> I'd like to use httrack to make speed comparaison between 
diferent ISP's,
> by fetching some homepages on a regular basis: it seems 
to be the best 
> tool to simulate a real browsers behaviour.  I could use 
wget, but it gets 
> all urls sequentialy, while httrack is doing it 'right', 
fetching many 
> objects at the same time.
> 
> How accurate are the infos from the 'HTTrack mirror 
complete in 12 
> seconds : 73 links scanned, 72 files written (187443 
bytes overall) 
>  [98779 bytes received at 8231 bytes/sec]' part? 
Especially the '8231 b/
> s' information? 

The reliable information are the "98779 bytes received" 
(which is the RAW data received (TCP data without the 
headers)), and "12 seconds". httrack then just computes 
the "8231 bytes/sec" part with a precision of 1 second and 
display the rounded result.

The "overall" stat is not really accurate, due to cache 
issues and compression/ssl issues.