This was the purpose of the various branching proposals. The main issue there are
not enough time/manpower resources to make any of the proposals work as it needs
build system changes, a full time release manager and packagers who want to deal with it.
In the end though this is a volunteer project done on the side by various people in the
Fedora community. If we were to move this elsewhere, it would still be a volunteer activity.
I wish I had a better answer but after seeing us break the world every RHEL-X.Y release.. I don't.
Indeed, it's a thankless job.
Sincerely, keep up the good work, and that goes for all maintainers too.
I've always advised those running 'downstream' versions wait to update until they are on the same, new 'Update' after any new RHEL release. It's a good rule to follow -- e.g., RHEL 7.7 just came out.
- "Still lurking" bjs
P.S. For those that don't know, starting a good 7 years ago I brought up the increasing 'breakage' with RHEL as Fedora EPEL build various packages for CentOS because CentOS rebuilding various RHEL add-ons. Simply put, at a half-dozen major Red Hat accounts, I had to warn EPEL would now conflict in various ways outside of the very RHEL core, so RHEL systems couldn't be subscribed to it, and each EPEL package needed to be vetted into an internal channel/repo (e.g., Satellite) before even considering deployment.
I also worked with both Red Hat GLS as well as Partners over the next 18 months to tell them to yank all sections regarding EPEL from Training, Enablement and other documentation, as only core RHEL compatibility was the goal, not all of RHEL.
So at that time I suggested the idea of forking into Emerging Technologies for Enterprise Linux (ETEL) from EPEL. That way, people who wanted newer Fedora stuff or, in the case of CentOS not rebuilding various Enterprise add-ons available for RHEL, or even newer versions of packages in RHEL add-ons, would have a separate repo from EPEL that was never supposed to conflict.
As Stephen, Kevin and others others pointed out for that to happen, increases of resources and volunteer hours would have to happen. And that's just the reality. Not everyone is going to be made happy, given the constraints. So we should all appreciate what we have. I'm just glad EPEL still focuses on at least core RHEL.