Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!brl-adm!brl-smoke!smoke!FISHER%RPICICGE.BITNET@wiscvm.ARPA
From: FISHER%RPICICGE.BITNET@wiscvm.ARPA
Newsgroups: net.micro.cpm
Subject: X-Acknowledge:
Message-ID: <3233@brl-smoke.ARPA>
Date: Fri, 22-Aug-86 07:18:19 EDT
Article-I.D.: brl-smok.3233
Posted: Fri Aug 22 07:18:19 1986
Date-Received: Sat, 23-Aug-86 05:02:04 EDT
Sender: news@brl-smoke.ARPA
Lines: 27

Subject: BITNET mail follows

Date: 20 August 1986, 15:20:07 EAS
From: FISHER at RPICICGE
To:   INFO-CPM at AMSAA.ARPA
Re:   Uuencode/uudecode character transparency over the net

I did some experimenting with the character translation that takes place
when one crosses a BITNET/ARPANET gateway.  Many characters were filtered
during the EBCDIC <==> ASCII conversions, particularly by the UCBJADE
gateway.  However, it appears that ALL of the characters used by the
uuencode/uudecode programs have unique representations on both sides
of the WISCVM gateway.  Also, the translation performed at WISCVM
seems to agree exactly with the translation VM on IBM hosts uses when
communicating with ASCII terminals.  (I do not know if the same
statement can be made for MVS hosts and the non-IBM BITNET machines.)

At any rate, at least for VM BITNET sites, ARPANET to BITNET via the
normal WISCVM path, then to CP/M via something suitable like KERMIT
should yield the desired results.

(**Note**:  The EBCDIC values used for brackets, braces, backslash,
etc, are not same values needed for a TN-train printer.  So, though
clear-text files may look funny when displayed on an IBM printer or
327X terminal, they should still arrive in CP/M-land correctly :-)

J.S.Fisher     FISHER@RPICICGE.BITNET