Initial revision

doc/programmer/difrac.tex

@@ -0,0 +1,52 @@
+\subsection{The Difrac Subsystem}
+Difrac is a F77 package for controlling a four circle diffractometer
+developed by P. Gabe and White for PC's. This package became only
+available very late. This package has been included into SICS for four
+circle diffraction. All the code for the difrac package resides in the
+difrac subdirectory to sics. For more information about the difrac
+system see the documentation coming with difrac. In order to
+integrate difrac into SICS the following steps had to be taken:
+\begin{itemize}
+\item Define an interface from SICS to DIFRAC. 
+\item Modify the DIFRAC hardware control to run the hardware through
+SICS.
+\item Adapt I/O
+\item Modify DIFRAC to be a subsystem instead of a program of its own. 
+\item Adapt measurement procedures to the special needs of TRICS.
+\end{itemize}
+
+\subsubsection{The Interface SICS to DIFRAC}
+This is implemented in difrac.c. It it is just a interface wrapper
+function which takes the command given and forwards it to the DIFRAC
+interpreter. A tricky bit about this ist, that this module has to keep
+track of which cconnection issued the command. This is done through a
+connection stack. 
+
+\subsubsection{DIFRAC HW control}
+DIFRAC controls the hardware through a couple of function originally
+defined in the file ralf.f. This file has been replaced by a file
+trics.f which in turn calls C-functions implemented in difrac.c which
+do the actual job. 
+
+\subsubsection{DIFRAC I/O}
+In contrast to many F77 prograsm DIFRAC has a very intelligent I/O
+system. All input is channelled through the function GETLIN in
+gwrite.f. This function has been modified to call a C-function which
+gets a line of input from SICS. This adpater is again in difrac.c. All
+output is printed first into a internal string buffer, COUT, and then
+printed using the function GWRITE. This function has been modified to
+call a C-function which writes the output to a SICS connection. The
+integration of DIFRAC into SICS was only possible because of the
+existence of this I/O system. 
+
+\subsubsection{Changes in DIFRAC}
+The DIFRAC main program was taken apart into a initialisation routine
+(difini.f) and a callable interpreter  function (difint.f). Further
+changes were required because TRICS does not do continous scans in the
+way a X-ray four circle diffractometer does. DIFRAC's concept of a
+scan drives a motor slowly and collects data. This has been modified
+to do step scans for TRICS. Some commands which do not make sense
+within SICS have been  disabled. Some little changes to measurement
+routines were necessary.
+
+