just a point of order here in favor of uuencode. people
have been using it on puny machines over crappy networks
for years now, and there's some versions of it (notably
for atari st's, if I recall right) that have elaborate
checks for corruption in it.
pushing base64 as the one true solution when it doesn't
even have error detection seems a bit presumptuous.
uuencode works just fine in its common environment
(netnews), there's many megabytes a day floating around
encoded that way, don't knock it so quick.
I thought that the IETF had a mandate to at least describe
existing practice before leaping off into new solutions.
--Ed