But what is "the one model of ownership for mail"?
It seems to me that there is no defined email ownership model on the
Internet; rather, there is a de facto one due to the limits of current
standards and implementations. Very much the same way that, before MIME,
there was no defined way to transfer non-textual data.
The fact that Internet email exists means that there's an ownership model. It
may not be written down anywhere, but it exists nevertheless. And parts of it
*are* written down and *are* well defined, and there is work already underway
to clean all this up.
Ownership of a message prior to sending it and after receiving it is pretty
clear already, I believe. While there are communities that seem unclear on
ownership after receipt, I see these communities as using a service other than
email in the conventional sense.
As for message in transit, the one thing that is clear is that there are no
clear ownership rights granted to the originator or recipient. That is, it is
wrong to assume that the sender can exert any form of ownership after sending
a message and it is wrong to assume that the recipient can exert any form of
ownership prior to reception.
However, it is also wrong to assume that the sender cannot exert any control
at all. Systems do exist that allow for recall of messages from a local
I've yet to see a system that lets a recipient exercise any form of control
over a message prior to delivery (this is almost a tautology since one
defintion of delivery is when the messsage is placed under the control of the
recipient), but anything is possible.
Given that there's considerable variation in the area of in transit messages,
the only thing to do is sit down and try to come up with way of describing what
is used in practice. And this is exactly what the IFIP EmailMgt people are
doing. They are attacking the larger problem of general email management, which
subsumes issues of ownership into a much broader context. If you want to get
involved in this work you are of course welcome to do so.
All this taken together constitutes a close approximation of a model. It may
not be clean and it may have considerable variability, but it is a model
It seems to me that one of the things that needs to be done is clarify the
ownership models currently in use, and decide which one(s) (if any) are
reasonable and useful and supportable.
I agree with the first and disagree with the second. Clarification and
documentation is fine and dandy, and that's already well underway. But getting
into decisions about what is reasonable and what is not is more than a little
problematic, in that it almost certainly involves declaring previously
legitimate approachs as now being broken. Experience has shown that attempts
to do this are not well received.