On Sep 18, 2014, at 2:53 PM, Monty Taylor <mord...@inaugust.com> wrote:
> Hey all, > > I've recently been thinking a lot about Sean's Layers stuff. So I wrote > a blog post which Jim Blair and Devananda were kind enough to help me edit. > > http://inaugust.com/post/108 > > Enjoy. I’ve read through this a few times now, and I think I can support most of it. It doesn’t address some of the issues we have, but most of the concrete proposals seem like they take us to incrementally better places than where we are now. I definitely like the idea of making the integrated gate different for each project, based on the other projects it actually integrates with. I could see extending the two-project gate idea for projects outside of layer 1 to include more than 2 projects eventually. The assumption that all layer 1 projects can depend on the other members of layer 1 being present may have ramifications for the trademark “core” definition. I don’t think those cause problems, based on the mechanisms for defining capabilities and designated sections being worked out now, but it’s worth pointing out as something we’ll need to keep in mind. The self-organizing groupings that Vish, John, and others have mentioned may well lead us to create additional trademarks in a more naturally evolving way than the single big mark we’re trying to squeeze everything into now. Does the unified client SDK fit into layer 1 as one of the “common libraries … necessary for these”? Or do we anticipate the services still using their own individual libraries for talking to each other? Having a quality designation will help distros and deployers. I’m not sure we want Cern specifically to be our arbiter of that quality, but I do like the idea of having users be involved in the determination. Maybe it's something the User Committee could help with in a more general way. This proposal only addresses some of the challenges we have right now. If we maintain a big tent approach, and I think we should, we still need a way to implement cross-project policies and initiatives outside of the scope of any one of our existing programs. I agree with Vish that we need a different name for Layer 1. A name that doesn’t imply “leveling” or “layering” would be good, since some of the cloud-native services don’t build on those layer 1 services. Doug _______________________________________________ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
