Bob Hinden wrote: > > Folks, > > In my personal view, while this is a nice theoretical problem, no > evidence has been presented that anyone building an > implementation is > confused about it. For example, were there ever any > interoperability > problems because someone did this wrong? Common sense would > indicate > that IPv6 has the same "endianness" as IPv4. > > I have hard time understanding why documenting this further this > would be a good use of anyone's time (i.e., individuals, IPv6 w.g., > ADs, IETF, RFC-Editor, ....).
Bob, since I brought this up to the wg, I'd like to say that this has been an issue repeatedly. In one case, interoperability problems existed already, when one organization decided to use little endian in certain variables of a multicast message (for navigation data). Fortunately, they corrected the problem rather than force everyone else to comply. But in several other cases, the question "will this be big or little endian?" has come up, and I had this handy-dandy RFC 791 appendix to use to put an end to the discussion. This has happened over the years, the most recent examples being just weeks ago. So yes, it's an ongoing problem. Not everyone out there who writes IP applications is familiar with IP conventions. These aren't necessarily people who write applications for the office or the WWW. And it is very difficult to convince those who already "have all the answers," and whose processor today may be a little endian machine, that really in the bigger perspective they're better off using the industry standard. As we are migrating to IPv6, the same questions will continue to come up. And as of now, other than referring people back to an RFC written for IPv4, or to a Wikipedia article kindly pointed out by Jeroen, there is nothing in writing to give any guidance for IPv6. Bert -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
