On Fri, Sep 14, 2012 at 2:08 PM, Johnny Hughes <johnny at centos.org> wrote: > I am trying to understand how this: > yum install epel-repo > (and installing an rpm maintained by CentOS that is disabled by default, > and requires editing after install) > > is any easier for users than this: > yum install \ > http://ftp.osuosl.org/pub/fedora-epel/6/i386/epel-release-6-7.noarch.rpm > (and getting an enabled repo file with no editing requried) > > The repo is still installed in one step, and no editing is required. > > I find it hard to understand how us including repo rpms, which have the > potential to become outdated and require editing after install are > somehow easier than installing from originator. We are also going to > install the user's pki files (which are in most repo "release" files), etc. > > I am fine to put the rpms in the extras repo if people really think this > is a benefit, and maybe I am not seeing something, but to me this does > not seem to help much. > > What am I missing? The difference is that you need to track down that URL to begin with. That's a significant amount of friction which consists of: switch to your web browser, locate the EPEL site, navigate the site to find the download link, copy/paste the download link into your terminal window. Most of that requires a bunch of mousing around, waiting for pages to load, etc... A simple "yum install epel-release" is quick and requires no context switches or anything else, which significantly reduces latency. I think the "CentOS edited" thing is probably out of the picture, as you'd then be just creating your own custom RPM anyway, which would be much more overhead. As far as the packages becoming outdated, you are right, and I think part of the assumption here is that a CentOS repo with these packages would be automatically mirrored from their parent repos. I just figured that was a given. ❧ Brian Mathis