Please go back and read the messages on the list for the past 10 months
before raising this issue again. (People who have been reading the list can
stop reading this message right here.)
To recap, the problem with the RFC934 scheme in a nutshell is that it
increases line lengths. This in turn may interact destructively with mailers
that perform line wrapping, causing the loss of boundary markers or the
addition of false boundary markers. Moreover, the general act of line
wrapping can also general spurious RFC934 boundary markers. These problems
are all operational facts rather than designer theories. I fought against
the present marker scheme but I could not deny operational realities.
The present scheme (along with its conditions for markers not appearing
within message bodies) avoids all these problems. In addition it is
quite a bit easier to implement (having done both RFC934 and RFCXXXX
implementations I have found this to be true).
Ned