Date: Fri 26 Apr 91 21:39:28-EDT
From: John C Klensin <KLENSIN(_at_)infoods(_dot_)mit(_dot_)edu>
Subject: Re: UUencode
To: ljm(_at_)ftp(_dot_)com
Cc: ietf-822(_at_)dimacs(_dot_)rutgers(_dot_)edu
[...]
I'm coming in a little out of context and hope I'm not going over old
ground but, given Leo's comment...
This is just a question, but do IBM mainframes of the VM/CMS and MVS
persuasions go on that "almost everything" list? I vaguely recall
being told that several versions and styles of uuencode won't survive
translation to ebcidic.
A related problem is that uuencoded ASCII mail that gets converted to EBCDIC
and back somewhere in its transit path may not survive intact. There are two
reasons for this: (1) trailing spaces may get truncated when converting
fixed-length space-padded records back to CR-LF- or newline- terminated
records, and (2) some ASCII characters that are used in uuencoded documents
are not represented in some versions of the EBCDIC character set, so there is
no obvious translation from EBCDIC to ASCII or vice versa. (In the EBCDIC
table I have handy, these characters include backquote, caret, left bracket
and right bracket.)
The character set used by BASE64, on the other hand, seems quite likely to
survive translation to and from any character set that includes a latin
alphabet.
Keith