I don't understand the complexity that comes from putting the
text/plain alternative inside the multipart/related instead of
outside. I also don't get why the convolution with external-body
references. Why isn't this example just:
================================================================
1 Content-type: Multipart/Alternative
1.1 Content-Type: Text/plain
Content-Description: plain text version for those who
cannot read HTML
1.2 Content-type: Multipart/Related; type=text/html
1.2.1 Content-Type: Text/HTML
Content-Description: HTML version of the same message,
which may contain external file references
<IMG SRC="cid:1.3@accurate.com">
1.2.2 Content-Type: Image/GIF
Content-Description: the GIF picture
Content-ID: <1(_dot_)3(_at_)accurate(_dot_)com>
================================================================
or, if you don't want to rewrite the URLs inside HTML, creating
a new type for transforming URLs:
================================================================
1 Content-type: Multipart/Alternative
1.1 Content-Type: Text/plain
Content-Description: plain text version for those who
cannot read HTML
1.2 Content-type: Multipart/Related; type=application/translate
1.2.1 Content-Type: application/translate; type=text/html
Content-Description: new media type for translating things
http://www.dsv.su.se/picture.GIF cid:1.3@accurate.com
1.2.2 Content-Type: Text/HTML
Content-Description: HTML version of the same message,
which may contain external file references
<IMG SRC="http://www.dsv.su.se/picture.GIF">
1.2.3 Content-Type: Image/GIF
Content-Description: the GIF picture
Content-ID: <1(_dot_)3(_at_)accurate(_dot_)com>
================================================================