🚀 go-pugleaf

In article <8904242330.AA20426@fenris.idt.unit.no> bjornmu@idt.unit.no writes:
>BTW, has anyone tried to compile DBW_Render (I'm talking about the
>modified version by Ofer Licht) with Lattice 5.02 and run it on a
>68020?  

Several people have asked me for the executables to dbw with
many interested in in-line FP.  If you have compiled dbw with
lattice or manx with 68881 code, please put a zoo of the
executable by anonymous ftp to bach.berkeley.edu in the 
directory /pub/incoming.

I would like to reiterate in public some things I have mentioned to
a couple of people about dbw.  My modifications to dbw fall into
2 categories.  
	1. changes in the internal representation and outputting
	   of scanlines.  The original code used ints but bit twiddled
	   them as if they were 16bits long.  This was for the normal
	   16bit manx libraries.  To make dbw behave "normally" --
	   while still perserving output that would be readable by
           the ray2 postproccessor, I changed many integer structures
	   to shorts.  This is the minimal needed to make dbw work
	   on un*x boxes.
	2. adding the capability for dbw to read scanline ranges
	   on the commandline.  This was done to facilitate
	   distributed processing with the "distpro" module.
As you can see, my mods were minimal.  I think most people are happy
about being able to run dbw on their office workstations.  Some people,
however, would like to see dbw recompiled to run on an '020 amiga with
an '881.  To those people I say "go fiddle with your compilers and
hack the code some more."  (if you do manage this, please send me a copy for
distribution purposes--see above).  To people who are "serious" about ray 
tracing I say "find yourself a LAN of MFLOP workstations and go distributing."
To the people who only have their original amigas with no add ons I say
"I'm sorry.  you will probably be wasting your time trying to do good
ray tracing on your machines.  It just takes too long!"
(to the eyes I say "why couldn't the distribution of photoreceptors in the
fovia be simpler to simulate?")

-Ofer Licht
ofer@gandalf.berkeley.edu