John R Levine wrote: >> the ietf's stated purpose is to ensure interoperability. for many years >> we all treated that as "make sure everybody agrees as to the meaning of >> what's on the wire." the .ALT (or .EXTERNAL or whatever) specification >> will not change what's on the wire, but its purpose is still to ensure >> interoperability. simply put, we want any internet-capable device to >> have the same experience when using non-internet naming. > > Sure, we all want that, but an IANA registry that only lets one entity > register each name would be counterproductive, because names have > semantic meaning.
you must have misunderstood me. i'm arguing for <localpart>.<uniquename>.EXTERNAL, where the <uniquename> is registered with IANA on a FCFS basis to prevent collisions, and <localpart> is not subject to standardization or even documentation (so, the defining RFC can be one page long and just say, "this is for Tor"). > Nobody cares whether they get rrtype number or port number 42103 or > 42104. But if two people show up with something that uses sex.alt, > and IANA tries to tell one of them to use sey.alt instead, fat chance. IETF is not a regulator and there are no network police or armed forces. you can use whatever identifier you want, and if you can convince others to share that use with you, that's all fine, even if it conflicts with the use many others do make, or might wish to make, of the same code point. IANA has no power to tell anybody anything. what IANA can do and should do is, allow those who want to cooperate and who want to avoid conflict, to register their unique identifiers so as to protect their applications and users against unintentional collisions with other applications and developers who also wish to cooperate and who also wish to avoid conflict. > ... > That's why I think that a FCFS non-exclusive registry could be useful > so that people who want to avoid collisions can find infor that lets > them assess the collision risk for their favorite names, and it > doesn't matter whether it's at IANA or somewhere else since it's just > voluntary advice. Since we are not the Domain Name Police, if people > want to create services with names that collide, they will do so. > It'd be nuts to pretend otherwise. i am not pretending otherwise, though i may be "nuts" for other reasons. since IETF seeks interoperability above all, a non-exclusive registry would serve no purpose. -- Paul Vixie _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
