On Wed, 29 Oct 2003, Tony Hain wrote: > So you want to standardize an implementation & optimization issue ... This > is an optimization that you may want to use in any implementation you do, > but it is not an issue we need to standardize.
Did you notice that the APIs are Informational documents? :-) The key point here is raise the awareness of implementors by first discussiing, then writing a draft, and last publishing it. Whether it says "do this, ignore that" or "you might want to do this, but be aware of foo" is pretty irrelevant to me. > In particular, it precludes > use of IPv6 if a local network has a name service intelligent enough to > provide LL information to the right place. Right, there are tradeoffs here. Unless there is concensus about that this is a generally desirable thing, one way forward could be publishing a tradeoffs document which would just bring up both sides of the argument, why you would or would not want to do it, and let the implementor decide what they feel is more important for them. > Please stop assuming that the > world will not evolve beyond where we are, and trying to prevent it from > doing so. I'm not sure if I'd classify the widespread use of link-local addresses as "evolving".. > Yes we do not want LL addresses to show up in the global name > space, and the current DNS is too lame to prevent that, but that is no > reason to specify that nobody can ever use it. My personal point is that LL addresses are highly problematic for apps, because they'll have to then (at least) parse the zone identifier as well, because e.g. if you are a host and have one physical interface and one tunnel interface, you cannot use link local addresses without additional disambiguation. The only way to ensure that link-locals work seems to be restricting them to the special applications which know how to deal with them. But it's no use trying to debate this argument over here. If sufficiently many folks think link-locals are a good thing, then we'd probably need a tradeoffs document instead of a recommendation/specification document. -- Pekka Savola "You each name yourselves king, yet the Netcore Oy kingdom bleeds." Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings -------------------------------------------------------------------- IETF IPv6 working group mailing list [EMAIL PROTECTED] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
