- Some option that allows grouping after one or more fields
  would be nice, this would also need sorting (like "group by
  field country, after all entries for each country print the
  number of addresses in this group").

- As of now there are some limits regarding the input file (number
  of entries, max number of fields per entry). These should be 
  eliminated by a more efficient memory management.

- The definition of a format should be easier (i.e. not require 
  three files). Maybe all of the definitions should be done in 
  one file, like this:

    <HEAD>
    this is the head
    </HEAD>

    <ENTRY>
    @0[40] - @1
    @3
	</ENTRY>

	<FOOT>
	</FOOT>

- A graphical user interface would be really nice, but also rather
  complicated ...

- Support more input files (dbm, gdbm ?). Don't know if this is really
  needed.
