
shallow96
	shalpre96  creates shobs.d

shobs.d:
  0.00499999989  0.00499999989  0.  0.00499999989  0.
    1    0    0    0    1    1    1    0    1    0
start.mod                                                                       
nnall.dsp
time.dat


ttinv96
	ttinvpr96  creates tobs.d

tobs.d:
    1    0    0    0    1    1    1    0    1    0
start.mod                                                                       
time.dat


The format of the time.dat file is

FILE_NAME OF TRAVEL TIME FILES each of which has entries such as

TYPE DISTANCE  TIME ERROR EVDP(km) RECDEP(km)
where
TYPE is P S SV SH
DISTACNE in km
TIME in sec
ERROR in sec
EVDP in km
RECDEP in km

---------
In the ttinv96 and the shallow96 programs the internal variables for the inversion will change
For surf96, rftn96 (not used) and joint96, the inversion was for Vs and Qs^-1

For the ttinv96 and the shallow96 the inversion will be for Vp and for Vs independently
This will change the meaning of the weights

---------------------
Look at VOLIV/src/srfpre96.f at subroutine getdsp

This is a routine that converts human input to the parameters required by the program
Thus the human dispersion line is

FILE TYPE CUG TX PER VAL  DVAL
SURF96 L C T     0    1.0000        3.2085       0.10000E-02
SURF96 L U T     0    1.0000        2.9010       0.10000E-02
SURF96 L G T     0    1.0000       0.41043E-02   0.10000E-06

The subroutine  getdsp read this and then outputs to the program file on unit=8, e.g., tmpsrfi.12
Note that getdsp permits lower case as well as upper case

We need a subroutine gettim
The output will be on file tmpsrfi.20

1 10.500000 1.950000 0.5 0.0 0.0

Where
0 = P
1 = SV
2 = SH
3 = S

We will use the lgstr, getblnk

subroutine gettim ( time.dat, ntobs)

ntobs = 0
open time.dat
open the tmpsrfi.20
for each file name in time.dat
   open the file name
   then read each line
       write one line to tmpsrfi.20
       ntobs = ntobs + 1
   close the file name
close time.dat
close tmpsrfi.20
return
    


--
At the upper level

    call gettim (time.dat, ntobs)
------------------------
change ttinvdr96 to recognize 1 2 3 for S

----------
lstinv needs more values for shallow
tmpsrfi.02 needs an oa(i) to complement the ob(i)

----------
in ssamat and stamat set up P and S rows properly

in  shamat why is there a close(1) but not and open(1)? is this from shallow96?
