On Tue, Sep 10, 2013 at 3:57 PM, <mohamed.boucad...@orange.com> wrote:
> I have considered that Lorenzo. “is not required to deploy IPv6” would be > accurate if this document is dealing only with dual-stack, but this is not > true for the IPv6-only mode. The set of SHOULD recommendations are > targeting that deployment model. > I disagree. By my reading, you can make a phone that works perfectly well on an IPv6-only carrier network without implementing #2, #3, #9, #10, #11, #12, #14, $15, #16, #17, #18*, #20, #21, #22, #23, #24, #26, #33, and #36. Some of those are MUSTs in this document. If you want to do IPv6-only on wifi you need either #9 and #10 (or both plus #11 as well), and either #20 or #21 (or both plus #23). But the other ones are not necessary to deploy an IPv6-only phone. One of your co-authors will be able to confirm this: I'm told there are multiple IPv6-only phones on T-Mobile USA today, and I'm sure none of them implement all the requirements in this document (or even all the MUSTs). [*] How did #18 even make it in? What use is a MAY in a requirements document? Of course implementors MAY do anything they want, unless they SHOULD NOT or MUST NOT.
