From: utzoo!decvax!ucbvax!ARPAVAX:C70:editor-people
Newsgroups: fa.editor-p
Title: Re: buffer structures
Article-I.D.: ucb.1972
Posted: Mon Sep 20 02:48:55 1982
Received: Tue Sep 21 05:53:23 1982

>From z@CCA-UNIX Mon Sep 20 02:46:18 1982
CCA EMACS uses an array of line pointers rather than a linked list of
lines; in such a scheme I don't see any problem in implementing
buffer bounds or in efficiently determining which of two buffer
positions is "first".

Actually, the only reason I haven't implemented buffer bounds in CCA
EMACS is that they're somewhat low in priority; they are definitely
on my list to be done, though, and their implementation is
straightforward.  If they had been designed-in from the start, they
would have been quite trivial to put in.  However, I started out with
Montgomery's EMACS and have had to retrofit a lot of stuff.  The same
goes for nondrifting marks; this is also a feature which has recently
been installed.

To determine the relative positions of two points may take two
comparisons in a line-oriented editor compared to one in a buffer gap
editor, but in terms of actual peformance this difference is
insignificant.  In actual execution, only one comparison is usually
needed anyway.  All in all, I am quite happy with the line oriented
approach, and I think it is a major reason why command execution
speed in CCA EMACS is generally quite fast.

	Steve Zimmerman