Sorry for always talking about httrack problems - it's a 
great tool! (That's why I spend so much time with it :-)

Just by accident I discovered, that httrack -g would 
download all the scattered java classes by parsing the main 
class in the <applet code=> parameter. This is exactly what 
I need - even though it's an unexpected behaviour of the -
g "just download files" option.

My problem now: My classes (see below) only work if they 
are located in the same directory. But httrack would 
download the rest of the classes in a nested folder 
hierarchy (from a java package structure?) and no -N1003 
would help to get all the class files to the same -O folder.

And there seems to be an problem with httrack, because 
winhttrack would download only correctly download 

<http://www.arachnoid.com/gravitation/index.html>

with all values set to default (because then the main class 
is at it's original position)

But one has got a problem finding the classes when changing 
the site structure to flat (everything in site_name)! In 
this case the classes from java parsing are hidden in a 
nested structure while the main class is correctly located 
at the uppermost folder level.

My suggestions:
-g option should not parse the main class but only download 
the explicitly listed files
-N1003 should store all classes found to the same -O folder
       (which works if done by hand in the above example)

Frank