Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!usc!bloom-beacon!eru!luth!sunic!mcsun!ukc!strath-cs!dcl-cs!gdt!gdr!exspes
From: exspes@gdr.bath.ac.uk (P E Smee)
Newsgroups: comp.sys.atari.st
Subject: Re: Decoded Net files
Message-ID: <1989Sep17.143908.2056@gdt.bath.ac.uk>
Date: 17 Sep 89 14:39:08 GMT
References: <8909080255.AA29852@ucbvax.Berkeley.EDU> <22004@cup.portal.com>
Reply-To: exspes@gdr.bath.ac.uk (P E Smee)
Organization: University of Bristol c/o University of Bath
Lines: 36

In article <22004@cup.portal.com> Bob_BobR_Retelle@cup.portal.com writes:
>It shouldn't matter what kind of computer does the transferring (VAX, IBM or
>whatever) since the whole point of UUENCODING is to convert a binary file int
>a standard ASCII text file that can be transferred by ANY machine, over ANY
>network...

Ahh, but the problem is with transfers which pass thru an IBM machine, since
IBM machines DO NOT SPEAK ASCII.  They speak EBCDIC.  There is not a unique
one-to-one and reversible mapping between ASCII and EBCDIC.  (And, to
further confuse things, there are a number of variants of EBCDIC, frequently
all used in different contexts on the same machine.)

The most frequent problem with transfers which have been passed through
IBM's is for characters which began life as ASCII circumflex (up-arrow,
little hat, whatever you call it -- ASCII 0x94) to come back out of the
IBM as tildes (squiggles -- ASCII 0x126).  This is because neither of
these graphics exists in EBCDIC.  EBCDIC has a (different) single    
graphic representing the (IBM) 'NOT' character -- and both hat and
squiggle tend to get mapped to that on input from an ASCII machine.

The simplest uudecode algorithm (which assumes that the file contains only
chars in the uudecode character set) will give an incorrect 6-bit value
when it decodes the resulting squiggle, as compared with what was in the
original file.  (There is an alternative algorithm which will do the
right thing, as will Dumas 'table' method.)

It is generally (but not invariable) true that if you push an ASCII text
file from an ASCII machine through an IBM EBCDIC machine and then back
to an ASCII machine, you will not end up with an identical file.  The
two character coding schemes do not match.

-- 
 Paul Smee               |    JANET: Smee@uk.ac.bristol
 Computer Centre         |   BITNET: Smee%uk.ac.bristol@ukacrl.bitnet
 University of Bristol   | Internet: Smee%uk.ac.bristol@nsfnet-relay.ac.uk
 (Phone: +44 272 303132) |     UUCP: ...!mcvax!ukc!gdr.bath.ac.uk!exspes