----- Original Message -----
Leif Johansson wrote:
> Let's please remember that the architechture devised here will primarily
> be used to
> run _applications_. Long experience shows that private addresses in any
> form are a
> bad idea on the Internet. This was discussed at length on the v6 wg
> meeting last week
> and I believe that deprecating sl is the best we can do.
>
> Let's please remember that the architechture devised here will primarily
> be used to
> run _applications_. Long experience shows that private addresses in any
> form are a
> bad idea on the Internet. This was discussed at length on the v6 wg
> meeting last week
> and I believe that deprecating sl is the best we can do.
>
I am willing to admit that there are big problems
with private addresses, but a simple search on www.yahoo.com for
+ipv6 +"local address"
The results are 13,400 pages.
We need to define this very clearly and get it out
there as the already too much confusion as to how it is recommended to implement
this. At least two people have posted that they don't think that the fe80::
range will be able to be used due to this prepublication of the concept (see
draft-itojun-ipv6-local-experiment-00.txt for example).
As a recommendation team we need to offer an
alternative not just a "the old way was bad so don't used it anymore" kind of
answer.
I understand that there was a discussion and vote
in SF, great. Who is going to put out a new draft-RFC that will explain this
clearly enough that end-user implementation people (think LAN installation) will
understand it and will be willing to work with it. They have a history of doing
the old way why should they change now?
