> I think the concept of <file>.html.tmp makes this
> program useless for me.
> 
> You cannot check that it works before it finishes. 
There is no concept. Tmp files are temporary binary files for internal
processing. They are gone when the actual file is download and processed. NOT
when the mirror is done.

> For a large site, this makes no sense.  
As opposed to saving GB of internal data in memory?
> a) So I would need to copy a subset of the site just
> to confirm that it will work?Ignore the tmp files. Look only at the htmls.

> b) If a site takes several days to complete then how
> can I shutdown my computer and continue the next
> day?1) Pause (minimize program,) hibernate if necessary, resume later.
Starts right were it stopped.
2) Cancel (press once only). Continue will reverify completed files (about 3
files/second depends on round trip time, independent of file size).