🚀 go-pugleaf

<food for line-eaters>

I have some experience with installing and using movie.byu, and a few minutes
to relate my impressions. Let me say that we also wrote some of our own
software for surface triangulation from contours and for rendering of smooth-
shaded surfaces before we got movie, so we have a basis for comparison.
The version we used was the VAX distribution tape of 1984 or 1985.  I
know of a more recent SUN version and a mini-movie intended for < 32bit
machines but have no experience with them.  BYU is good about distribution
and runs training courses.  In my experience they aren't very
reachable for phone support.

Movie.byu is written in FORTRAN, the code is somewhat understandable
and sparsely commented (as I recall).  It is fairly large in terms
of source lines.

-rwxr-x---  1 noel       258594 Apr 16  1985 command.f*
-rwxr-xr-x  1 noel        26390 Apr 16  1985 compose.f*
-rw-r--r--  1 noel        34662 Apr 16  1985 device.f
-rwxr-x---  1 noel       149146 Apr 16  1985 hidden.f*
-rwxr-x---  1 noel       166702 Apr 16  1985 mosaic.f*
-rwxr-xr-x  1 noel        26897 Apr 16  1985 movlib.f*
-rwxr-x---  1 noel       140973 Apr 16  1985 section.f*
-rwxr-x---  1 noel        21793 Apr 16  1985 title.f*
-rwxr-x---  1 noel       235563 Apr 16  1985 utility.f*
Our Makefile in part:

display: command.o hidden.o device.o grin.o f_to_c.o movlib.o ${GRLIB}
	f77 -o display command.o hidden.o device.o grin.o movlib.o ${GRLIB}

mosaic: mosaic.o device.o grin.o movlib.o ${GRLIB}
	f77 -o mosaic mosaic.o device.o grin.o movlib.o ${GRLIB}

compose: compose.o device.o grin.o movlib.o ${GRLIB}
	f77 -o compose compose.o device.o grin.o movlib.o ${GRLIB}

title: title.o
	f77 -o title title.o movlib.o

utility: utility.o  movlib.o
	f77 -o utility utility.o movlib.o

all: display mosaic compose title utility

We use movie to take 3-D objects (nerves and other neuroanatomical structures)
represented by stacks of contour boundaries, convert them to triangular
surface-patch representation (using the MOSAIC program),
and to DISPLAY them with translucency, color,
shadows, etc. on an 8-bit deep raster display.

It was fairly easy for me to write a driver for my display (Grinnell).
Reconfiguring movie for the large objects (10s of thousands of points)
we use was a bit of a pain as the array sizes were all numeric constants
and interrelated and documentation of what to change was only so-so.
(are they using PARAMETERS yet?).

In general the programs are highly interactive (read lots of typed commands,
in a fairly hairy multi-level command hierarchy) which gives you the control
you need but it takes a LONG time to get nice pictures out.  You can speed
up computation by previewing at 256x256 resolution, but you can expect
to spend 10-15 minutes on a VAX 780 to calculate a moderately complex
color picture at 512x512 (doesn't matter how many shades).

All input and output files are ASCII and human-readable.  This means they
take up a lot of space and are easily transportable.

We never tried to animate sequences of views, but I gather this works
fairly well if you have the CPU/human time to do it. We never used TITLE
to insert text or COMPOSE to join model files together, preferring to
join them in ASCII form before feeding to movie.

Our version had some bugs regarding shadows and translucency was hard to
get right.  I believe they have fixed their bugs and are continuing to
develop the system.

I will be happy to give more info on what I know if it is relevant...
what kinds of things do you want to know about?
You can get a book from them describing the operation of the programs
for some amount of $$.  I will post their address if you want.

-- 
-- 
-- Noel Kropf	{philabs,rna}!cubsvax!noel	914-739-7655
--		 cubsvax!noel@columbia	(Internet, ARPAnet)
--		1002 Fairchild; Columbia University; New York NY 10027
--
-- I love a tasty tune - una salsita saborosa !