david@mks.com (David Rowley) (09/22/90)
A while back I submitted a proposal regarding the creation of a
public domain rendering library, with portability and modularity
in mind, perhaps based on existing standards such as renderman.
The responses I received lead me to believe that many are frustrated
by the lack of a common framework for work in this area, and that
such an approach could work out quite well.
Here are the responses I received:
From: Craig Kolb <craig@weedeater.math.yale.edu>
In fact, Pat Hanrahan, Mark VandeWettering and I have been discussing
exactly this.
The first step will probably be a Call For Submissions of ray/primitive
intersection routines, as we'll be working the raytracing portion
of the project first. At the very least, we'd like to collect
efficient ray/primitive intersection routines so that people won't
be forced to reinvent the wheel.
On top of this library, we'd like to build an extensible raytracing
skeleton.
After this is done, I'd like to develop a hybrid rendering package framework
and implement a shading language as well. (Something like Renderman,
but not Renderman -- it's got a number of problems...)
This is still all quite preliminary, and is the result of some
conversations Mark and I had over the net last year and a brief chat I had
with Pat at siggraph. We're supposed to get together to discuss the project
early next month.
If you have ideas, suggestions, comments, or want to volunteer, by
all means keep in touch... There's plenty of work to be done, and it
could be *very* nice if done well.
Cheers,
Craig
From: spacm1!sysmgt007@elroy.Jpl.Nasa.Gov
It sounds like a good idea to me. While I do not have much experience in
writing raytracers, I do have MUCH experience programming, have used
DECwindows, and have a general understanding of much of the disccusion in this
group.
In particular, I work on a VAX/VMS system and would be able to check out the
software (and make neccessary corrections) to make sure it would run under
VMS.
If you would like my help, let me know.
Dean (The Shadow)
From: mg5184@mars.njit.edu (Michael Gaines)
Fortunately it's on its way. I've been working on just that same system
for the last three months now. Less a language parser and some other tidbits,
it's designed to be a PD animation package for the Mac series. I can't tell
you the name or anything else because I'm waiting for copyright info and
other info from Mac to return. I hope to have the whole thing out by mid
September.
Mike
From: Jeff Butterworth <butterwo@kimchee.ncsc.org>
Your idea sounds wonderful. Since, as you pointed out, there is already a
fairly large body of PD rendering software in existence, I suspect that the
main problem will just be in planning this beast and getting a few of the
original authors to help out.
Although I won't promise that I can help out very much with the programming,
(I'm starting grad school at UNC-CH next week.) I would like to learn more
about your project. I could even do simple things like help track down
pieces of PD code that may be needed, like a marching cubes implementation, etc.
Please let me know if this thing takes off. I love free software and
standards too! (Long live X, C, and Unix!)
Jeff Butterworth
From: Peter Miller <peter@pdact.pd.necisa.oz.au>
I like the idea. I have a ray-tracer of my own,
which I think is pretty good, but it too is has limitations.
I think a very nice raytracer could be produced,
given the resources which appear to be out there.
Count me in, if the project ever looks like it will get off the ground.
I like the FSF slant, too, although a lean towards being
a little less restrictive.
From: jrc%concurrent.co.uk@RELAY.CS.NET
A great idea.
It would be good if the framework could be object oriented (whatever
that means :-) ) so that new objects could be added with the minimum of
fuss.
It would also be nice if it was written in one of the more modern
languages eg. Modula-3, GNU g++. Both of these are widely available.
If you set up a mailing list please add my name to it.
Good Luck - John Connett
From: dgh@ecr.mu.oz.au (David Hook)
G'day,
The last time I read an article that had this for a title
it was by Jonathan Swift and was about eating Irish babies!
Seriously though...
Myself and a couple of people here would be interested in
helping in such a venture, we have been tossing the idea
around here for some time, although I'd recommend that
official word was got from Pixar first. Just one thing
however, we would prefer that the work was done in the
public domain, rather than under *any* copyright including
the gnu one. There is already enough paranoia in the
industry.
Let me know what happens.
From: Jeremy Wohl <cfreas@csserv2.ic.sunysb.edu>
I'm very glad you have articled my frustrations over the last year,
as I began learning about photo-realistic graphics. Several group graphics
efforts on CompuServe (fractint, etc.) might serve as a model.
Last year, I spent the $15 for Pixar to send me the Renderman reference
manuals. As I don't have more than 1.5 years experience I am siding with
this standard. Perhaps in my naivete, I trust the peer review and industry
feedback it has gone through.
My current development environment is OS/2. I'd be happy to port to
this OS and provide some sort of graphical front-end. Might we also provide
a GUI-independent 3-D editor?
Thanks for an excellent proposal!
Jeremy Wohl
---------------------------------------------------------------------------
Well, there seems to be a fair bit of interest -- what's the next step ?
Craig, would you like to coordinate such a project ? Is there interest
in a mailing list ?
--
ll // // ,~/~~\' David Rowley
/ll/// //l' `\\\ Mortice Kern Systems Inc.
/ l //_// ll\___/ 35 King Street North, Waterloo, ON, Canada N2J 2W9
O_/ 519/884-2251, FAX 519/884-8861, david@mks.com