From: utzoo!decvax!cca!cosell@Bbn-Unix@sri-unix
Newsgroups: net.unix-wizards
Title: '\n'
Article-I.D.: sri-unix.3151
Posted: Tue Sep  7 19:30:20 1982
Received: Wed Sep  8 03:16:07 1982

From: Bernie Cosell 
Date:  5 Sep 1982  3:23:32 EDT (Sunday)
In a recent revision of the ASCII standard (where `recent' probably
means between five and ten years ago), `Newline' is a defined
alternate meaning for'\12'.  I haven't seen the real spec in
a long time now, but as I recall they had the good sense to take
all of the `standard' usages of various control chars and make
them alternate meanings for the chars.  (for example, I believe that
the flow control functions of ^s and ^Q are now explicit).

In some sense, this means that the VT100 folk got it wrong.
The key on the vt100 that was clearly intended
to carry the `newline' function sends 015 instead of 012.  As I understand
the spec you have two choices for `newline': 015-012 or just 012.
DEC (and a lot of others, of course) mostly `made up' the convention
of using 015 for newline.  It hardly has a persuasive argument in
favor of it beyond the largeness of the key on many terminals.

   [historical note: I believe that this `convention' came into being
    in the days of halfduplex model 33's: if you had the user use
    CR for the newline function, the system had merely to echo a LF
    and everything was fine.  If you had the user use LF for newline,
    then the system had to echo CR RUBOUT, since it took a char of
    padding to do a CR, normally nmasked by the LF character (didn't you
    ever wonder why the sequence is CR-LF and not vice-versa?).  I've
    fooled with such a scheme and I can report that the extra char
    is really a pain it can really upset you rhythm, dsuch as it can be on a
    110 baud, noisy, mechanical kludge.]

    /Bernie