added comments

git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/Boundary/trunk@32 6a38eb6e-646e-4a02-a296-d141613ad6c4

1 files changed, 68 insertions, 0 deletions

diff --git a/src/RadiationBoundary.c b/src/RadiationBoundary.c
index 706979a..c65f971 100644
--- a/src/RadiationBoundary.c
+++ b/src/RadiationBoundary.c
@@ -1,6 +1,74 @@
 #include "cctk.h"
 #include "cctk_parameters.h"
 
+/*@@
+   @file    RadiationBoundary.F
+   @date    Mon Mar 15 15:50:14 1999
+   @author  Gerd Lanfermann
+   @desc 
+            File contains the C routine, which performs the 
+            actual GF field assignments for the radiation boundaries.
+	    Since this is now in C, there is really now reason to have 
+	    the wrapper around. Anyhow, for now we keep this struture.
+   @enddesc 
+@@*/
+
+
+
+/*@@
+   @routine RadiativeBC
+   @date    Mon Mar 15 15:51:57 1999
+   @author  Miguel Alcubierre
+   @desc
+            Implements radiative boundary condition.
+            The boundary condition that is implemented is:
+
+
+               f  =  f0  +  u(r - v0*t) / r
+ 
+
+            which leads to the differential equation:
+
+
+               (x / r) d f  +  v d f  + v0 x (f - f0) / r^2  =  0
+                 i      t         i         i
+
+            where xi is the normal direction to the given boundaries.
+
+            That is, I assume outgoing radial waves with the correct 1/r
+            fall off, and the correct asymptotic value f0 (f0 = 1 for 
+	    diagonal metric components, and f0 = 0 for everything else),
+            but at a given boundary I only worry about derivatives in
+            the normal direction.  This works very well.
+
+            The only problem with this condition is that it does not
+            use the physical characteristic speed, but rather it assumes
+            a wave speed of v0, so the boundaries should be out in
+            the region where the characteristic speed is constant.
+            Notice that this speed does not have to be 1.  For gauge
+            quantities {alpha,phi,trK} we can have a different asymptotic
+            speed, which is why the value of v0 is passed as a parameter.
+
+
+	    Some remarks on the C version for Fortran programmers: 
+	    this is 1:1 translation of the 
+	    F code, which used the very nice looking F array assignments 
+	    x(2,:,:), etc. In C this needs to written as loops, but to 
+	    my surprise it doesn;t look to bad actually. F statments like
+	    x(1,:,:) becomes x(xgp0) (the integer index "xgp0" meaning
+	    the zero'th gridpoint from the boundary), x(2,:,:) --> x(xgp1),  where: 
+	    xgp0 = CCTK_GFINDEX3D(GH,0,j,k) 
+	    xgp1 = CCTK_GFINDEX3D(GH,1,j,k) and i,j are looping of the grid sizes.
+
+   @enddesc 
+   @calls     
+   @calledby   
+   @history Cactus 4.0 by Gerd Lanfermann
+ 
+   @endhistory 
+@@*/ 
+
+
 #define SQR(a) ((a)*(a))
 
 void RadiativeBC(cGH *GH, int *doBC, int *sw, int *lsh, int *dxyz, int dt,