Dirhtml is quick and easy, but for some reason, the html that it creates stops
Httrach dead in its tracks just as quickly as <file:///C:/>. Even using the
"build files recursively" option, where the links point to actual files,
instead of just directories.

I have harvested files from comercial web sites using
<html://www.somesite.com/directory> with httrack, where I assume there is no
actual html for the directory either.  It allows me to skip the index and
image html and go directly to the jpegs. The output from my browser appears to
be of similar composition for <file:///C:/> as it does for
<html://www.somesite.com/directory>.