> How can I tell HTTrack to download missing files in a 
> project without updating it nor reloading/deleting 
already 
> loaded files ? In other words, how can I tell HTTrack to 
> browse and parse downloaded files then download missing 
> ones (I'm thinking about images that are refered by the 
> HTML files) ?
"Continue an interrupted mirror" - should do the trick. You 
can also erase bogus files and then restart the mirror.
The operation should be fast, even if the engine will have 
to rescan the complete local structure (to rebuild all 
links, and so..)