> I think this can be easily prevented by creating a hash 
> (MD5) code of all downloaded files. If HTTrack downloads 
a 
> file whose hash code matches a previous downloaded file 
it 
> can just point the link to that other file. 
> Am i right?
Hum, yes and no. Yes, the method would work - BUT httrack 
names links BEFORE actually downloading them ; and knowing 
that a link is the same as another obne is in fact useless 
AFTER downloading it, unfortunately.
Yes, the engine could generate on-the-fly links, waiting 
for all pending files to be completed ; but this would 
cause other problems, and would slow down the whole 
download process in a terrible way.