One key point for the split implementation is the management of the overflow
space.  Overflow pages fall into one of three varieties (implicitly).

a) currently in use,  b) never used, and c) previously used but currently 
unused.

We must manage all three varieties in our program;   the first variety will
always appear on a chain,  the second by means of an offset pointer, and the
third via a list or array.  The list can be managed on disk as a linked list or
as an array in main memory that we will write to disk when the file is to
be closed.  This may not fit into a fixed sized header, so the previously used 
but currently unused data will be written to the end of the client data. 
(or even better within the pages that are being listed!)  Then we only need to 
write probably a page at most.   Maintaining the array in our program will 
result in fewer disk operations than writing the linked list to the file.

Please use N = 1 in this exercise.  


The makefile according to the handout is to make three programs.  
It will suffice if it makes the executables  lhAccess and lhIndex.

interactive version:

        % lhAccess  Datafile name  

                author:   <type in author as it appears in the index>
                
                the program returns the relavent summary record offset values.

                author:


Or for a non-interactive version:

        % lhAccess  <datafile name>  <author name>

                outputs a list of summary record offset values.


For index construction:     (this program is not really interactive!)

        % lhIndex <data file name>  <alpha max>