Subject: Re: XON/XOFF
To: None <dribbling@thekeyboard.com>
From: Greg A. Woods <woods@weird.com>
List: port-i386
Date: 10/20/2000 22:13:47
[ On Friday, October 20, 100 at 17:45:10 (+0000), dribbling@thekeyboard.com wrote: ]
> Subject: XON/XOFF
>
> So it was developed for TOPS?  Still surprising that it
> shunned software handshaking though, since it was commonly
> supported by DEC hardware.

Things are, as always, a little more complex than they first appear on
the surface!

As I understand it Emacs was first implemetned on ITS, and grew out of
ITS TECO, and in fact was originally a collection of "Editor Macros" for
TECO put together by Richard Stallman.  If I remember correctly TOPS was
not well thought of around MIT, which is of course why they developed ITS.

Probably the best paper describing the history of all things "Emacs" is
by Bernard S. Greenberg, the author of Multics Emacs and the fellow who
undoubtably took Emacs away from TECO and gave it a new life as an
editor of its own with its own extension language (i.e. he can more or
less be credited with the idea of turning Emacs inside out so that it
was no long an editor written within another editor, but rather a
separate stand-alone editor that itself can be used to implement, or at
least emulate, other editors [Stallman could easily have been thinking
along the same lines, but I believe it was Greenberg who first chose
lisp as the new extension language and the first to actually implement a
standalone editor that mimiced ITS Emacs]).  You can read his paper
here: <URL:http://www.best.com/~thvv/mepap.html>.  For those of us who
used each of early GNU Emacs, early Gosling Emacs, and Multics Emacs
(and in my case all at about the same time), this paper is a fascinating
insight into the editor that was often at the centre of our computing
lives (and indeed may still be there -- note the X-Mailer header in this
message!)  In some ways I'm glad I never used ITS Emacs, though I'm sure
it would have taught me its own lessons too!  :-)

> That's a good point about dropped characters.  So I'm right
> in thinking that it's Emacs that doesn't cope well with
> software handshaking, rather than unix?

It's not that Emacs doesn't cope well with software handshaking -- it's
just that when software handshaking is used the user will loose two easy
command characters for Emacs and indeed the default XON/XOFF characters
are, as I explained earlier, very commonly used by users of the default
Emacs key bindings.

Unix copes as well, or better, with software flow control than any
operating system can.  It will not cope well with lost flow control
characters, though a smart user or agent will usually work around such
problems with ease.  Indeed the Unix "stty ixany" feature makes software
flow control much more friendly and usable than it is in many other
systems.

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods@acm.org>      <robohack!woods>
Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com>