On 02/16/2011 10:50 AM, Larry Vaden wrote: > On Wed, Feb 16, 2011 at 10:29 AM, Akemi Yagi <amyagi at gmail.com> wrote: >> >> Apparently there was some confusion around the release of SL 5.6 >> alpha. Troy Dawson cleared it up in his post to the main SL mailing >> list: >> >> http://listserv.fnal.gov/scripts/wa.exe?A2=ind1102&L=scientific-linux-users&T=0&X=78ED2C774C38226BC0&Y=amyagi%40gmail.com&P=6965 > > I am not confused; as a former summer employee at ORINS and ORNL when > in college 4 decades ago, what's good enough for the national labs is > good enough for me _now_. See > <http://ftp.scientificlinux.org/linux/scientific/5rolling/i386/SL/> > for the gory details of changes since Jan 14. > > Back in the Sputnik days, one had to leaf through meters of Nuclear > Science Abstracts to find applicable gems; these days, Troy announces > on distrowatch. > > I think Troy's (ANL) and Matthias' (CERN) approach is VERY GOOD, > having released preliminary alpha/beta/gamma code a week after RedHat > and we have benefited from that with the only inconvenience being that > the install process starts with boot.iso and required mirroring the SL > repository in order to avoid load on ANL, which is probably > unnecessary given they are on Internet2. > > IMHO, "Complete and correct" doesn't exist in the sw world (vs. > hardware) and Troy's and Matthias' approach is very reasonable and > timely. > > Which leads me to another favorite point: has anyone calculated the > average age of RHEL at release time? There is nothing wrong with their approach. However, CentOS has dozens of internal servers and millions of machines that update against CentOS repos on our trees that mirrored external of CentOS. We do not want to distribute things we think are broken or not complete. WRT the age of RHEL ... that is what enterprise Linux is. Fedora (or Ubuntu non LTS, or opensuse, or Debian SID, or any number of other alternatives) exist if you don't want the more stable (ie, older) items. Again, nothing wrong with their approach (I like Troy in any dealings we have had), however it is not what CentOS does or is going to do. When we release, we basically loose meaningful access to our machines for a week as dozens of internal servers, hundreds of external mirrors, and millions of individual machines get updated. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 253 bytes Desc: OpenPGP digital signature URL: <http://lists.centos.org/pipermail/centos/attachments/20110216/cf37ddbd/attachment-0005.sig>