Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!rutgers!njin!princeton!udel!mmdf
From: TROTH%TAMCBA.bitnet@jade.berkeley.edu (Rick Troth)
Newsgroups: comp.os.minix
Subject: RE: EARN - BITNET
Message-ID: <3964@louie.udel.EDU>
Date: 5 Sep 88 01:18:50 GMT
Sender: mmdf@udel.EDU
Lines: 20


        The problem in Europe (EARN) is largely because IBM has recommended
two different character sets,  both called EBCDIC,  in the US (BITNET) and
most of Europe.  There is work being done to correct the problem.  But I am
surprized to hear that you are having trouble from the fixed-vs-variable
record conventions:  try uuencode/uudecode.  (white space should be white
space,  don't count on a number of 0x20s or 0x09s)

        Historically,  EBCDIC did not have things like square brackets and
the hat (circumflex) accent,  and ASCII did not have things like a cent sign
or logical-not symbol.  Now there are extensions to both EBCDIC and ASCII that
include these and other characters supporting about forty Latin oriented
languages.  But the smooth transition to these new character sets is difficult
becuase of  "National Replacement"  sets,  which have existed in both EBCDIC
and ASCII for years.  Pound (number),  cent,  exclaim,  brackets and others
all get messed-up among the different countries.

"Those who do not understand UNIX                                  - Rick Troth
 are condemned to re-invent it                             TAMCBA VM Operations
 poorly"        (stolen quote)                    Texas A&M College of Business