ietf-822
[Top] [All Lists]

re: internal external body

1993-03-22 17:22:26
--- Begin Message ---

- Is there any reason that the body of a message/external-body entity
  has to be outside of the top-level message? If not, then what is the
  right way to handle it ?


Here is an additional issue relating to this question.

My understanding of the spec is that content-id is a required header for the 
message/external-body. This is provided for "caching" purposes. Is the 
content-id header part of the message/external-body headers or the encapsulated 
message headers?  The spec seems to indicate that content-id is part of the 
message/external-body headers.  

This seems kind of strange to me since it causes caching to be associated with a
specific reference to the external body rather than to the external body itself
(which might have multiple references within a message and across different
messages).

This was discussed by the group some  months ago with several people touching 
on 
the use of content-ids for intra message references.  The issue of caching a 
reference vs caching the item was also brought up but did not seem to reach 
closure.

It seems to me that associating the content-id with the encapsulated headers
(i.e.  with the external body) allows more efficient caching.  Independent of
that issue, use of an external-body reference to a part that resides in the same
message requires the content-id to be part of the encapsulated headers.  If you
then have a "required" content-id header on the wrapper headers it seems kind of
klunky.

What am I missing?

Gabe
--
Gabe Beged-Dov                   e-mail: gabe(_at_)cv(_dot_)hp(_dot_)com
Hewlett-Packard
Open Systems Software Division, 
Corvallis, Oregon                phone: (503)750-4415






--- End Message ---
<Prev in Thread] Current Thread [Next in Thread>