Well, I agree with Christian's responses. We need to prevent panic
(people rushing to switch off FEC0 immediately) and we need to prevent
people continuing to write code to support it. I find it hard to see
any wording that is better than the current draft. The IETF often
specifies what must or must not be implemented.
I also don't think we should rewrite all the RFCs that refer to SL.
I have no problem with listing them, as in
Note that the following documents refer to link local addresses
and will require appropriate updates: [RFC xxx], [RFC yyy],...
Brian
Alain Durand wrote:
>
> Christian Huitema wrote:
>
> >>Suggested text to address both comments:
> >>
> >>
> >>
> >>"The special behavior of this prefix MUST no longer be supported."
> >>
> >>
> >
> >Well, we did not intend to force every implementation developer to go
> >fix the problem immediately, recall the products that have already
> >shipped, etc. The "new implementation" piece is meant to convey that
> >meaning. Just dropping it does not really solve the issue.
> >
> What you are describing here will be accomplished by using a SHOULD
> statement
> instead of a MUST.
> "The special behavior of this prefix SHOULD no longer be supported"
>
> >>>Well, we still have link local scope, so there is still that. Do you
> >>>suggest that we write a line for each of the RFC that currently
> >>>
> >>>
> >mention
> >
> >
> >>>site local and explain how to change them?
> >>>
> >>>
> >>>
> >>Precisely. There are not that many of them. If you go through the
> >>archives,
> >>you'll find a post where I made the list of places that either
> >>
> >>
> >mentioned
> >
> >
> >>FECO:: or site local. Or a simple grep in the RFC pages will do.
> >>
> >>
> >
> >I would rather use a catch-up phrase than an exhaustive list. Something
> >like:
> >"The special behavior of this prefix MUST no longer be supported in new
> >implementations or in new protocol definitions. References to this
> >prefix should be removed from IETF documents when these documents are
> >revised."
> >
> Well, the list of specs that are using SL is really not that long...
> and if you do not say how "new" implementation should handle those
> "old" specs now that SL are gone, you only do half of the deprecating job.
>
> - Alain.
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[EMAIL PROTECTED]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
