MIME vs. DIME vs. ...

I think the issues are: footprint of code, reliability,
consistency. MIME with content-length isn't MIME;
MIME multipart is difficult to create 'reliably'
(because of the difficulty of choosing separators that
are guaranteed to not appear in the body), MIME multipart
is difficult to parse reliability (long history of
coding errors with separator-based file formats.)

The "blocks architecture" doesn't provide for a single
MIME body. The architecture is a protocol architecture
rather than a data/message architecture.

http://www.ietf.org/draft-herriot-application-multiplexed-04.txt

is a pretty simple chunked encoding of multipart/related.

   contents = *chunk finalChunk
   chunk      = header payload CRLF
   header     = "CHK" SP messageNumber SP length SP isMore CRLF
   messageNumber   = 1..2147483647
   length   = 0..2147483647
   isMore       = "MORE" / "LAST"
   payload    = *OCTET
   finalChunk = finalHeader CRLF
   finalHeader  = "CHK" SP "0" SP "0" SP "LAST" CRLF

The write-up in Herriot's draft gives some use cases.