Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!spool.mu.edu!uunet!europa.asd.contel.com!wlbr!WLV.IMSD.CONTEL.COM!mcc
From: mcc@WLV.IMSD.CONTEL.COM (Merton Campbell Crockett)
Newsgroups: comp.binaries.ibm.pc.d
Subject: Re: tenex vs. binary (was: Re: GARBO.UWASA.FI)
Keywords: ftp
Message-ID: <1991Mar20.124134.11589@wlbr.imsd.contel.com>
Date: 20 Mar 91 12:41:34 GMT
Article-I.D.: wlbr.1991Mar20.124134.11589
References: <1991Mar19.220315.20176@cbnewsh.att.com> <5461@vela.acs.oakland.edu> <5464@vela.acs.oakland.edu>
Sender: news@wlbr.imsd.contel.com (news)
Distribution: unlimited
Organization: Contel Federal Systems
Lines: 58
Nntp-Posting-Host: wlv.imsd.contel.com


 In article <5464@vela.acs.oakland.edu> w8sdz@rigel.acs.oakland.edu (Keith Petersen) writes:
 >In my previous posting I wrote:
 >> Remember references to big indian versus little indian byte-order?

 >In respose I received via email:

 >ressler@cs.cornell.edu (Gene Ressler) writes:
 >> I've always heard it referred to as big-Endian vs. little-Endian.
 >> Machines where the most significant bits (big end) of a word occur in the
 >> first byte are the former and conversely...  Just another piece
 >> of mindless trivia.

 >raymond@math.berkeley.edu (Raymond Chen) writes:
 >> Unless this is a reference to something I'm not familiar with, perhaps
 >> you meant `big-endian' and `little-endian'.  The reference is to the
 >> paper by Jonathan Swift about the two countries who fought a war over
 >> whether one should places one's soft-boiled egg in the egg cup with
 >> the little end or the big end pointed upwards.

 >That's an interesting story about the soft-boiled egg.  :-)

 >I've seen some long and involved discussions on the net about big end
 >versus little end.

Ahh!  It comes back to me.  On DECSystem 10 and DECSystem 20, bit 0 references
the most significant bit or sign bit of the 36 bit word.

The following is from is a response to an earlier reply which was bounced
because of an unknown host.

References: <12011@uhccux.uhcc.Hawaii.Edu> <1991Mar19.110958.17619@uwasa.fi>

 In article <3067@bimacs.BITNET> you write:
 >>   No the files are _not_ corrupted.  Most of these complaints are a
 >>result of the user's own failure to use the binary mode (tenex in
 >>the case of Simtel20) in the transfer from the ftp site to his/her
 >>own site, and/or his/her failure to transfer the files in the binary
 >>mode from his/her mailframe to the PC.

 >Excuse me, what is the difference between binary mode and tenex?
 >Are they compatible?
 >What ftp sites use them?
 >(etc...)

Binary mode and tenex are NOT compatible.  DECsystem 10s and DECsystem 20s
require the use of TENEX (DECsystem 10 Exchange).  These systems are based
on 36 bit words.  A "byte" is a variable length entity on these systems and
can be 1 to 36 bits in length.  "Bytes" are allocated starting from the least
significant bit of a word.  Multiple "bytes" can be stored in a single word.
"Bytes" cannot span words.  Any high-order bits that are not allocated to
a "byte" are ignored.

TENEX specifies an 8 bit "byte"--4 "bytes" to a word.  In binary mode, you
get all 36 bits of the word transmitted to you as a bit stream.  The high
order 4 bits which have an undefined value will be transmitted in the stream.

Merton Campbell Crockett