Dave Crocker wrote:
> Ahh. I see. You have a complaint about working group process. No, I have an issue with agenda railroading. I feel no need to run for mommy's skirts when things don't go my way, however. > The key point is that, after two years of debate, it is time to declare > that changes to the core DNS protocols, including changes to official > character representation by that core, is no longer the technical path > being pursued for near-term support of internationalized characters. Perhaps this can be represented through some measure other than an unsolicited agenda reconstruction. > What is wrong with that approach is that it does not work. > > Please review the last 11 years of IETF history, in particular the > results of its trying to support multiple "complementary" (but actually > competing) specifications, versus the results of choosing a single > specification. SMTP, IMAP, NNTP are complementary. RIP, OSPF and BGP work well together when needed. HTTP and FTP aren't burning each other down. Golly, it almost seems like complementary is the norm. > Your assessment that it does so at some sort of unspecified cost to some > set of unspecified protocols, that will be created at some unspecified > times and for some unspecified set of functions, is quite seriously > irrelevant to the current requirements. Protocols which comply with BCP18 will use UTF-8 for all protocol data _except_ domain names, where they will require ACE. Going with ACE-only directly equates to two mandatory encodings. This is an incredibly high cost to lay on the Internet for the sole purpose of backwards compatibility, particularly given the fact that there is no reason to limit the Internet to just one choice. Other costs were itemized in the last thread, but we can go through it again if you are up to completing the discussion. -- Eric A. Hall http://www.ehsco.com/ Internet Core Protocols http://www.oreilly.com/catalog/coreprot/
