On Thursday 02 December 2010 14:38:39 Thomas Thym wrote:
On Wednesday 01 December 2010 16:51:49 wrote Slaya Chronicles - Geeko Acolyte:
I may be wrong but RH is where it is today is partially due to CentOS.
CentOS, by more traditional business models should have killed RHEL years ago. I mean, who would want to pay for something (SLES/SLED) you can get at no cost (CentOS) right ?
The opposite is true now. What we have now is a sizeable pool of RHEL-type system trained IT guys. And guess what these people will recommend for paid-for support Enterprise Linux?
=Asking= I can clearly see the advantage of a LTS system.
No doubt.
My impression of that discussion is, that it is worth it to ask Novell what they need to support the openSLES idea. For the openSUSE-Novell-openSLX-every community member ecosystem this seems by far the better alternative.
=Thinking further= Another idea which came to my mind. I see manly two types of openSUSE users with their special needs. 1. Stability, security, long-term support, not much work maintaining or migrating the systems. This will apply to many servers or desktops from business users (esp. bigger ones). 2. Stability, security and software that is up to date. They update their systems regularly. (Power-users, Home-users, etc.)
In my opinion the amount of people wanting a mixture is small (but that is just a guess). So what does the user want? 1. => He/she wants an LTS version and is happy when a new version is published, when end of life of his version is approaching after 3 years. openSUSE LTS/openSLES would be the choice. 2. => They will be happy to use openSUSE Tumbleweed. It fits exactly their needs.
=Consequences= When the majority of users is satisfied with LTS and Tumbleweed why not change the release cycle to focus our resources for these new tasks? Now we have 8 month until the next release. We support every release for 18 month (2 releases and 2 months). => we have 2.33 main versions to maintain.
So why not expand the release cycle to 2.5 years and support it for 3 years?! This would be fine for group 1. For the second group we offer Tumbleweed based on that LTS main release. To fill the gap between we could provide "snapshot"-CDs of the Tumbleweed tree (if really needed) to offer a live and install medium for new installations / new users. There may be a 3rd tree like factory / testing. That might reduce the number of repos as well. You could get rid e.g. of all the openSUSE 11.x/KDE_Factory repos. And we would have again about 2 main versions to maintain.
Practically, a long time support is possible already. I offered multiple times to create an OBS project where people could still maintain (security) updates for openSUSE 11.0 for example. However, so far no real group got established. So it looks in first place as a lack of workforce to me. But again, there is nothing what stops you to do this already _right now_. However, I don't see that this would obsolete KDE:Factory openSUSE_11.1 repos for example. Because you still want to offer the user the choice to have either something rocking stable with little changes and on this other hand with some bleeding edge component. bye adrian -- Adrian Schroeter SUSE Linux Products GmbH email: adrian@suse.de -- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org