On 21 Oct 2011, at 08:21, <teemu.savolai...@nokia.com> <teemu.savolai...@nokia.com> wrote:
> Do you agree that nodes' behavioral differences between "foo" and "foo." > names is out of the scope of this particular MIF draft? In my view neither the draft nor MIF should be encoding any changes to client name resolution behaviour that depend on the "bareness" of a supplied name. Existing OSes have many varying algorithms for deciding which name service(s) to use to resolve any particular name, some of which have user-configurable knobs (e.g. "options: ndots" in /etc/resolv.conf). By all means use server selection once the OS has decided that "DNS" is the correct name service, but that algorithm shouldn't IMHO impact or otherwise specify the prior name service selection algorithm. > There could perhaps be another draft, which would say that if name is "foo" > it should not be appended with search lists but "foo." might? If anything it would be the other way around - 'foo.' is explicitly already full qualified, and MUST NOT be appended with search lists. Ray _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
