> There is always manual configuration or some proprietary configuration > method (for example some sort of 'over the air' configuration for > cell phones using SMS). Anyhow, isn't an implementation decision > on what to implement and what not to? All we can do is make recommendations > on what implementations should do.
If part of evaluating the applicability of dns-discovery is based on it being mandated to be the last resort, then it seems like there had better be a standard non-last resort which is tried before the last resort. Otherwise the statement about "last resort" is vacuous and should be removed from the document. Assuming that the statement stays in the document ... Saying that devices which implement DNS discovery MUST also implement a manual configuration method, and only if that manually configured stuff is not set (or fails?) does it fall back to using dns-discovery, would give some meaning to "last resort". But that would place a rather strict requirement on minimal IPv6 implementations since they would need a mechanism to enter the manual information. While cell phones might have ways to manually configuring this that doesn't require a user keying in IPv6 addresses on a keypad, an IPv6 toaster doesn't have such a mechanism. Hence for some devices it might make more sense to require DHCP as the non-last resort. So perhaps requiring either a DHCP method or a manual configuration (where the latter might imply a keypad) is the way to provide meaning to "last resort". Erik -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
