> Don't experts use GUIs? Aren't experts tempted 
> to abuse these tools? This begins to smack of
> elitist design. 

No, no, don't be misunderstood. I think the problem raised by "imacat" is
real: not allowing beginers to use potentially dangerous options, *they are
not aware of*. Downloading with high speed can be harmful for a regular
server, and most people are just not aware of this fact. Nowadays, you can get
high bandwidth with domestic DSL lines - such as 20Mbits for reasonnable fees.
People may not understand that downloading at these speeds can be costly for
many servers, and will select a rate of "10MB/s" without knowing that it can
clobber the remote server and hog its network.

By "experts", I mean people that are aware of the consequences of using very
high transfer rates and/or multiple connections (>10 simultaneous
connections)

A load test, for example, can use 64 simultaneous connections, and several MB
of bandwidth.
An automatic mirror, done each night, can also use similar settings. Web
archives can also use time to time aggressive settings to make a "realtime
snapshot" of a website.

By using a _very_ explicit option, tagged as "dangerous", people will NOT be
able anymore to ignore what they are doing.

The'll still be able to abuse. But I think that the problem will be minor.

This is exactly like "ping -f" or "nmap -T insane": you can overload a server
with these two commands, more than any offline browsers will ever be able to.
But these commands are explicitely tagged as "dangerous", and not accessible
immediately to beginners.

You'll still be able to use the "full potential" of HTTrack. But using some
magical and clearly identified options.

I think this is a good and reasonnable solution for everyone (beginners,
experts, web administrators)