At 11:45 AM +0900 1/26/01, [EMAIL PROTECTED] wrote: >The problem is we don't know yet what the long term solution would be. >In that context, I'm not sure whether the proposed solution is a >solution for a short >term (it could lock us up longer than we expect as Klensin pointed >out) though I agree that the ACE-based approach is a simple solution >to allow users to use IDN with the current DNS protocol. Well, I certainly won't speak for John, but but I don't remember him speaking about a short-term solution locking us up. He said that an even half-successful short-term solution will make it less likely that users will pressure us to to a good long-term solution. He also warned strongly about any short-term solution that changes the DNS in a way that makes it hard to implement a long-term solution. I agree with both of those. The ACE-based solutions proposed so far will not change the DNS in a way that makes it hard to implement any long-term solution. The ACE-based presentation-layer solutions don't change the characters allowed or disallowed in host names (a through z, 0 through 9, hyphen), with the terribly minor exception of host names that begin with a particular weird prefix. Even in the presentation layer, these solutions don't prevent later presentation-layer changes for different encodings. Of the two themes of long-term DNS proposals so far (marked binary names such as UDNS or IDNE, or a radical change to the DNS such as a new class), neither is affected by a presentation-layer ACE-based solution, other than the desire for them will be slowed. It is easy to imagine that a long-term UTF8-based solution would completely coexist with an ACE-based solution in the presentation layer and on the wire, with no conflict at all. Even the directory-based solutions that have been discussed but not well documented would not be affected by an ACE-based solution, and in fact might be helped, because people will see that the Internet is truly international much sooner. (As you can tell, this is, so far, my preferred long-term solution.) Am I missing some significant effect that ACEs might have on other long-term solutions? --Paul Hoffman, Director --Internet Mail Consortium
