On Thu, May 21, 2015 at 1:36 PM, Martin Bjorklund <m...@tail-f.com> wrote: > Andy, > > I don't think the implementation burden on the server is that heavy. > A YANG 1.0 compliant server today supports: > > leaf a in module A of type foo from foo@2001-01-01 > leaf b in module B of type foo from foo@2002-02-02 > > I.e., it supports different leafs in the combined data model referring > to different versions of the same type. > > A YANG 1.1 server would also support > > leaf a in module A of type foo from foo@2001-01-01 > leaf b in module A of type foo from foo@2002-02-02 > > Of course, how much a specific implementation is affected will vary, > but I don't think the concepts are *that* different. >
I am not that concerned about the implementation details. Our server fully supports multiple revisions at once. I am more concerned with the increased complexity due to multiple imports of the same module. I think Y45-03 is sufficient and Y45-04 covers 1 or 2 corner-cases that don't justify the increased complexity. In the hands of experts it won't be a problem, but most YANG modules are not read or written by experts. We don't have any experience using multiple revisions at once. Even the example touted as the reason we need Y45-04 was pointless (because ip-address the same in both revisions). I am not convinced it is really needed. I am much more comfortable adding a new feature like "actions", because there is significant deployment of this solution approach. > > /martin Andy _______________________________________________ netmod mailing list netmod@ietf.org https://www.ietf.org/mailman/listinfo/netmod