes) and have no "original" to refer to.
OK. If the source data in such a situation is completely unobservable
from the net, then I'd consider whatever byproduct is made available to
be the original. No audit trail necessary there.
Dan