Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id RAA15693; Thu, 1 Jun 1995 17:44:24 -0400 X-Resent-To: drums@CS.UTK.EDU ; Thu, 1 Jun 1995 17:44:23 EDT Errors-to: owner-drums@CS.UTK.EDU Received: from CU.NIH.GOV by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id RAA15686; Thu, 1 Jun 1995 17:44:21 -0400 Message-Id: <199506012144.RAA15686@CS.UTK.EDU> To: drums@CS.UTK.EDU From: "Roger Fajman" Date: Thu, 01 Jun 1995 17:42:14 EDT Subject: Re: address syntax > IPv6 is the big unknown looming here. > > I saw one specification for "native" IPv6 addresses which had > hexadecimal numbers and colons. The idea of permitting colons inside > a domain is a complete nonstarter--it breaks existing code for > dealing with route-addr. That's what it uses. The basic syntax is groups of four hexadecimal digits separated by colons. Leading zeros in a group can be omitted. "::" can be used once in an address to represent enough zero groups to make the address 16 bytes long. I sent ABNF for it to mailext a few months ago. I can dig it out if it's wanted. There was a fairly lengthy discussion within the IPv6 WG in San Jose about the effect of using colons in the addresses on other things. Domain literals and URLs were mentioned. Various alternatives were discussed, but there were problems with all of them. The decision then was to stick with colon and let other things adjust. As far as I know, that hasn't changed. Domain literals with colons inside the square brackets are certainly unambiguous syntax, but I don't know the effect on parsers in existing implementations. It's likely that it will require some changes to any implementation to accept IPv6 domain literals, whatever the syntax, if they are allowed at all.