On Mon, Mar 09, 2009 at 02:11:13PM -0400, Edward Lewis wrote: > I'm not sure (as this is all second hand arguing). ;) I have gotten the > feeling that some folks, perhaps John/perhaps not, have been arguing that > the specs in 1123 were sacred and ought not to change lest other bad > things happen. IOW, maintaining the status quo of a "working Internet" > was imperative.
No, I think John is arguing that IDNs at the top level are a good idea, but opening things any wider would be bad. I think his position is very similar to what Patrik argued, which was that we ought to require people to show that there is not a harm before allowing the innovation at the top level, rather than presuming that there's no harm and reacting if we run into one. > I suppose my confusion is now a bit deeper. If A-labels conform to the > rules in 1123 and all U-labels can be translated to A-labels, is BiDi an > issue (for the DNS)? A labels do _not_ conform to 1123. 1123 says "alphabetic", and all A-labels have at least two hyphen-minus ("-") characters in them. That's part of why we have a problem. > If A-labels are what is in the DNS, isn't the DNS taken care-of? Maybe. That's what I was asking in idnabis about. It isn't clear to me yet whether it is possible for an A-label (which is always xn--[output from Punycode algorithm]) to end with a digit instead of always a letter. I _think_ it's always a letter, but I'm not totally sure. You can definitely have digits inside a Punycode-encoded label, because I've seen them. We may want to restrict any A-label that happens to end with a digit anyway, because of what might happen if that label is exposed in A-label form, but in a BiDi context. > BTW, I was under the impression that ICANN wanted a root zone without > any confusion, no homonyms, etc., but the IDN test bed has that - the > two Chinese-language[0] entries are pronounced the same but written > differently. I am happy to report that I do not know anything about what ICANN wants in respect of the contents of the root zone, and that I don't yet think I need to learn. A -- Andrew Sullivan a...@shinkuro.com Shinkuro, Inc. _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop