I probably don't agree that it is not a reliable way to draw conclusions
about backbone performance. The index has drawn criticism because it
points up winners and losers and that's about it. There is no way to do
this without criticism, and indeed it should be part of the process.
The end user experience will reflect the sum of that, plus dialup variables
which we simply do not include in this work.
Jack Rickard
If you had actually read some of the responses without having
made up your mind beforehand (which, as I recall from when I
was a journalism student, is a sure sign that your pretenses
of objectivity are lost), you might notice that both "winners"
and "losers" are pointing out the SAME flaws in the study.
Please go re-read the archives with a more open mind.
I probably don't agree that it is not a reliable way to draw conclusions
about backbone performance. The index has drawn criticism because it
points up winners and losers and that's about it. There is no way to do
this without criticism, and indeed it should be part of the process.
Jack, let me point out that I am one of your largest critics on this one,
and I'm not even up and running yet. I can't win or lose, its just a bad
setup. It is not a reliable way to draw conclusions about backbone
performance. (Note the sentence ends there, there is no probably, there is
no agreement or disagreement, simply a statement. You may believe what you
like.) This /is/ a study of how quickly backbone companies web servers
serve web pages to the locations that you were testing from. I would bet
that Netscape, Playboy, Sun, and a few other people who are very concerned
about their web presence would have better "backbones" than any backbone
provider according to your statistics. Daniel McRobb's post hit the nail
on the head. If KC over at NLANR says that your study is good, I'll
believe her. If /anyone/ who's opinion I trust says your study was good
I'll at least reconsider my opinion. As it stands, I'll stand by my
statements.
The end user experience will reflect the sum of that, plus dialup variables
which we simply do not include in this work.
Justin "I've had more end users on my network than you have ever seen, or
possibly even read about" Newton