On Sun, 3 Nov 2002, Steven M. Bellovin wrote:

# I wonder if the issue is system performance vs. program performance on
# a microbenchmark.   That is, a whole system might perform faster if
# only one copy of libc had to be resident, but each individual command
# might be a bit slower.

That's certainly a trade-off, and a confusing one to most folks,
especially those who are sensitive to the speed of interactive/scripting
sessions.  How can you tell, for example, that the system is faster when
all the commands run sluggish?  [exaggeration, but run with it, here.]

				--*greywolf;
--
"Windows/NT - From the people who brought you EDLIN".