>   Is there any way to test a downloaded mirror 
without 
> connecting to the site ?  That is, just scan the 
> internal mirror and report any missing files or 
broken 
> links ?
Err.. to report internal links, which point to 
external sites? To report broken links, the engine 
will have to check them remotely!
If you just want to test *internal* broken links, 
either look at the hts-log.txt file, or mirror.. the 
local site in another project!

Example:
<file://c:\my%20web%20sites\foo\index.html>

(drag and drop the top index.html file in a new mirror 
URL window)

Ensure that default filters (+*.gif and so on) have 
been wiped out, or the engine will attempt to connect 
to the outside world.

>Also, I don't quite understand the 'Separated 
> Files' option works.  Could someone explain that 
> please ?
This is ONLY useful to get binary or html files, 
WITHOUT parsing them, and without crawling into. It is 
similar to get files one by one, using a browser.
Not very useful, in fact ;)