Path: utzoo!yunexus!geac!syntron!jtsv16!uunet!husc6!bloom-beacon!gatech!udel!mmdf
From: kvancamp@ardec.arpa (Ken Van Camp)
Newsgroups: comp.sys.amiga
Subject: Re:  Sculpt-4D Upgrade
Message-ID: <4779@louie.udel.EDU>
Date: 11 Oct 88 20:48:29 GMT
Article-I.D.: louie.4779
Sender: mmdf@udel.EDU
Lines: 66


 >I feel that ray tracing is too exotic a process
 >to be valuable to production on a small machine. Scanline renderers are MUCH
 >faster and produce virtually the same result.

I hope you didn't mean that first statement as strongly as it came out
sounding... there are many things that you just can't do with scanline
rendering.  Nonetheless, I agree that it has been under-utilized on the
Amiga.  And when it has been utilitized, it has been brain-damaged.  Why
can't these people do smoothing in scanline mode?  It's a lot easier to
implement than in ray-tracing mode...

 > I think they should also offer some kind ascii format of their objects. This
 >way someone could write a myriad of programs to manipulate te objects which 
 >could supplement the main programs.

What's that?  You don't read IFF?  Script files in Sculpt-3D **do** work, 
and you can generate the geometry with it (no thanks to the documentation).  
I had to play around with it quite a bit before I realized what the 
documentation was missing.  The correct way to generate a face (and all its 
vertices and edges) is with a line like: 

(x1,y1,z1)-(x2,y2,z2)-(x3,y3,z3)-(x1,y1,z1)

That's right, you have to repeat the first vertex!  The documentation
doesn't make this clear, but I finally figured out that you're just
generating edges when you specify coordinate pairs, and then Sculpt-3D
checks all the edges and constructs faces only when it finds three
connected edges.  This (and the fact that it checks every vertex to
make sure it isn't a repeat) makes scripts **extremely** slow to load.
As I recall, there were also a couple of obvious omissions in the script
language -- I think there's no way to set the observer location or target,
and maybe one or two others.

Anyway, I've given up on the script files and have written some routines
that generate the IFF scene files, and they're **much** easier to work
with.  Since my surface modeler reads ASCII files, and I wrote a program
that converts my surface model files into Sculpt-3D scene files, in
effect I have what you have asked for: a way of inputting Sculpt-3D data in 
an ASCII file.  It's not finished (currently only writes geometry info), but 
if you want a copy anyway you can mail me a disk with return postage (P.O.  
Box 784, Stroudsburg PA 18360).  (What's that?  You've never heard of my 
surface modeler?  That's because it's just a small feature in a larger, 
commercial product -- to be released "some time next year".) 

 > What is 3-Tuple? Is it a complete system like Sculpt- Animate? Who is Fred
 >Mitchell? Has he said anything about reflection mapping?

Yes, it is a complete system, written by Fred Mitchell, president of 
MitchellWare.  It hasn't been released yet, but I spoke with Fred after last 
month's JAUG meeting and he said he had already implemented texture mapping.  
I'm not sure if reflection mapping is the same as texture mapping, but 
3-Tuple sounds to me like a very powerful package.  There was actually a 
"review" in last month's Amiga/World, but it was VERY deceptive and I 
wouldn't pay much attention to it.  (Either Fred lies alot about features, 
or the copy Amiga/World got was really old, or the reviewer didn't have the 
slightest idea what he was talking about.) Instead, just use the address on 
the review to write to MitchellWare and ask for a brochure.  It tells much 
more.  

                            --Ken Van Camp 
ARPANET:  kvancamp@ARDEC.ARPA   -or-   kvancamp@AC4.PICA.MIL
BITNET:   (use above through normal gateways, like UBVM.CC.BUFFALO.EDU)
USENET:   uunet!ardec.arpa!kvancamp@UUNET.UU.NET

          'Tis better to Shell, than to shell out the dough.