On Mar 30, 2010, at 12:40 PM, Edward Lewis wrote: > Because it recently dawned on me that biasing in favor of v6 is the > root of the evil, not the run out of v4.
Biasing stupidly is the root of all evil. Stuart Cheshire explained how to make this work, and it's going to be a standard feature in MacOS 10.7, but there's no reason why we can't fix ssh now. Your syscall, or library call, should be saying "here's a name, give me a connection to it," not "what's the IP address of this name" and then "now make a connection to that IP address." Your argument boils down to saying that as apps become IPv6-capable, they should still prefer IPv4 if it's available. But that's just squandering an opportunity for win. The right thing to be doing is fixing applications so that they do the *right* thing, not fixing them so that they do something very much like what they were doing before, only with a different bias. The question of whether to break AAAA records is one small piece of this larger discussion, and if you make a choice here that prevents the right thing being done at the higher layer, that's a *really* bad outcome. _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop