> I did as described above (i.e. replaced 
> www.somewebsite.com/somedirectory with 
> www.somewebsite.com in the url settings of the 
> previous project), and here I got a great strike 
from 
> HTTrack:
> It deleted all previously downloaded files and began 
> to download them again.
> So it did not work, and all files are lost. Why? 

Well, it deleted all files *first* ? Or after an 
unsuccessfull first attempt? The purge code is placed 
after the mirror code, and therefore the engine never 
clean anything before downloading ; BUT if the mirror 
did not download files that were present, they'll be 
erased.