Re: Update of the MIME-MHS Specs

[Keith Moore <[e-mail]>]

> I suggest that the splitting and recombining messages is the wrong solution
> to exchanging large messages. No matter how small you dice the message there 
> will always be a node or recipient who can't accept pieces and put them 
> together.

Nobody thinks that message/partial is an ideal solution.  It's designed for
exceptional cases where the sender knows that the recipient *can* accept
the pieces and put them together, and that the mail system *can* likely
transmit the pieces and get them together intact (perhaps with some delay
between sending each piece.)

(maybe the next MIME RFC should make this very clear.)

> I suggest that a better usage of store and forward messaging for large
> information objects is to deliver knowledge about the information object to
> the recipients and provide another communications channel for recipients to 
> retrieve the information object.
> 
> Why can't we automate this manual process? Let's get the larger than 1MB
> messages out of the store and forward path and make better use of our 
> resources.

Uhhh...unless I completely misunderstand what you are talking about,
MIME already has this capability.  It's called message/external-body.
Yes, it does get used a lot more often than message/partial.

Keith Moore