UNI2TRI - convert UNIformly sampled model to a TRIangulated model

uni2tri <slothfile >modelfile n2= n1= [optional parameters]

Required Parameters:
n1= number of samples in first (fast) dimension
n2= number of samples in second dimension

Optional Parameters:
d1=1.0 sampling interval in dimension 1
d2=1.0 sampling interval in dimension 2
f1=0.0 first value in dimension 1
f2=0.0 first value in dimesion 2
ifile= triangulated model file used as initial model
errmax= maximum sloth error (see notes below)
verbose=1 =0 for silence
=1 to report maximum error at each stage to stderr
=2 to also write the normalized error to efile
efile=emax.dat filename for error file (for verbose=2)
mm=0 output every mm-th intermediate model (0 for none)
mfile=intmodel intermediate models written to intmodel%d
method=3 =1 add 1 vertex at maximum error
=2 add vertex to every triangle that exceeds errmax
=3 method 2, but avoid closely spaced vertices
tol=10 closeness criterion for (in samples)
sfill= x, z, x0, z0, s00, dsdx, dsdz to fill a region

Notes:
Triangles are constructed until the maximum error is
not greater than the user-specified errmax. The default errmax
is 1% of the maximum value in the sampled input file.

After the uniform values have been triangulated, the optional sfill
parameters are used to fill closed regions bounded by fixed edges.
Let (x,z) denote any point inside a closed region. Values inside
this region is determined by s(x,z) = s00+(x-x0)*dsdx+(z-z0)*dsdz.
The (x,z) component of the sfill parameter is used to identify a
closed region.

The uniformly sampled quantity is assumed to be sloth=1/v^2.



AUTHOR: Craig Artley, Colorado School of Mines, 03/31/94
NOTE: After a program outlined by Dave Hale, 12/27/90.