Added RBSTFIT0 to l1_utils.f90 a variant of the robust line fitting routine

	Added RBSTFIT0 to l1_utils.f90 a variant of the robust line fitting routine
	RBSTFIT that assumes that line is defined by a single scale parameter. i.e.
	Y = SLOPE*X as opposed to Y = SHIFT + SLOPE*X

l1_utils.f90

@@ -117,6 +117,76 @@ END SUBROUTINE RSTFIT
 ! -----------------------------------------------------------------------
 ! -----------------------------------------------------------------------
 
+SUBROUTINE RSTFIT0(X,Y,NDATA,SLOPE)
+!+
+!  Purpose:
+!     Robust straight line fit using simple L1 algorithm method.
+!
+!  Description:
+!     This routine fits the line Y = B*X to a set of points X,Y by the
+!     criterion of least absolute deviations.
+!
+!     This version calls on the L1 subroutine as taken from the TOM archives
+!     (referred to as TOMS algorithm 478). This is a fast stable and simple
+!     routine which is satisfies our needs. See l1.F95 for further details.
+!
+!     Notice the solution is not guaranteed to be unique and in such cases the
+!     answer is determined by the order of the data points given. However this
+!     case is sufficiently rare to not worry about. 
+!
+!  Arguments:
+!     X = REAL VECTOR(NDATA) (Given)
+!        X data points to be fitted
+!     Y = REAL VECTOR(NDATA) (Given)
+!        Y data points to be fitted
+!     NDATA = INTEGER (Given)
+!        Size of X and Y vectors
+!     SLOPE = REAL (Returned)
+!        Slope (or gradient)
+!
+!-       
+   IMPLICIT NONE             ! No implicit typing
+   
+   ! Arguments
+   INTEGER NDATA
+   REAL X(NDATA),Y(NDATA)
+   REAL SLOPE
+
+   ! No of parameters to solve for
+   INTEGER,PARAMETER::N=1
+   
+   ! N2 = N+2
+   INTEGER,PARAMETER::N2=N+2
+   
+   ! Local variables :
+   INTEGER M,M2
+   REAL TOLER
+   REAL A(NDATA+2,N2) ! A has extra rows and cols to store return values
+   REAL XS(N)         ! solution vector
+   REAL B(NDATA)      ! Y data input vector
+   REAL E(NDATA)      ! workspace vector
+   INTEGER S(NDATA)
+   
+   ! ???
+   M = NDATA
+   M2 = M+2
+   TOLER = 10**(-12*2/3)  ! safe value below which is deemed zero
+
+   ! fit function Yi=a*Xi+b*1.0, so data columns are Xi and 1.0 for i=1,ndata
+   ! These are placed into A array whilst Y values are plced in B array
+   A(1:M,1) = X(1:M)
+   B(1:M)   = Y(1:M)
+
+   ! Call L1 routine, solution returned in XS vector
+   CALL L1(M,N,M2,N2,A,B,TOLER,XS,E,S)
+      
+   SLOPE = XS(1)
+   
+END SUBROUTINE RSTFIT0
+
+! -----------------------------------------------------------------------
+! -----------------------------------------------------------------------
+
 SUBROUTINE RSTFIT2(X,Y,NDATA,X2,Y2,NDATA2)
 !+
 !  Purpose: