APPENDIX. PROGRAM LISTING

The program listing is

# Extend a few traces off the end of a CDP gather using the vspray operator.
#
#	<in.H Test np=6 eps=1. vspace=vspace > out.H
#	np is the number of padded traces off end, nx+1,...,nx+np
#	vspace is a space of slowness squared.
#
# keyword missing velocity
# JFC
#%
integer nt, n2, n3, esize, nx, nv, np, nxp, niter
from history:	integer n1:nt, n2, n3, esize
			nx=n2
from par:	integer nv=2*nx,  np=6, niter=2
nxp = nx + np
allocate:	real xx(nt, nxp), vv(nt, nv), pp(nt, np)

subroutine dumb( niter, nt, nx, np, nxp, nv, xx, vv, pp)
integer	it, nt, ix, nx, iv, nv, ip, np, nxp, iter, niter
integer rect1, rect2
real	xx(nt, nxp), vv(nt, nv), pp(nt, np)
real 	t0, dt, x0, dx, s0, ds, p0, dp, sum, eps, smax
temporary real             dpp(nt,np), spp(nt,np)
temporary real  rr(nt,nv), dr (nt,nv), sr (nt,nv)
temporary real  ww(nt,nv)
temporary real  wr(nt,nv)
temporary real  rrr(nt,nv)
from history:	real o1:t0, d1:dt
from history:	real o2:x0, d2:dx
from par:	real eps=1.
call sreed('in', xx, 4 * nt * nx)
call hclose()
smax= 1./ 1.0 **2
s0 = -smax / 5.
ds = (smax-s0)  / (nv-1.)
dp = dx
p0 = x0 + nx * dx
do it=1,nt  { do ip=1,np  { xx(it,nx+ip) = 0. }}
call vspray( 0,  nt,dt,t0,  nx,dx,x0, xx,  nv,ds,s0, vv)

do it=1,nt  { do iv=1,nv  { ww(it,iv) = abs( vv(it,iv)) }}
rect1 = nt/10
rect2 = 3
call triangle2( rect1, rect2, nt,nv, ww, ww)
sum = 0.
do it=1,nt  { do iv=1,nv  { sum = sum + ww(it,iv) }}
do it=1,nt  { do iv=1,nv  { ww(it,iv) = 1. / (ww(it,iv)+eps*sum/(nt*nv)) }}

do it= 1, nt {
	do iv= 1, nv
		rr(it,iv) = - vv(it,iv)
	do ip= 1, np
		pp(it,ip) = 0.			# padded data initialize
	}

do it=1,nt  { do iv=1,nv  { wr(it,iv) = rr(it,iv) * ww(it,iv) }}

do iter= 0, niter {
	do it=1,nt  { do iv=1,nv  { rrr(it,iv) = wr(it,iv) / ww(it,iv) }}
	do it=1,nt  { do iv=1,nv  { dr(it,iv) = wr(it,iv) * ww(it,iv) }}
	call vspray( 1,  nt,dt,t0,  np,dp,p0, dpp,  nv,ds,s0, dr)
	call vspray( 0,  nt,dt,t0,  np,dp,p0, dpp,  nv,ds,s0, dr)
	do it=1,nt  { do iv=1,nv  { dr(it,iv) = dr(it,iv) * ww(it,iv) }}
	call cgstep( iter, nt*np, pp, dpp, spp,
			   nt*nv, wr, dr,  sr  )
	do it= 1, nt
		do ip= 1, np
			xx(it,nx+ip) = pp(it,ip)
	}
call srite('out',    xx, 4 * nt * nxp)
call srite('vspace',rrr, 4 * nt *nv)
return; end

The subroutines appear in previous sep reports and perhaps in this one. The interactive report (if any) should have pushbuttons to bring up the relevant library subroutines here.