On 2015-07-08 10:08:07 -0400 (-0400), Monty Taylor wrote: > On 07/08/2015 01:51 AM, Tristan Goode wrote: [...] > > So let's give up naming things like toys in our play crib, start acting > > like grown ups and using semver or plain old integers. At worst, we might > > have some people see bugs in version 13. > > We already use semver. This is not a realistic solution to anything. To be slightly more specific... http://git.openstack.org/cgit/openstack/cinder/tag/?id=7.0.0.0b1 http://git.openstack.org/cgit/openstack/glance/tag/?id=11.0.0.0b1 http://git.openstack.org/cgit/openstack/neutron/tag/?id=7.0.0.0b1 http://git.openstack.org/cgit/openstack/ceilometer/tag/?id=5.0.0.0b1 http://git.openstack.org/cgit/openstack/heat/tag/?id=5.0.0.0b1 http://git.openstack.org/cgit/openstack/nova/tag/?id=12.0.0.0b1 http://git.openstack.org/cgit/openstack/trove/tag/?id=4.0.0.0b1 http://git.openstack.org/cgit/openstack/horizon/tag/?id=8.0.0.0b1 http://git.openstack.org/cgit/openstack/sahara/tag/?id=3.0.0.0b1 ...et cetera. Is the suggestion to take the mean of those release numbers to determine an "OpenStack" release identifier, or maybe a sum total? (This is of course a rhetorical question.) We're naming a release cycle which tracks the development process of a bunch of different releases of discrete components each with their own version numbers. Also, lets stop insulting people on public mailing lists by implying that they or their processes and customs are infantile. It's not constructive at all. Further, we should pick one mailing list on which to have this discussion rather than cross-posting to four. -- Jeremy Stanley