I would like to see the test run again with the following change.
From each provider test the response time of the other 28 sites and not
the providers own web server. Then average the response times for these
other 28 web servers and report that average response time from that
provider. The providers with good connectivity to the rest of the net
should have lower average response time.
P.S. One might also be interested in the top one hundred web sites
average response time.
Peter Cole of Telescan, Inc. (281)588-9155
Better computing through lack of sleep.