Received: from PACIFIC-CARRIER-ANNEX.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA06342; Sun, 10 Dec 95 14:54:41 EST
Received: from colossus.cse.psu.edu by MIT.EDU with SMTP
	id AA08245; Sun, 10 Dec 95 14:54:41 EST
Received: by colossus.cse.psu.edu id <78425>; Sun, 10 Dec 1995 14:46:11 -0500
Received: from plan9.att.com ([192.20.225.253]) by colossus.cse.psu.edu with SMTP id <78423>; Sun, 10 Dec 1995 14:45:55 -0500
From: td@plan9.att.com
To: 9fans@cse.psu.edu
Date: 	Sun, 10 Dec 1995 14:23:48 -0500
Subject: re: Adding pcdist to CD-ROM file system?
Message-Id: <95Dec10.144555est.78423@colossus.cse.psu.edu>
Sender: owner-9fans@cse.psu.edu
Precedence: bulk
Reply-To: 9fans@cse.psu.edu

Luther Huffman, Jr. avers (aside):
>p.s. To whomever designed catclock, I enjoy it. =;-)

It's an X program that I converted to use libg.

Despite the inclusion a polygon scan-converter
(libg doesn't have one), the conversion reduced the
program from several thousand lines to 331 (plus bitmap
data).  The only feature it lost was the alarm.

I should say that the polygon tiler is of extremely high
quality.  It uses only integer arithmetic, handles either
of the common rules for identifying the interior of
non-simply connected polygons, clips in an extremely simple
manner, and guarantees, among other things, that if you cover
a region with a bunch of abutting polygons, it will touch each
pixel exactly once.  There is a comment in the source code
indicating the couple of things I think would have to be done
(none of them difficult) to turn it into a production-quality
library routine.  (Actually, I think it should be in the library
as-is.)

