5 - How does it work?

   0) g2h reads the input file and creates a directory named exactly as the name of the input
      guide file. In this directory, g2h will put all nodes of the input guide file, translated in html
      files, named exactly as the nodes names, suffixed by ".html". 
      When a link button is encountered, g2h has for now, three alternative:

   1) The link button is a node: g2h creates an html file named exactly as the node and puts in
      it the converted text as step 0.

   2) The link button is another file: g2h try to read "@database" as first characters in this
      file. If it success, g2h create a new directory named exactly as the name of the new guide
      file to convert, and converts it's nodes recursively with the same options you entered for
      the first call to g2h as step 0. If it fails, go to step 3.

   3) If g2h didn't read "@database" in the linked file, it deduces :) it's not a guide file.
      Then, it suppose that's an image file and make a special html file named exactly as the link
      name but suffixed with "img.html". In this file, the name of the image file is inserted as a 
      IMG SRC html tag.

      NB: I know that other file than guide files or image files are supported by AmigaGuide, but
      g2h just do the above things...

   IMPORTANT : All files created are in lower cases and only the last 25 characters of a link are kept
   to make the file names not too long for AmigaOS. So, don't worry if some files names looks to have
   lost their begining...