Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!um-math!hyc
From: hyc@math.lsa.umich.edu (Howard Chu)
Newsgroups: comp.sys.atari.st
Subject: Re: uudecode problems / sending across BITNET
Message-ID: <608@stag.math.lsa.umich.edu>
Date: 30 Mar 89 05:21:13 GMT
References: <22.MAR.1989.14:48:15.FX@UK.AC.DARESBURY.NNGA>
Sender: usenet@math.lsa.umich.edu
Reply-To: hyc@math.lsa.umich.edu (Howard Chu)
Organization: University of Michigan Math Dept., Ann Arbor
Lines: 37
UUCP-Path: {mailrus,umix}!um-math!hyc

In article <22.MAR.1989.14:48:15.FX@UK.AC.DARESBURY.NNGA> FX@nnga.daresbury.ac.UK (Dave Love) writes:
>The problems with uudecode look as though they may have originated
>through transfer of files across BITNET.  It is a well-known problem
>(although I don't know if it's universal) that the characters ^~{} at
>least get mangled going through a bitnet route, even if the sender and
>receiver are ascii machines.  Thus sent ^ can become recieved
> ~, sent ~ recieved % ...  Of course, if this message has passed through
>bitnet to you that won't make much sense!  The characters that I know
>cause trouble are uparrow/caret, tilde and curly brackets.  You cannot
>undo the transformation automatically since you don't know if % was sent
>as such or sent as tilde.  Moral -- exchange such files via internet or
>uucp.  If anyone out there has the clout to persuade the bitnet
>authorities to sort this out, please do so.  It plays havoc with
>resaerchers trying to exchange TeX and PostScript if nothing else.

It's a nice thought, but will probably never happen. BITNET is wedded to
RSCS and EBCDIC, which is probably the major source of the problems right
there - until very recently, there was no one-to-one mapping for ASCII/EBCDIC
translation. Aside from that, there's no such thing as "standard EBCDIC" - IBM
changes the definition with almost every new hardware release. As such,
persuading the bitnet authorities is probably a hopeless case. We recently
had to change the EBCDIC on our mainframe to comply with the recent ISO
definition, (aka IBM CP37, I believe.), which is supposed to now be The
Standard EBCDIC. That was, to say the least, a royal pain. Curly brackets,
square brackets, tilde, Icelandic Thorn (what the heck is an Icelandic
Thorn? I dunno, but your lower case and upper case versions are swapped.
Oh. Gee, thanks for letting us know.) and a bunch of other characters got
relocated, all in the name of standardization. We are now unable to communicate
these characters to other Bitnet hosts because none of them have undergone
the painful transition yet...

Bitnet is evil. EBCDIC is evil. IBM is evil.
Anything built entirely upon and still relying upon punch card technology
is utterly evil.
--
 -=- PrayerMail: Send 100Mbits to holyghost@father.son[127.0.0.1]
 and You Too can have a Personal Electronic Relationship with God!