On Fri, 2006-10-13 at 20:34 +0530, Rahul Sundaram wrote: > >>> There are 2 advantages. First, you can control what machines > >>> act as mirrors and thus not have to arrange anything with > >>> other parties to get access to the logs. > >> How would a separate repository solve the problem? The repo information > >> would still be distributed and we still need to coordinate with the > >> mirrors to get more information. > > > > It should not take many machines to robustly distribute a single > > file. Keep all mirrors on machines you already control. > > Red Hat hosting all the mirrors in the same place would mean that its > not geographically distributed and it would also mean taking up all the > bandwidth costs. They don't have to be in the same place - they just have to be on machines where there is no question about your access to the logs. And how much bandwidth does it take to distribute one file that changes once a month? > > How can I keep my information from being collected in logs on > > repositories where I actually need the files? You may claim > > you have a way that will keep you from using them, but how > > can you prove that to me? > > What exactly is your concern from the project collected anonymous stats > from mirrors logs on the server side? I cant see any reason why you > would want to avoid that. If you do, use a local mirror. I don't personally have a problem with this so I can't elaborate, but I've seen others suggest a privacy concern that should probably be respected. I thought you did mention some mechanism to opt out anyway. What I'm saying is that the only really effective mechanism is to not need to connect to the site at all, which would be the case with an optional repository. -- Les Mikesell lesmikesell@xxxxxxxxx
