structure of, errors in generated iv

[Seth Teller <teller@lcs.mit.edu>]

sean, will --

upon closer inspection, i'm unable to view many of
the 3D.iv files (example: 3D.iv in directory

/scratch/willtor_extrude_MIT/willtor_new_run/test_run2/mit_W7

).

that file starts with a huge list of vertices --
something like 100,000 or so (!) -- then a long
list of individual, switch-enclosed face DEFs
followed by color USEs and face USEs.

i would suggest restructuring things to have
   1) *fewer* vertices per coord3 block --
      a few hundred at most

   2) *more* faces per instance block --
      a few dozen or hundred at least.

of course, "block" is a vague notion-- blocks should
be some meaningful grouping of source geometry.  for
example a block could be all of the geometry in a 
named space.  or a block could be all of the geometry
that fits into a few hundred vertices (once this limit
is reached you could close off the block, issue its
USEs, and start a new one).

clearly this is another thing to think about and get
right.

also, there are many error messages in the W7.?.errors
files in that directory -- unknown group, etc.  can
you track down the cause of these?

prof. t.