On Thu, Dec 04, 2003 at 01:42:36PM +0200, Pekka Savola wrote: > On Thu, 4 Dec 2003, Tim Chown wrote: > > On Thu, Dec 04, 2003 at 01:25:35PM +0200, Pekka Savola wrote: > > > > > > Also, there are basically two versions of "DHCP": the one specified in > > > RFC3315, and the "stateless DHCP", in IESG review at the moment. It > > > is not clear to which you're referring to here. > > > > Does that matter to the client? > > The sentences start basically like, "If the node implements DHCP, it > MUST/SHOULD do foo". > > Does a stateless DHCP count as implmenting DHCP? Is stateless DHCP > non-compliant with Node Requirements?
OK, so the node may implement the full DHCPv6 spec (for address and other info) or stateless DHCPv6 (only for other info). The implementation of how other info is obtained would be the same. So I agree we should say something like "If the node implements stateful address configuration for DHCPv6 then" i.e. put the language in terms of client functionality rather than whether the server is full or ststeless? Tim -------------------------------------------------------------------- IETF IPv6 working group mailing list [EMAIL PROTECTED] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------