>No, we haven't done a rigorous analysis to determine where the >performance issues lie in the installation process, though some people >have done some ad-hoc work and made some proposals based on it. I think >we'd all be very interested in your notes.
One thing that I always find intriguing is the amount of time that passes by with "no disk I/O". This certainly points to something other than "contents" file rewrites. Casper
