ietf-822
[Top] [All Lists]

Re: UUencode

1991-04-28 18:03:24
No, uuencode is the right solution for this simple reason that it is
very commonly available on almost everything, DOS, OS/2, Macs, various
UNIXs, even VMS/VAXen.  Currently, I can send a uuencoded message to anyone
and be confident of their ability to decode it -- choosing uuencode as
the basis for an automated encoding facility ensures that even those
with 'old-style' MTA/UAs will be able to understand the message.

Er, no, you cannot be so confident. UUENCODE made some poor choices in terms of
the characters it uses, and the result is that a lot of uuencode messages get
wrecked during e-mail transit...

...As for the presence of lots of implementations being a compelling reason
to use a poor encoding, an implementation of the BASE64 is only a few dozen
lines of code (my code that handles BASE64, HEX, BASE85, and UUENCODE is
something like 500 lines). A complete utility is only a few hundred lines.
We're not talking about a huge programming investment here....


Sorry, I seem to have not been clear.  UUENCODE (or to a lesser
extent binhex) is the superior default choice because of the huge
installed base of users to whom it is available and who are familiar
with its use.  This is how I send binary information today.  It works,
albeit painfully from a user interface perspective, and I've never had
anyone ask how to decode my message.

Providing a nifty mail system that automates the task of sending an
executable to a coworker, but that requires the recipent to buy new mail
software as well (or manage to get and compile a copy of the BASE64
sources -- Oh, you didn't get a compiler with your system?), would be much
less desirable than providing a system which could be installed and used
one node at a time.

enjoy,
leo j mclaughlin iii
ljm(_at_)ftp(_dot_)com


<Prev in Thread] Current Thread [Next in Thread>