>> (2) SHOULD RFC 6126, IPv4 subset; > Why not MUST? [...] I don't think the prodding should be done by causing > unnecessary pain for average consumers.
Fully agreed, but I'm not sure what is the WG's thinking about IPv4. RFC 7368 (Homenet arch) is conveniently vague about IPv4. My reading is that the authors of RFC 7368 expected that router vendors will use multiple NAT for IPv4, and that they will ignore whatever we mandate. OTOH, the HNCP draft does (mostly) the right thing wrt. IPv4, so I guess we could ignore RFC 7368's lack of assertiveness here and say that RFC 6126 style IPv4 is MUST, except if IPv4 is not supported at all. (I'm not overly keen on the "MUST implement, SHOULD deploy" style of compromise.) > It's ok for vendors to ship with IPv4 disabled by default That's policy, in my opinion it's HNCP's job, not Babel's. Babel should run with IPv4 unconditionally enabled, if HNCP chooses to publish an IPv4 prefix, who is Babel to disagree? I guess we're in full agreement here, just looking for the right wording. -- Juliusz _______________________________________________ homenet mailing list homenet@ietf.org https://www.ietf.org/mailman/listinfo/homenet
