In message <20050425180959.GA20220@netbsd.org>,
Bill Studenmund writes:

>On Sun, Apr 24, 2005 at 12:53:01AM -0700, Jonathan Stone wrote:

>
>Would you please stop this line of thought? Yes, David said, "IPv6," but
>as best I can understand his desires and most of the other desires I've
>heard mentioned, the initial, core, key thought wasn't that "what IPv6
>does is right," it's "I hate what IPv4 is doing."


Bill,


No, I think I should not drop that line of thought. There's a
fundamental point here: IPv4 is defined by RFCs, in which the Host
Requirements (RFC-1122) are a key part.

It doesn't matter how much you, or David, or I, or Jason, or anyone
else hates or dislikes how IPv4 is specified, or how much we'd like to
change it. If those wishes happen to violate the specification, then
those wishes are out-of-court.

Unless, of course, one goes through the appropriate IETF channels to
change those specifications by issuing a new standards-track RFC.


As it happens, I have considerable sympathy (and some admiration) for
David's attempt to meet Stuart et. al's I-D by treating zeroconf
(link-local) addresses analagously to IPv6 scoped addresses.  I'd have
to check the I-D thorougly to be sure such behaviour matches the I-D.
(I suspect it doesn't actually meet the requierments to explicitly ARP;
but I'd have to check David's patch to be sure).


But when it comes to ideas which do violence to how IPv4 is specified
to work, then its: BZZZT, thanks very much for playing, game over.

Is that something you can agree with?