* Christopher Hicks ([EMAIL PROTECTED]) wrote : > On Wed, 27 Nov 2002, Haesu wrote: > > I think both of these are fine ideas. We just need someone willing to > step forward with an implementation. :-) > > Everybody has expressed a lot of factors of what makes a good mirror, but > if we combined those into a numerical "mirror quality quotient" and listed > that information on the site it would naturally help people find the right > mirror for them. To start things off I propose: > > b + 48/r - a/4 > > Where: > - b is bandwidth in Mbps > - r is how many hours between rsyncs > - a is how many point releases your apache is behind
the first two parts are useful, the third is totally irrelevant. Will people please get off the idea that (a) mirror maintainers have any control over what services are running on their server, and (b) that it makes any odds at all to the quality of the mirror. As an example of this, www.openbsd.org runs on a solaris box, but that doesn't seem to affect the mirror quality. -Thom
