Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!usc!ucsd!nosc!crash!jcs
From: jcs@crash.cts.com (John Schultz)
Newsgroups: comp.graphics
Subject: Re: Fastest 2D line clipper
Message-ID: <2148@crash.cts.com>
Date: 8 Apr 90 21:59:56 GMT
References: <2130@crash.cts.com> <21266@eerie.acsu.Buffalo.EDU>
Reply-To: jcs@crash.cts.com (John Schultz)
Organization: Crash TimeSharing, El Cajon, CA
Lines: 27

In article <21266@eerie.acsu.Buffalo.EDU> prussak@spica.cs.buffalo.edu (Michal Prussak) writes:
>
>In my understanding, Cohen-Sutherland is not presently the state
>of the art algorithm. The best that I know is a Liong-Barski algorithm.
>I don't see any point comparing your algorithm to an outdated one.
>
  The CS algorithm may not be the state of the art, but the FC clipper
was tested against the Liang-Barsky algorithm, which is a parametric
method.  The LB algorithm was 2-12x slower. The Cyrus-Beck parametric
method was much slower than the LB method. There was also an algorithm
published in ACM Computer Graphics, Volume 21, Number 4, July 1987,
pp. 253-262, developed by Tina M. Nicholl, D.T. Lee and Robin A. Nicholl.
They state that their method is faster than CS, and LB. They mention
the SPY algorithm as well, but it is unclear which is faster, as they
don't show any real statistical comparisons.  Further, they do not give
a complete source code example, so I didn't bother to implement and 
test it. Nice structured paper, though.
  Also, the FC or SPY algorithm was created by Sobkow,Pospisil and Yang.
I just fixed a few errors in the C version then rewrote it in assembly.
I have not heard of the Liong-Barski algorithm- what type of algorithm
is it? I'd be interested in seeing it (source, references, etc).
Of course, state of the art is high-speed hardware clipping, which is
available only on state of the art workstations or graphics boards.



  John