Harald T. Alvestrand has asked me to write something about this.
After thinking much about this, and considering both my own earlier
proposal and Ed Levinsons counterproposal, I have concluded that
a third alternative would be more suitable than our earlier
alternatives. The third alternative is shown by the following
example:
.... main message heading ...
Content-Type: Multipart/related
... Content-part 1 ...
Content-Type: Multipart/alternative
... Content-part 1.1
Content-Type: Text/plain
... plain text version for those who cannot read HTML ...
... Content-part 1.2
Content-Type: Text/HTML
... HTML version of the same message, which many contain ...
... external file references, as shown in the example ...
... which follows on the next line ...
<IMG SRC=HTTP://www.dsv.su.se/picture.GIF>
... Content-part 2 ...
Content-Type: Image/GIF
Content-Location: HTTP://www.dsv.su.se/picture.GIF
... the GIF picture
Here is a comparison of the new notation suggested above for some
common combinations of user software:
User software New notation using Notation using CIM URL-s
Content-Location
Mail system which has no The user can manually The user will have to
knowledge of the save the Text/HTML modify the HTML text
Text/HTML Content-Type content in a file, and and replace the CIM URL-
open and view it with an s with file references.
ordinary Web browser.
Mail system which uses The user can view the Will not work unless the
an ordinary Web browser message, including mail system goes into
as a helper application pictures, with no the text and changes the
to view Text/HTML problems. (The pictures CIM URL-s to something
Contents will actually be else.
downloaded from a remote
WWW server, but that is
normal WWW client
behaviour.)
Mail system which uses a The user can still use Either the mail system
proxy HTML server an ordinary Web browser, will have to modify the
the proxy server will CIM URL-s, or the Web
recognize that it has browser must recognize
the picture locally and CIM URL-s and know how
not retrieve it from its to retrieve them from
remote location. the mail system.
Mail system with built- No problem No problem
in Web browser.
Thus, in the first three of the four cases discussed above, my new
proposed format is easier to implement and use than the earlier proposals
using CID-s which I and Ed Levinson wrote. I am not against CID-s and
MID-s as a long-term proposal, but in the short term, my new proposal
makes things simpler in three of the four cases.
(Note: A similar idea was put forward by Ron Croonenberg in a message
two weeks ago.)
----------------------------------------------------------------------
Jacob Palme <jpalme(_at_)dsv(_dot_)su(_dot_)se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/~jpalme