[comp.graphics] NURBS data available

fish@cs.utah.edu (Russ Fish) (03/27/91)

Greetings, computer graphites and NURBologists!

There has been interest in a trial distribution of some NURBS geometric models
created with the the Alpha_1 geometric modeling system.  As a first whack,
Beth Cobb and Robert Mecklenburg packaged up half-a-dozen fairly simple
models.  

We've done a lot of strange mechanical parts, but these are all models of
"common household objects", in the tradition of the Utah teapot and
volkswagen.  The geometric constructions of most of them show up in the
src/shape/examples directory of the Alpha_1 distribution tape...


To quote the beginning of the README file:

    This directory contains six simple models described as NURBS.  Each file
    contains the following notice, followed by one or more spline surface
    descriptions:

	    These models were created using the Alpha_1 geometric
	    modeling system at the Computer Science Department,
	    University of Utah.  We would appreciate your
	    acknowledging the source of the data in publication
	    of work utilizing these models, or portions of them.

    The text format in which these models are distributed is intended to be
    mashed into whatever form you want in order to read the data into your own
    software.  No parser or filters for this text form is (nor will be)
    provided by the University of Utah.


Here's a listing of the data file names and sizes.

    % wc -c *.txt
       13960 coke.txt
	5572 dart.txt
	6136 goblet.txt
       11475 pencil.txt
       30299 scissors.txt
	2447 spoon.txt
       69889 total


The compressed tar file is 28106 bytes long, available for anonymous FTP from:

    graphics/nurbs-models.tar.Z on CS.UTAH.EDU or UUNET.UU.NET .

(I understand UUNET subscribers can get it via UUCP from them, as well.
Thanks to UUNET for adding this to their archive.)

I also dropped some corresponding Alpha_1 .rle.Z pictures onto a
~ftp/graphics/images directory on cs.utah.edu .  The picture files are much
larger than the data files, of course...


If you don't have FTP or UUNET UUCP access and want a copy of the data, send
me a message and I'll consider posting the data to an appropriate newsgroup or
e-mailing it to you.  Please address queries or requests to me directly rather
than posting to the net, since I don't normally read news much anymore.


We'd also be interested in hearing about it if you like the data and/or want
to see more.  People here have done a *lot* of NURBS models over the years...

Cheers!
-Russ Fish			fish@cs.utah.edu		(801) 581-5884
--

-Russ Fish			fish@cs.utah.edu		(801) 581-5884

fish%gr.utah.edu@cs.utah.edu (Russ Fish) (03/27/91)

Hi, folks.  Seems we had a problem with missing data in the first rev of the
nurbs-models.tar.Z file, so I tracked it down and tried again.

The file is now 68537 bytes, instead of 28106 bytes, so you can see how much
got eaten!  The new version is now on cs.utah.edu ~ftp/graphics, and I assume
the copy at uunet.uu.net will be updated soon as well.  (Thanks, UUNET!)

The individual file sizes inside the .tar.Z file are now:
    2881 README
   58495 coke.txt
   25079 dart.txt
    6136 goblet.txt
   74777 pencil.txt
   30299 scissors.txt
    9215 spoon.txt
  206882 total


In case you're interested, here's an explanation of what happened: We pass
structures of C++ objects between programs inside Alpha_1 using a machine
independent binary object stream.  There is a utility to dump that out in a
simple-minded temporary text form or read it back in, which is used primarily
for debugging since the text form will change as we evolve the object
structures.  (The binary form provides type migration via polymorphism based
on interpreting object structure descriptions embedded in the stream.)

Rather than forcing the rest of the NURBS world to read our binary form, we
tried a hack: passing the text form through a script to strip out all of the
topology and attribute information, leaving just the surfaces in a slightly
prettified form.  (Hence the somewhat curt statement in the README file:

    The text format in which these models are distributed is intended to be
    mashed into whatever form you want in order to read the data into your own
    software.  No parser or filters for this text form is (nor will be)
    provided by the University of Utah.

Unfortunately, the attribute stripping was a bit too enthusiastic...


Further bugs, problems, queries, etc. to me via e-mail, please.

(somewhat) Chagrined,
-Russ Fish			fish@cs.utah.edu		(801) 581-5884