On Sat, 16 Mar 2002 22:20:56 EST, Keith Moore said:
it would be quite reasonable for applications that already had a
deeply-wired assumption that character strings were in EBCDIC.
Which is why on BITNET and related networks, the applications
generally exchanged data in EBCDIC even between ASCII-native hosts.
Damn you Keith. I had *almost* managed to repress the joyful memories of
trying to debug what happened when a document managed to start off life
as ASCII, get translated to EBCDIC for Bitnet, to/from ASCII again for
an intermediate hop, back to EBCDIC again for another hop, and then back
to ASCII at the tail end of the journey. And yes, that translate count
is correct, and that document was... umm... interesting. ;)
However, the day-to-day joys of Bitnet ASCII-EBCDIC translation and what
it did to uuencode was one of the *primary* motivating reasons why Base64
happened instead of uuencode. Let's not lose sight of that debacle. ;)
Computer Systems Senior Engineer
Description: PGP signature